Please welcome Principal Military Deputy Assistant Secretary of the U.S. Navy for Research, Development, and Acquisition, Vice Admiral Seiko Okano.
Good morning. Thanks for having me here. Under Secretary Phelan’s leadership, the Department of the Navy is changing how we do business. We're done waiting for purpose-built government solutions when the best technology in the world's already proven in this room. We need speed, we need scale, and we're gonna get this by deploying commercial AI directly into the maritime industrial base. This is what ShipOS is and why we're here together today. What you're about to see on the screen are capabilities of ShipOS, the Department of the Navy's AI-enabled operating system for the naval shipbuilding enterprise. We started deliberately, two shipbuilders, three public shipyards, and 18 suppliers, all focused on submarine construction and maintenance.
That's not the full industrial base that the Navy has. That's just the beachhead. From that beachhead, the agents are already working, surfacing schedule opportunities, flagging material risk, identifying capacity that can be reallocated before it becomes a gap. Secretary Phalen has been unambiguous. We measure by outcomes. Does the fleet have what it needs? ShipOS will enable us to answer that question continuously with real-time visibility across shipyards, key suppliers, and production schedules. I wanna show you two workflows inside the platform. Now we're inside a supplier's view of their production timelines. Anyone in this room who's worked in manufacturing or defense acquisition knows what an engineering change notice means in practice.
It means a phone tree, emails, somebody manually updating the production schedule, calling procurement, tracking down the shop floor supervisor, trying to hold all of it in their heads long enough to figure out what breaks. In naval shipbuilding, a single change notice can cascade across dozens of suppliers, hundreds of work orders, and a production timeline measured in years. Watch what happens when the same change notice enters ShipOS. What used to land on somebody's desk as a problem now arrives at, as a decision with context, options, and trade-offs already mapped. Every purchase order downstream of that change is immediately flagged. Every production stage affected is visible.
The system doesn't wait for a human to map the cascade. It already has. ShipOS opens the change notice, and the AIP agent starts running, parsing source documents, cross-referencing the bill of materials, identifying every downstream dependency simultaneously in parallel. Think of it as a playoff run. Every play calls the next. A missed block in the first quarter doesn't just lose a down, it changes what's available in the fourth. Now, here's where the human comes back in. ShipOS presents three paths forward. Act now, minimal schedule, and cost impact. Defer, defined cost growth, defined schedule risk. Reject and escalate, full manual review, maximum exposure. Each path shows the trade-offs clearly, days, dollars, risk score.
The program manager sees the full picture, makes the call, and approves. Agents execute. Systems update. Notifications go out. The production plan is revised. The result, a decision that takes minutes with a complete audit trail that documents itself. Now I wanna show you something that will look very familiar to everybody in this room, an inbox. This is the ShipOS intelligent comms pipeline, a live view of incoming supplier communications automatically triaged as they arrive. Some threads are already being worked by agents. Some are flagged for human review. Some are already resolved without anyone touching them. Now watch what happens when we open one.
A message from a lead mechanic on the shop floor. Two issues. Equipment shows signs of wearing, pulling more power than baseline, same pattern as a failure they had had before. And a material shortage, parts needed for work starting Monday that haven't been confirmed yet. 16 workers potentially standing idle if it isn't resolved. ShipOS reads the message. It resolves shop floor shorthand, informal references to locations and equipment back to the actual assets in the system. It pulls the relevant telemetry. It finds the matching failure patterns and historical maintenance records. It checks inventory, confirms the shortfall, locates an inbound shipment that covers the gap, and stages a preventative maintenance work order. It drafts the response.
One email, two problems that would typically require separate people, separate systems, and hours of back and forth, handled in the time it takes to read the message. That means every supplier, every subcontractor, every program manager, regardless of their own digital maturity, can participate in an intelligent, connected workflow simply by doing what they already do, sending emails, sharing documents. ShipOS meets the industrial base where it is, and it lifts the entire enterprise. It doesn't ask them to change how they work. It meets them where they are and makes every message they send more actionable than it would ever been on its own.
What you're seeing now is where it all comes together. Every automated change assessment, every intelligently routed communications, every agent orchestrated action, all flowing into a single dynamic production schedule that reflects the true state of the program. A more accurate schedule means risk is identified earlier, intervention happens sooner, schedule recovery is reduced, cost growth is contained. Ultimately, it means this. Capability to the fleet delivered at speed, not the software, not the operating system, the ship in the water, the submarine at sea, ready to execute when the nation calls. The Department of the Navy has made a strategic bet.
The faster path to a stronger fleet runs directly through a smarter, more agile industrial base. ShipOS is how we turn that decision into results. The next era of naval power will be shaped not only in the fleet, but in the shipyards that build it. We've seen how AI has transformed the fight. Now it's transforming how we build. ShipOS gets more capability as our partners work with us to deploy it, each optimizing their own operations, maintaining full control of their data and security, and working alongside systems that already have.
The Navy has made its commitment and is transforming how it acquires and builds capability for the fleet. Rebuilding the maritime industrial base will require the innovation, partnership, and leadership represented in this room, and that work is already underway. Thank you.
Please welcome President and Chief Executive Officer of World View, Ryan Hartman.
World View exists to inspire, create, and explore new perspectives for a radically improved future. I know that's very aspirational, and so it's my job here today to make that real. Right now above us, high in the atmosphere is the stratosphere. It's an operable domain. Most of human history has largely ignored the stratosphere, too high for aircraft to loiter, too low for satellites to orbit, just an atmospheric layer that you learn about in school, not well enough understood to operationalize. World View exists to change that. We build stratospheric platforms that can stay over an area for a very long period of time, 30 days, 60 days, or more.
We do this by leveraging four directional winds that exist in the stratosphere. By controlling our altitude with ballasted air, we can freely navigate in the stratosphere, and we can do what we call station keeping. It's simple in principle, but complex in execution because you're steering a month-long mission using the physics of the atmosphere. Navigation becomes a vertical decision that creates a horizontal result. Stratospheric sensing delivers imagery that is higher resolution than any satellite with much longer endurance than UAVs. UAVs can get close and they're precise, but their endurance is limited to hours. Satellites can be global and predictable, but they also have fixed orbits and limited revisits.
I'm happy to report that the stratosphere is now open for business. For our customers, this isn't a debate about which domain wins. It's about getting the predictability of a satellite, the precision of a UAV, and the enduring presence of a stratospheric platform, all stitched together in one operational picture. Because when minutes matter, data isn't the product, decisions are. Here's the core idea I want you to take away from us today. The stratosphere is operational, and AI-driven workflows built on Palantir are how we scale it. Let me show you why that matters, beginning with a hard constraint we used to accept.
Previously, from initial flight plan to launch, it can take us no less than two weeks. Two weeks works for a scientific payload, but 2 weeks is unacceptable for an operational demand that shifts by the hour. Our first focus with Palantir was simple: compress time without compromising safety, compliance, or mission objectives. That's where our journey with Palantir began. In January, we partnered with Palantir. In a very short period of time, we delivered AI Flight Director powered by Palantir Maven's mission planning. Here's what that means. A stratospheric mission isn't a point-to-point flight. It's a multi-week, sometimes multi-month operation conducted at the edge of Earth's atmosphere, riding winds that don't care about our schedule.
Our customers obviously do care. They care about exact mission criteria, things like coordinates, dwell time, sensor coverage, and the ability to adapt when the world changes. We use AI Flight Director to simulate optimal trajectories that meet customer objectives while operating in a dynamic, sometimes unpredictable environment. We do it in a way operators can trust. They can audit it, and then they can execute it. Now the second piece with Palantir. Planning is only half of the battle. I'll tell you, the real fight is disruption. A sudden SIGMET, a storm cell that accelerates, a wind shift that turns a safe trajectory into a risk, or an intelligence briefing, and the AOI has shifted based on new intelligence.
Systems must be immediately retasked to transit to the new location, and historically, that becomes a scramble using spreadsheets, computer modeling, phone calls, manual replans, and time that you just do not have. With Palantir, we connect live telemetry from the balloon's sensors and combine that with live meteorological streams and operational constraints, and then we make the operation AI aware in real time. Instead of an operator hunting for the right data, Maven's mission planning agent surfaces the alert, it contextualizes it, and proposes a set of actions. Approve a reroute that an AI Flight Director already modeled. Adjust to catch a different wind band.
Update a constraint and see downstream effects instantaneously. This is where the stratosphere ceases being a platform that collects data and starts becoming a platform that participates in decisions. The outcome isn't just speed, it's safety and predictability at scale. At the end of every flight, we take what happened, we take what we learned, all the things that we decided and what resulted, and we write it back. With that, every mission, the Ontology becomes a living memory of the operation. Past events, decisions, and outcomes enriching every future flight plan and execution. That's what we've built so far, AI Flight Director that compresses time and makes execution adaptive.
Now let's talk about what happens next, because this is the part that changes the economics. It's easy to celebrate a single launch, but the scaling story is completely different. Palantir-powered workflows don't just make one launch faster. They make dozens or 100 simultaneous launches possible with the same operational efficiency. What they do is operationalize a swarm. A swarm is not a marketing term, it's a control problem. When you have many assets in flight, you need fleet orchestration, automated deconfliction, dynamic retasking, and prioritized attention so humans are focused where judgment is required. We can now see the path to that system.
Imagine this, multiple stratospheric platforms in flight, each streaming sensor data, each running edge inference, each capable of being retasked in minutes. A potential object of interest appears, the platform detects it, the system distills it, and an AI agent contextualizes it. The operator gets a recommendation with the rationale, and the action is executed. That's the future of in situ decision-making. It matters because it changes what's strategically possible. We had a roadmap to get there in 2030, and we can now see that possible in 2026. That's why we're partnering with Ondas. This partnership immediately creates a multi-domain ISR alliance and a suite of offerings that combine World View stratospheric balloon systems with Ondas’ expansive portfolio of unmanned systems.
This partnership accelerates something we already believe, that the future is multi-domain, executed as one coordinated system. Here's what changes. The same Palantir native workflows that plan and operate Stratollite become the backbone for orchestrating multiple domains, stratospheric platforms, UAVs in groups one, two, and three, and space-based assets. Today, across the industry, systems have always had their own walled garden. Separate data feeds, separate tools, separate teams, separate timelines, and customers end up patching together fragments just to infer what's happening. We're building a system of systems, unified intelligence with a single pane of glass. One operational picture, one workflow language, one set of decision loops that can task the right asset for the right moment.
If you want persistence, the stratosphere holds the station. If you want precision, a UAV gets it done. If you want global context, space completes the frame. The operator shouldn't have to care which domain the data came from. They should care whether the right decision is made fastest and it's accountable. Let me close where I started. We exist to inspire, create, and explore new perspectives for a radically improved future. That future requires persistent awareness, resilient economics, and decision advantage. Battles are won with power. Wars are won with economics. Stratospheric platforms are low cost, persistent, hard to counter without expensive specialized systems. With Palantir, we turn that persistence into operational tempo. With Ondas, we expand that tempo across all domains. The stratosphere is open for business.
Please welcome Vice Chairman of Freedom Mortgage and Chairman of Motor, Michael Middleman.
Good morning everyone. At Freedom Mortgage, our mission is to foster homeownership in the United States. Homeownership is the key to Americans building financial wealth for themselves and their families. At Freedom as the largest FHA and VA lender and servicer, we know a few things. What we do know is that Americans very much live paycheck to paycheck. They have less than five days worth of daily expenses available in their bank account. We also know that Americans have the ability to borrow up to 90 % or more than 95% of the price of their home, their new home, or one they may move to .
H ave access to borrowing rates that the most sophisticated financial institutions have in the country and in the world at a fixed cost that allows these Americans to realize all of the return upside and therefore build wealth for themselves and for their families that they can pass along for multiple generations. Homeownership is the key to that. We also know that one of the largest issues that our industry faces today is an affordability issue. Right now, we are not getting enough homeowners or potential homeowners into a home and their next home. This backlog is serious and substantive.
At Freedom Mortgage, it's our mission to make sure that we can drive the homeownership rate in the United States up even further than it's ever been and help solve for this affordability issue. In order to do that, we have to leverage the most state-of-the-art technology. Ever since this whole AI craze has been unleashed on the world, we saw an opportunity to really solve this critical issue that we know is facing our nation today. We have spent a lot of time, energy, and resources over the last couple years trying to figure out how to do this at enterprise scale.
I'll tell you, it's hard. It's not an easy thing to do. It's not until I got connected with the team at Palantir here to realize how powerful Foundry and AIP can be, as in just the first 90 or so days, we've built such unbelievable, amazing things that should really give us a ton of promise to help solve this issue. It's really been truly amazing. Therefore, we struck a partnership between Palantir and Motor, where I serve as the chairman of one of the fastest-growing outsourcing and technology firms for the mortgage industry and for financial services, to co-build with Palantir Freedom in a pilot with.
I'm sorry, as we built, co-built with Freedom Mortgage and Palantir in conjunction with Motor to bring Freedom Mortgage as our first customer and bring these capabilities to life, not just for the customers that we serve at Freedom Mortgage, but also bring to the broader industry so we can act as a facility for the industry to bring lower borrowing costs, lower rates, and more affordability to the American people. Let's take a look at some of what we've built so far. Okay. So this represents the first set of use cases that we started on, which represents one of the biggest issues that we face from a cost perspective.
From a regulatory perspective, there's massive costs and infrastructure that we have to have. We're constantly being audited by external regulators and agencies. We have internal auditors. We have managers, supervisors, inline QA, QC on a regular basis to make sure that we're following each and every rule that we have to and delivering the best customer experience that we can. What we've done here is created a system that goes back to source docs and creates rules that you can trace all the way back to source docs and information to make sure every loan that we're doing every single day is traceable back to the rule and the source code.
And we have the ability to not just do that in a cohesive and well-orchestrated and efficient manner from a cost perspective, but also from a change management perspective. Our regulations and our rules and our operations and products change on a very regular basis. This could be IT projects that take months or years that we can now take down to minutes, hours, and days. This is a huge thing for not just Freedom Mortgage, but the entire industry as well as any policymakers as they look to institute new rules to help Americans and the American economy faster. Let's take a look at one of the other use cases that we've got. This is our next generation doc extraction.
We have hundreds and thousands of documents that we look at every single week and every single month that tie back to all these rules that we have people saving to folders and check-in and typing in source information to our systems of record and doing all kinds of stuff. Today, we have— with what we've built and where we're going, a complete catalog of every single document and how it works back through our Ontology back to our operations in a very easy and cohesive fashion that allows us to not just save, understand, and contextualize what's in those documents, but turn it all into data that we can create actionable outcomes that drives the efficiencies and costs down for our customer.
Finally, one of the most important things that we've really looked at and focused on, and one of the things that really helps and is one of the first use cases a lot of AI companies have focused on, has been the ability to conceptualize and take calls and turn calls into data. We take in excess of 500,000 calls every single month, and what we try to do here with this use case is connect with our thousands of agents that are out there and bring in what is the current data and information and historical information of our existing customers or anybody that calls in.
What that customer is asking for or looking for today, combined with what kind of value, proper help could we offer based off of the market environment and rules we're able to operate under today. We can contextualize all that information and in a very automated fashion, help assist our agents bring this value in a much more efficient and customer service-oriented way to our customers, delivering tremendous value. It's really all of these things coming together into a flattened Ontology, where each call, each document, each phone call or mobile or whatever situation we have .
In dealing with an event with one of our customers , or somebody externally , to substantiate what we're doing, is a huge event that is now well and uniformly operating inside of the Ontology, which is going to be a game changer for our industry. I believe, and we believe at Freedom Mortgage and at Palantir, this strategic partnership will build the most profound and impactful technology for our industry , that we think will help solve the problem around affordability, drive borrowing costs lower, and have a significant impact to our industry and to Americans in their pursuit of home ownership. It will be the best thing that we've seen in decades. We think that impact is real, and we are on the precipice of bringing that to life right now. Thank you.
Please welcome Chief Operating Officer of SAP, Sebastian Steinhäuser.
Hello. Awesome to be here. SAP software is powering about 80% of global GDP, and our mission is clear, to help you run better and improve people's lives every day. Who of you feels that change is only accelerating for your business? I guess we are all navigating the same reality. Supply chains need to be rewired overnight. Customer expectations change faster than ever, and technology waits for no one. At SAP, we believe there's only one way to get your most mission-critical backbone systems ready to navigate that reality. They need to run in the cloud, modern and up to date, with AI embedded into your flow of work.
That's what our offering RISE with SAP is about. It's not just about modernizing ERP system, it's really about getting ready for AI-powered operations. Who of you is ready for an ERP migration? Yeah, I guess so. I think for most of you, the prospect will feel like Frodo when Gandalf told him, "Hey guy, small guy, you walk straight into Mordor, climb Mount Doom, and destroy the enemy." Why is that? I mean, in many cases, the migration of an ERP system costs ten times the amount the software costs. For many customers, moving their systems is not easy. Why?
Fragmented data landscapes, siloed processes, change management, complex governance. For that, we are partnering closely with our customers to help them on every step of their journey to move with confidence. For common scenarios, we have our own migration tools. For the most complex ones, the right partnerships matter. That's where Palantir comes in. Together, we focus on the most challenging migration scenarios, complex SAP to SAP migrations, non-SAP to SAP migrations. For those scenarios, Palantir's AIP brings agentic capabilities that help customers understand existing landscapes faster, cut through complexity, and make better decisions faster.
In as little as 10 days, you can receive a validated view of what's possible, a sub six-month migration assessment so that you can get to a clear execution plan and a faster go, no-go decision. This is not just theory. We have tried it out with some initial joint customers. Early results from these engagements show more than 99% validation accuracy within just two weeks, more than 70% timeline and cost reduction for the migration, actually based on the customer's view, not on ours, and 2/3 less scarce expert involvement in the migration. Let me give you two examples.
One Fortune 500 company ran a discovery sprint, validated the plan and executed in under four months. Another was able to move from actually doing about five migrations every two weeks to doing dozens and dozens and dozens in a week and less. We are excited for what's next. For us, it's very clear. For SAP, the destination is in the cloud with AI for all of you, and our partnership with Palantir helps you get there much faster. We are united in this mission also together with our ecosystem. I see our friends from Accenture are coming next. We have some exciting news to share there as well soon. Excited to listen to Tracy too. Thank you. We are excited to continue this partnership together. Thank you.
Please welcome Advanced AI Delivery Lead for SAP at Accenture, Tracy Venable.
Hi. Good morning, everyone. Happy to be on stage and following Sebastian in his announcement with SAP and Palantir, and I'm thrilled to announce the partnership that Accenture is going to expand with SAP and Palantir going forward. We're living through one of those moments in time when the ground beneath our enterprises is shifting in an acceleration by forces of AI and technology. Expectations placed on organizations have fundamentally changed. Companies are being asked to move faster, operate smarter, and deliver greater value than ever before. This moment reminds me of a classic book that I read a long time ago. It said, Who Moved My Cheese? Some of you may know this story.
It's about some mice who wake up one day and discover that the pile of cheese that they've relied on has now gone away. At that point, some of the mice freeze, some of the mice look at their friends next door and complain about the fact that the cheese has moved, and some of the mice put on their shoes and enter into the maze, knowing that the way to find the next opportunity is to go looking for it. Right now, the enterprise technology landscape kind of feels a little like that maze to me. Let me start with what we're hearing directly from our clients today.
As you heard from Sebastian, on one hand, all of our companies are trying to modernize our core, moving from ECC to S/4HANA, reducing the number of technical debt items that are available. The other side, we have to increase our AI capabilities and be able to modernize and find the value that all of our suites are looking for. The challenge isn't the ambition that we have. The capital and the speed at which our organizations can reinvent themselves is where we have a problem. What many leaders are realizing is that this modernization is both a technology and a human challenge.
It needs to be carefully shepherded along a journey that many of us have worked to see the value ahead. There's a lot of work ahead and a lot of value ahead, but it's not gonna be simple. Boards are asking us a very pragmatic question: How do we modernize our SAP landscape, move to AI, and reshape our workforce at the same time? How do we do that faster, with confidence, and with measurable value along the way? Our view at Accenture is that the answer is simply not to add AI on top of your SAP journey. The answer is to change the SAP journey altogether.
We need to build organizations that support the future of our enterprises alongside this technology that we're going to be implementing. Companies that are surging ahead aren't running pilots. They build their foundations early and are scaling enterprise-wide. The gap is widening. I spoke to a few of you on the bus this morning on the shuttle, and I hear that this is, I think we're all kind of in the same place. That's why SAP and Accenture has partnered with Palantir. Humans in the lead, but the full weight of Accenture's experience to deploy at pace never before possible. Like Sebastian said, this isn't a theory.
We've had joint teams working side by side, SAP and Accenture architects alongside Palantir engineers, and we're building something entirely new together. We are taking the joint innovation to market not just as an accelerator, but as a transformation engine. SAP sets the future-proof destination. You have SAP in the cloud, you're infused with AI, and Accenture will bring deep industry experience and the ability to turn strategy into results at scale within the unique organizational structures that you all have. Palantir is providing the AI and data foundation that delivers real-time value. This will help our clients lead, let them see their enterprises clearly before migration starts, and guide decisions before transformation begins, and during and continue and optimize after we go live.
You'll have a foundation that is set, all while accelerating delivery and reducing your risk. The result is you don't just finish a migration, you turn an SAP journey into an operational monetization program that compounds your value over time. What does this partnership mean for our clients? Faster, lower risk cloud ERP migrations, better business decisions, and earlier value realization. A year from now, Accenture SAP engagements should all be able to say, "We delivered our clients' businesses value at pace we made possible with Palantir and data capabilities. We were those mice that picked up our shoes, and we walked into the maze, and we discovered a pile of cheese, that allowed us to survive.
We did it faster, we did it smarter, and we're building on everything that came before us." This is our new vision as Accenture, and we have 88,000 employees currently busily working on SAP implementations and have built a Palantir business group that is focused on working with our Palantir partners in order to accelerate the combination of the two. With SAP, Palantir, and Accenture, this is just the beginning, and we can't wait to get started. Thank you. I was asked to stay as I am the last speaker in the first half of this keynote, and so I get to announce the best part, coffee break. All right. Enjoy. We'll talk soon.
Welcome to AIPCon 9. Let's go look at the magic. The theme of AIPCon 9 is magic. It's not because we're gonna try to trick you. This isn't about cheap tricks and pulling the wool over your eyes. This is about real engineering excellence. It's about something that is so crazy that you didn't think was possible, it feels like magic. That is what it's all about. All right, let's take a look at ShipOS. There's some pretty cool stuff here.
Chad.
Patrick, good to see you.
Likewise.
What do we got going with ShipOS? Maybe tell us a little bit about what's going on here.
We have both, ShipOS, so I'm happy to talk about ShipOS, but also playing in the background here is a lot of the work that we've done on Project Maven.
Okay
To really try to communicate how we're connecting the Foundry to the fleet to the flight line.
Yeah, maybe for people that don't know what Maven is, just a short description of what Maven is.
Yeah, our participation in Maven began by deploying computer vision models in support of intelligence analysts who needed to make sense of the world faster. Today it is now, I would think about it as a battlespace awareness tool that users can interact with to either make decisions within the environment or push the ability to make decisions through tactical data links that exists today, really collapsing kill chains for operational commanders to make decisions.
Yeah, it's really about all the way from the data to the logic, when I think about it, and then to the action, right?
Yes.
It's all orchestrated, and there is really the final action to make it real in the real world is all orchestrated through Maven.
It is all orchestrated through Maven.
Oh, that's pretty cool. As we're, like, deploying Maven out into the real world, like, what are some of the things, the benefits? I mean, it's great. Like, I always want more information, but, like, what is the actual tactical advantage that we're seeing with it?
If you think about the world today, I think one of the things that I saw in the work that I did in support of Maven is operators are searching through haystacks looking for needles, and Maven is really reducing the hay in the haystack.
Yep.
It's making operators more effective. I know that Shyam talks about it as, how do you wrap an Iron Man suit around an intelligence analyst or a fires chief? This is enabling them to identify the points of interest or the objects of interest that they care about and rapidly build plans of action, not only around tactical action, but around operational and theater level missions that they need, might need to execute.
Yeah, I think I saw a stat somewhere about where they were talking about normally we would have 2,000 intelligence officers actually trying to do the targeting and looking at stuff. Now that's 20, and they're doing it in rapid succession as well. Like, that, like, doing more with less is really enabling the warfighter to really keep everyone safe and really, you know, go after the mission.
Yeah, I've really been impressed by some of the users that we have in platform. I think the best builders in Maven are the uniformed users who are downloading their intelligence and targeting brain into the workflows and really upskilling both the rest of the workflows and then making decisions faster and even more safely.
It's funny 'cause that's what I see both, you know, on the USG side, but also on the commercial side. It's like I get this question all the time of like, "What are the best builders?" It's the domain experts that actually know what they're doing, that we give them and enable them to have that autonomy to actually go build the workflow that they need. 'Cause they're the experts, and they've been doing it for 20 years, right?
Yep.
It's like now we've enabled their dream vision, right?
You see this beautiful democratization of knowledge across all of the theaters of operation that the U.S. has.
Yeah, that's it. It's the stuff that was institutional, the phone calls, the emails.
Yep
All those things, like actually bringing that dark matter of data and space into the light, into, like. It's really about making more data computable, that we can now have this richer information space, like what is the ground truth as being represented in the platform.
Exactly. While Maven, I think, is how we think about how do operators take action on something, those operators are unable to take action if they don't have the assets that are ready for the fight, and that's what ShipOS is about, is how do we generate the necessary combat power for, especially for our United States Navy to be effective.
Let's shift to that a little bit about maybe, like, what is ShipOS? Will you give just a short description of, like, what is ShipOS? What are we really doing?
Yeah. ShipOS is, I think it's a really ambitious vision set by the Secretary of the Navy, to that point, generate submarine combat power. I'm a prior submariner, so I'm a little bit biased in this. My wife was a submariner. My brother is as well. But I think that the U.S. submarine is the most capable tactical and strategic asset that we have in the country today. We need to make sure that they're on the field. ShipOS is broken into three primary domains. The first is how do we make sure the submarine maintenance is happening on schedule, so we're getting boats back in the water for the deterrence that we need?
The other two pillars are how do we support the U.S. private ship builders, as well as the supply chains that function and provide the inputs that those ship builders need to be successful, inside of an ecosystem where everybody can share data securely and collaboratively, to provide, again, that combat power to the sea.
That's interesting. Yeah, the readiness piece of it's like I have to have this thing in danger of being used, right? That's the deterrent piece. Also, okay, I need to build these things and get this, like, the complexities of that supply chain. I mean, submarines are incredibly complex, and the supply chain of subassemblies and people and groups, it's like this one thing downstream can totally hold up me building more things. I think that ShipOS is about connecting everybody to be on that same back plane, right?
Yeah, I think that's exactly right. One of the things, I can't remember if it was a ship builder or shipyard that said it to me, but how many parts do you need to make a submarine? You need all of them. You can't put a submarine to sea without everything working properly together.
Yeah. I see these stats behind you here. There's 200 hours to 15 seconds in reduction in BOM approval time, so bill of materials. As I'm changing bill of materials, that would normally be engineering review process, manual things, emails back and forth. 200 hours. I mean, I can understand that. But 15 seconds, that means, like, we're reviewing the bill of materials, we're reviewing everything automatically, and it's just happening.
That. Yes. That's exactly right. That's, like, the power of the Ontology, which I'm certain everybody is probably sick of hearing us talk about how do you encode all of the business logic and all of the raw data that you need in order to create a decision? Those 200 hours is that, like, the manual process, the manual churning that I need to do, that a computer can actually do pretty well. As you can see, the decision itself takes about 15 seconds, but the preparation for making a decision used to take 200 hours.
Yeah. I mean, there's just so much paperwork, and like, you can't get it wrong, right? Especially when you're on a submarine below the water and you're in war, like, it has to be right. I think that it's, like, not only that, but the accuracy that I can get to that. Then the faster contract review cycles. I think this is, like, an important thing of, like, understanding the unstructured data that are in contracts or things, like actually we can make sense of that
Isn't it, making more data computable, right? Like, between that and the reduction of late orders, I'm assuming these are all that connected supply chain. It's understanding how things are flowing, right?
This is that connected supply chain. This is understanding how does a ship builder, how does the Navy provide better inputs to the supply chain, so that the supply chain can move faster and less encumbered by needing things like funding in order to, like, put items on the factory floor.
It's not just the, "Hey, I need to get this stuff," it's also giving the demand signal back to your supply chain to say, "Hey, I'm gonna need more of these things. Start. Get the raw materials so you can build these subassemblies for me.
Yeah, you can get going.
It's really the backward and forward communication across my supply chain, right?
Yeah, exactly. You need to have this, like, harmonious loop, I'd call it, cosmic harmony of supply and demand both speaking to each other.
Yeah. Then these last two things down here, 96% project backlog clearance by November 2026. Like, that is where I'm stuck on a project because of some delayed piece or some kind of, like, knock-on effect that I'm delayed. By us connecting those things, we're finding those bottlenecks, getting ahead of them before they actually delaying things.
Yep, that's related to this reduction in material planning time.
Yeah.
Where if I know the material that I need to order ahead of time, I can actually clear my backlogs faster as well. You're seeing this, again, this really good, interrelated, aspect of how do I make better plans and how do I clear my backlogs.
Yeah, 'cause I mean, it's like one thing to make a plan. Something always happens, right? It's really about how do I execute.
Yeah.
I talk about this all the time because it's like demand planning, demand sensing, and it's actually, like, your demand planning is probably not your forecast accuracy, it's your ability to execute a plan, and I think that's the operational nature of Palantir. I wanna drive operational workflows to that when shit happens, I actually can go make the best decision. You, you're closest to the problem, can make that globally optimized decision, right?
Yeah. What we're learning about at some of the shipyards where we're working, one of the shipyard workers said it best, is like, "It's not about making a perfect plan, it's about making the least shitty plan that I can.
You can execute. An executable plan, right?
Yep. Every day, they're working on a bad plan, and they know it, but how do we give them the information to make the best decision that they have or that they can with the resources that they have?
Yeah. This is awesome. Well, Patrick, thank you for taking the time today.
Yeah, thanks.
Appreciate it. ShipOS really is the epitome of how do I create a connected ecosystem where our companies or organizations or institutions like NATO, but also Airbus and others, where I can create this ecosystem where everyone can share across supply chains, securely owning the process and the data. How do I do that at scale? Like, only Palantir can do that. All right. Teton Ridge. Here we are. Hey, guys. How are you?
Good to see you.
How you doing?
You guys are partnering with TWG and Palantir and building some AI models to help you guys with bull riding?
We're taking video and data analysis from the world's oldest sport.
Yeah
Trying to build better bull riders and better interfaces for fans to understand bull riding and western sports in general.
We've just got a sample camera set up at the bulls we were just looking at. Like, I have bull rider analytics and the bulls themselves, so if I can understand the movements. Like, what does that give you?
We're really trying to modernize the sport, you know, and this is gonna be a game changer for us.
They can improve. The common myth around bull riding, you just find the craziest guy in the room and strap him to a bull.
That's him.
It's a sport like every other sport.
Yeah.
There are minute body position details. There are all these different things that contribute to success in bull riding. There's never been any sort of measurement of that.
Yeah.
What we're doing now is we're capturing the skeletal points of the bull and the rider and then doing math calculations, basically around how those positions change during a ride. In the eight seconds of a bull ride, it's tracking where every one of those points go relative to where it started. Part of our job, as the broadcast partner on the rodeo side is to build more rodeo fans. Right now, if you go to a rodeo, you're gonna have a great time and you're gonna see some really spectacular stuff.
Yep.
You're not really gonna understand what you're watching. Nobody understands how the bull ride is judged and why it's easy to ride a bull, easy for guys like this, not me that jumps up and down in place versus the dynamic moves that are made in a bull ride that gets a lot of points. Statistics, information, analytics drives fandom.
Yeah.
That's what we're trying to do with this tool. Not every bull bucks the same.
Yeah.
Not every rider rides the same. This will give Colby an opportunity to dig one layer deeper in the analytics.
Yeah.
It's gonna give a new layer of insight into, you know, what is creating success, in particular match-ups for particular guys. Overall, what's creating success for that guy. You know, there might be something that, you know, that Colby might would miss in a film session with a rider that this technology is gonna.
Catch
Give him the opportunity to catch.
Oh, I mean, that's a good point. I mean, that's a lot of what we see across, whether it's the military side or commercial, it's how do I help people be more strategic by taking the data and the tactical pieces so that you're focused on the strategy about matching up and who goes what. I think that's really cool that now you're bringing this to, you know, the, one of the oldest sports around, right? Thank you, guys, for walking me through this today. This is awesome.
Thank you.
Yeah.
Yeah, thanks.
Appreciate it.
Very cool. This is pretty cool seeing everything come together. Very much when we talk about Palantir as an artist colony, this is what no one else would be doing. Hey, good to see you guys.
Hey, great to see you.
Hey.
Good to see you.
What do we got here? We got an Armada container.
Yeah.
What else we got going on?
What we're talking about today is the Sovereign AI operating system, which is a co-developed reference architecture from NVIDIA and Palantir. It's speed of time to deployment or the time that we can get workloads deployed and operational to provide business value. We're trying to shrink the gap between raw GPU hardware and getting operational, getting business value out of that hardware that you have.
If I need to go deploy Palantir to a remote location in the world on a ship, on an oil rig, wherever.
Yep
We've got a reference architecture with Dell and NVIDIA. We've got your Armada container here that brings it all together, and we can literally just ship it anywhere and have Foundry running anywhere. Like really forward deployed compute.
Yep. The Armada really gives us the opportunity to put us places that might not have a data center, where you might not have reliable communications, or I think the other use case is where you need low latency, basically where you need to process data and move your compute near where the data's being collected, so.
What does the Armada do for us? Like, what are you guys doing in this whole picture?
This is a cruiser. It's a 20-foot Dalean.
Yeah.
It's a full stack modular AI data center, and then we have this reference architecture now with Palantir, NVIDIA, and Dell, where we're bringing this to the edge.
Yeah.
When I say the edge, I mean literally anywhere.
Yep.
Deploying at speed and scale. That's, I think, what this is really about, is Sovereign AI that can be deployed in days anywhere in the world and scaled up as needed. You can run all of your workloads air-gapped locally at the edge.
Wow. Okay, so that's pretty cool to have the capability of bringing all of these things together. Like, what is the value proposition you're seeing when customers can deploy this out to the edge?
One is latency, and you think about situations where split seconds matter when you're using your data, where the utility of the data goes down significantly the longer time goes on. A good example of this is an emergency response scenario, and we, you know, are working with Alaska to enable real-time intelligence at the edge. They previously had over 24 hours of latency to process data from drones.
Yep
You need your data in split seconds if you're responding to avalanches or floods or wildfire fires or other natural disasters. The benefits of it are, one, you get real-time intelligence, so that's the latency piece. The second is you get true sovereignty and sovereignty down to the site level, the ability to deploy these behind the firewall at the edge, still managed centrally. That is unique. Nobody's done that before. Together we're bringing also complete cybersecurity as well.
Typically to run these AI workloads like you have with Palantir and AIP and Foundry, which is now what we're enabling with this architecture anywhere in the world, it's also more cost-effective to do it at the edge. Rather than sending all the data back, you just send the metadata back to the cloud.
That's pretty cool.
Yeah.
I really get that intelligence at the edge.
Yeah.
This isn't just simple AI models. These are some of your NVIDIA models, right?
I think that's the important part, is that we can contain and put all of the infrastructure within these scalable units. I have all my data there. I have full control of my data, obviously within Ontology.
Yep.
It's sort of the perfect substrate for building agentic systems. This will deploy with our open source models, and we can also work with other foundation model providers, put these, you know, at the edge as well, and I think this is important as you're thinking about what type of data you're concerned keeping control over, taking action on.
Yeah, which is interesting, 'cause I mean, our whole ethos at Palantir is how do I take the data, the logic, and then turn those into actions.
Yeah
Meaningful actions in the real world, and I think the latency, the bringing all of this together is, like, so powerful to drive those actions anywhere in the world. How do we deploy these things? Maybe we can talk about that for a second.
Sure.
Like, we're using AIP to actually deploy these things, right?
A while back, NVIDIA and Palantir started working on Chain Reaction.
Yep
We've focused a lot on delivering energy to data centers, building data centers.
How long does it take us? I mean, I heard we've had a lot of time compression here, right? What would it look like?
Well, this one was a little bit magical. We've managed to deliver this sort of soup to nuts in about three weeks, and I think that's really 'cause of the partnership between Palantir, NVIDIA, Dell, and Armada as our container partner as well. That is kind of what we're trying to do with this reference architecture, is take all the guesswork and requirements gathering out so there's a preset recipe, and then we can just deliver directly out to the field, deploy with Apollo, get Rubix installed in order to, you know, get provisions and microservices up. It pre-installs all the NVIDIA libraries. We're really trying to just accelerate the full life cycle and the full stack build-out, all the way from ordering to running your first workload in AIP.
Yeah. That's cool.
Yeah, I think this is a good example of what happens when you have a set of companies, again, Palantir, NVIDIA, Dell, Armada, that come together with a common mission for customers. It really is like magic, you know, going from a concept to deployment of AIP and Foundry anywhere in the world in weeks.
Yeah.
That is magical.
It's how do I take this that's six months down to a few weeks?
Yeah.
The compounding effect that has for customers is really the magic.
Yeah, with data and AI, speed is everything.
Yeah.
This is all about speed to deployment, speed to scale.
These are always large capital investments and, you know, making sure that you get real business value out of it as soon as possible.
Yeah
As soon as you’ve paid for it is what customers are always concerned about, and this, we hope, helps alleviate some of that concern in how will I get my workloads running in the minimum possible time.
Very cool.
Yeah. What I'm excited to see, too, is just the impact that this is gonna have on the ground when you think about a lot of these scenarios, whether you're talking about disaster response or on a ship, you know.
Yep
If you're talking about an oil rig or a mine. You know, you have terabytes of data in each one of these locations. Now unlocking that power of the data at the edge for the first time, huge opportunity.
There's sort of two sides to this. There's really the build part and then the maintenance and use part of it.
Yep.
Here on the supply chain part, what we're doing is we're basically figuring out all the pieces and parts that are needed to build the reference architecture within the Cruiser. We wanna monitor our supply chain, and this gave us a really great opportunity to build agentic logic on top of Ontology. We can pull in where we get new supply chain updates, emails.
Agentic AI running Nemotron open weight models to actually help me go deploy.
To go build an AI factory.
Factory.
Its inception.
Right.
Yeah. Yeah.
I mean, we're a little bit of inception here.
Yeah
This is cool. The fact that I've got Nemotron models integrated with the AIP, I've got my build materials, the Armada components, all these things being mapped out so I get. Everything at the right place at the right time that I need to go whatever, go after my use case.
That's right.
Okay, that's pretty cool.
Maintenance is really about monitoring this long term. Quickly, what we did in the reference architecture, there's basically three T-shirt sizes, small, medium, large. What we deploy in this Galleon is the small.
Yep.
It consists of four 8-way, so a total of 32 B200 GPUs, which are latest and greatest.
Very cool.
Capable of running big, big models.
Yep.
We have a control plane of servers here. We have kind of the live, active health here that we're monitoring, and then we also wanted to build a digital twin. Let me load that. What you see here is that's actually the 3D model of the Armada container you're looking at, and what we wanna do is build out the whole architecture in a digital twin and be able to monitor it live.
I mean, it really goes the last mile here, right?
Yes.
I think, I mean, it's a theme I've seen across all of the customers here at AIP, is I need to make more data computable.
Yeah
I need to be able to represent the ground truth of what's really going on.
Yeah.
I just, I still love the inception of AIP to deploy AIP out of the edge.
Yeah.
This is very cool. Well, thank you guys.
It's cool.
For walking me through all this. Like, I mean, I can't wait till I can go show up at customers with a container full of GPUs, and we go build really cool shit.
Yeah.
Like, that's awesome.
Anywhere in the world.
Anywhere in the world.
Anywhere in the world.
Cool. We have TeleTracking and Carilion Clinic here. Thank you for joining me. There's a big partnership between the three of us in really driving some, like, differentiated care. Maybe you guys can give me like, just a little bit on what are we doing together?
Imagine turning on air traffic control in an airport who didn't have air traffic control.
Yeah.
That's what this is doing. For the referring provider, there's a view in there that they can see all of their patients and what stage the referrals are in, and the referred-to provider can also see the demand coming into them and seeing what they may need to change and looking at those models.
When you put this in context of the comprehensive solution that we've been working, along with Carilion and 1,000 other hospitals, on the inpatient side, and you extend it to what's happening in the ambulatory side, that transparency and that visibility is driving a completely different way of operating, and I think we're knocking on the door of price transparency, in healthcare.
Well, thank you for taking the time to show us a little bit of what you guys are doing and talking through this. I really appreciate it.
Thank you.
Thank you.
Thank you for joining me from Tampa General Hospital. You guys have been really a great partner for us a long time here. Maybe you can give us a little bit of a idea what the use cases we're gonna talk about today.
Yeah, absolutely. This is our care progression navigator, and this is a tool that we built out initially for our case managers as a way to improve their efficiency before multidisciplinary rounds. Prior to building this tool, they would spend an hour and a half going through all of their patients to understand what was going on with them, digging through the chart and finding the information before going to multidisciplinary rounds. Now we've uncovered all of that information for them in two seconds opening up the tool.
Right. They're digging through files before, probably taking notes on paper. I think we've seen.
Taking notes on literal paper, showing up to multidisciplinary rounds with like a notebook.
It's the 1 + 1 equals 3 here that I'm actually human and AI working together is really the benefit.
Absolutely. It's really interesting because when we first went live with this, one of our ICU charge nurses saw it. She was like, "I need to use this for charge rounds." Because they would spend so much time. They had literal pieces of paper, and they would spend an hour and a half, two hours going through charge nurse handoff from day to night or week to week. They saw this tool, and they were like, "We need this tool." It reduces cognitive load, you know, tenfold.
Well, thank you for taking the time today. This is really cool work, really meaningful.
Absolutely. Thanks for having me.
Please welcome Chief Information Officer, Defense & Systems at GE Aerospace, Jess Salzbrun.
Good morning. Every two seconds, an aircraft powered by GE Aerospace technology takes off somewhere around the world. GE Aerospace powers two out of every three aircraft in the U.S. military fleet, and we have been powering the aerospace industry for over 100 years, grounded in a culture of innovation. In 2024, we became a standalone fit for purpose aerospace and defense company, which means our focus has never been sharper and our mission has never been more important. That mission. We invent the future of flight, we lift people up, and we bring them home safely. For our defense business, that means the war fighter.
A workhorse of our military fleet is the J85 engine. That engine powers the T-38 aircraft that the U.S. Air Force uses to train their next generation of pilots. This is an incredibly complex piece of hardware. It has over 6,000 individual piece parts that are managed across multiple different acquisition agencies. Keeping these engines available is critical to keeping planes in the air and pilots trained. This is a problem at the heart of U.S. national security and readiness, and at GE Aerospace, we feel that accountability deeply. A few years ago, we decided to take a radical new approach to sustaining these engines, to combine the strength of our hardware with the power of software.
Let's take you back in time, not just to 2024 when our journey with Palantir began, but all the way back to the 1960s. The integration of hardware and software has been a slow burn for over 60 years. In 1965, the Gemini spacecraft had the first ever onboard computer in an aerospace application for their lunar mission. That advanced into the 1970s, where we had fly-by-wire capability, computer-mediated flight controls. In the 1980s, that advanced to flight management systems. In the 1990s and 2000s, we saw sensor integration and predictive maintenance capabilities. With AI, the pace of software colliding with hardware is accelerating faster than we have ever seen it before.
GE Aerospace is uniquely positioned to capitalize on that convergence. As I shared, we have a legacy of over a century of hardware dominance and the software maturity of a decade of production AI powering the most advanced fleet management capability in the world on our commercial fleet. When solving the J85 problem, we knew we had the technical ability, but could we do it fast enough? Could we finally break through the structural barriers that have defined the defense industry for far too long? We needed a new approach, and so we partnered with Palantir to rebuild the operating system.
The J85 is where we began. We faced a 3x ramp in sustainment demand, getting more planes flying than we ever had before. A single part of those 6,000 piece parts that make up that engine, a single bolt can prevent a repair from getting a plane or getting an engine back on a plane and getting more pilots trained. To break these constraints, we started to build an Ontology. We pulled in structured data, unstructured data, bill of materials, schematics, supplier data, technician notes, all coalesced into a single object model. This was intended to surface those problem parts, which parts are gonna prevent a repair and getting more pilots trained.
It wasn't just the parts that were going to cause a problem today. We were identifying parts that could cause a problem three, six, 12 months down the road. The thing that I love about the solution that we built is we weren't just providing visibility. This wasn't just a dashboard pointing people in the direction of the parts that were going to be constrained. Rather, it was recommending actions with the power of AI. Here is a problem, and here's how to solve it. This quickly turned into a pilot with Palantir, the Defense Logistics Agency, and the U.S. Air Force, which is now under contract focused on customer outcomes around fleet management and supply optimization, all built on that unified Ontology.
For the first time, it wasn't taking months of back and forth painful communication to share information with our government customers. Operators finally had real visibility into exactly what was bottlenecking Air Force readiness and what actions they needed to take to improve that readiness posture. This put capabilities into the hands of our customers, but we quickly realized the value of this disruptive approach to breaking through supply chain constraints. It is no secret that the industrial supply chain has been incredibly constrained over the last several years. We thought if we could have this much impact with our customers, why would we not also apply this to our internal supply chain constraints?
We called Palantir, and in true Palantir fashion, they sent six FDEs the next week to Cincinnati to surge against this problem. We connected disparate systems, ERPs, planning systems, PLMs, supplier systems to extend our Ontology and encode the mechanics of our supply chain. What used to take multiple people, this is embarrassing to say, but I'm sure that we all relate to this, every Monday, eight hours, their entire day was spent stitching together data into Excel spreadsheets to identify a handful of parts that then they had to go problem solve against.
Those same leaders are now fed constraints directly in near real time, and their time is actually spent doing the things that they're good at, which is solving those problems, not just identifying them. The impact was staggering. The combination of technology and the rigorous application of FLIGHT DECK, which is GE Aerospace's proprietary lean operating model, we surfaced constraints, we solved those constraints, and in 2025, we output 26% more engines to our commercial and military customers than we did the year before. The same architecture that served the warfighter was also serving the rest of our business and all of GE Aerospace's customers.
In 2024, we laid the foundation with J85. We structured the complexity. We surfaced the right signals. We put recommendations in front of operators, empowering the best in our people with technology. In 2025, that focus evolved inward to expanding the Ontology, and with it, our impact on our customers. The expertise of our employees is an unrivaled asset, and it is wasted on rote, repetitive tasks. We empowered our people to identify and solve the right problems, putting their expertise to maximal use. In 2026, we are building a rich and powerful automation architecture with orchestration agents that are continuously monitoring and synthesizing signals, routing to functional expert agents across fulfillment, MRO, customer service, all compounding into multiple workflows across the value chain.
AI isn't replacing our expertise, it's amplifying it. It's freeing our people from the manual work of data wrangling, unlocking new degrees of freedom for our business, and enabling compounding outcomes for our customers. J85 was the start, but our Ontology scales across every program, every fleet, every platform. We are now scaling the product we built for J85 to new fleets across the world. At GE Aerospace, we relentlessly pursue cutting-edge technology to advance our mission, inventing the future of flight, lifting people up, and bringing them home safely. GE Aerospace is a hardware company, but we are also a software company, and now we are an AI company, not just in ambition, but in practice today, powered by GE, accelerated by Palantir. Thank you.
Please welcome Executive Vice President and Chief Digital and Information Officer of The Joint Commission, William Walters.
Good morning.
Thank you. I know you just had your coffee break, so show of hands, who's heard Joint Commission? I'll get some more engagement. Oh, fantastic. Wow. I expected only Dr. Weber's hand on this side of the room from Tampa General, so awesome. For those of you who don't know, we accredit the vast majority of American health systems. When I say accredit, really our mission is to ensure the quality and safety of healthcare in those organizations for all. Unironically, I sat in a room much like this one in Palo Alto 12 months ago, my first AIP, having been at The Joint Commission as the CDIO for two months, and having been staring down the barrel of 400 homegrown applications.
Many of which, like Jess said, Monday mornings, Friday afternoons, digging through spreadsheets to get data out of those, and realizing, you know, the burden that I had in integrating those systems to the benefit of thousands of healthcare organizations and to their quality. It clicked for me at AIP, and maybe it'll click for you here today, those of you who are new to Palantir Foundry and the Ontology itself. A part of our mission, and really the primary part of it, is we deploy surveyors to healthcare organizations, hundreds a week. These are doctors, clinicians, specialists. These are folks who visit these thousands of healthcare organizations every two to three years.
There's a lot of variability in who they are, their specialty, their licensure. We do life safety things. We have people who literally crawl into, you know, HVAC systems looking at things, and simultaneously, you know, neonatologists who are asking hard questions about how you deliver, you know, neonatology care. So that's who we are. Also we, like all of healthcare, and I feel like a lot of businesses, we're nonprofit, and so we operate at about a 2% margin. Not a ton of cash to be throwing around, you know, to solve some of these problems, right? I think many of us have this dilemma.
What I realized sitting in that room, and then a month later I kind of briefed our executive team on the path I wanted to take with Palantir, is modernization was not an option. It was mission critical. We had to modernize a system. We were gonna continue to erode away at who we were, our expenses, at the disservice and really of our healthcare organizations of which we accredit. What you're gonna see today is why and what we've built with Palantir. Truly the vision we had then was to build this digital control plane, connecting every part of the organization. You'll see it kind of flanked to my left and right here.
Full disclaimer, healthcare data is sensitive, as are our accreditation cycles and windows. It's a surprise inspection, if you will, when we show up at our healthcare organizations. This is not real data, is really the net of that. When my boss watches this back, he'll know that we didn't put data back there, or any of you that raised your hand that might have a survey coming up, also not your data, right? Anyway, exciting, right? We talked about Ontology. We've talked about a lot of those things today. No different for us.
What you see on the screen is exactly that. It is our control plane, our operating system. We call it Reforge. We took a page out of The Lord of the Rings theme you see today, where the sword was reforged into, you know, the weapon that ultimately slays Sauron for the nerds in the room. But that's no different for us. We're reforging that application portfolio to solve these problems, right? This is my favorite part, 'cause this is all real. These are things you'd see if you could zoom in, are truly things that were in hundreds, I mean, thousands of tables.
Things that had business logic, things that were stuck in people's heads. These point solutions that left how we schedule our surveyors, what their availability would look like to be available for a survey, as well as our standards. Literally, the book on what healthcare standards are is one we've written and interpreted based off CMS guidance into operational workflows. Similarly, 4.5 for deployed engineers, Joanna, Arthi, Pradeep, Avid, and the half came in a summer. Hailey, if you're watching, job offer still stands. She's at Vanderbilt. In days, she took something that we were trying to build in years and built it for us.
In essence, taking the first floor of our building as a library in Chicago and contextualizing them in the Ontology in Foundry, and then we made them publicly available. Go on our website now, you'll see them there. Again, Hailey, job offer still stands. Anyway, we needed all that visibility. As it's been discussed in depth today, that Ontology does exactly that for us, institutional memory, shared enterprise language right across our organization, and truly the control plane for our digital products. All that with my core principles of being brilliant at the basics, which includes security, building a modern platform and allowing us to have advanced data and analytics capability.
Alrighty. Here's our standards. This is quite literally what you'd see on the website if you go now, but this is it built in Foundry. This is real, every bit of this. There are no secrets in what safe and effective high-quality healthcare should be, right? Like, this is an open book test. If you're delivering healthcare, please go and read it. These are directly enabled into our workflows now. Truly, for the first time in a while, those who we accredit or otherwise, those who, you know, wanna know what we do, have public clarity on requirements and in real time. When they change, guess what? They change here too.
So really exciting for us. Again, built in days. And I joke with a couple of our partners, where we used to look at our calendars to kind of make change, we now look at our watch, right, for some of these. And so while our scheduling application did take eight months to build because of all that complexity, we have literally surveyors who are in an RV and travel the world, and then we gotta find out where they are in the RV, put them at the closest airport, and then fly them to their healthcare org. That new Ontology allows us for that complexity and availability.
Awesome. Here's availability. I share this because imagine hundreds of surveyors distributed across the world with all the variables in which I discussed, you know, their credentials, what specialty they are, what licenses they have, where they're allowed, which states they're allowed to practice in. It's now all connected, all live. This availability gives us a new strategic lever we never had. I'm not kidding. It used to take us weeks to sit in a room, our scheduling team, things in their heads, things in spreadsheets. It would take weeks to build out our planned survey schedule. It now takes minutes, three minutes. Literally, you run it. We call it Solver. You click the button, it runs, voila. 600 surveyors get scheduled in seconds. Gives us ultimate transparency.
We know where they are, which is also important, and ownership. It's really exciting. The other thing this allowed us, and we've actually used it real time, many of us traveled here. There's weather today. There was weather a couple weeks ago. Did anybody experience winter storm, whatever it was called? I think it had a name. We simultaneously were able to cancel surveys, right? Similarly, there's a cyber event in healthcare right now. You know, should we make this decision, we could go on a map, draw a circle, surveys get canceled immediately. In the past, you had to call those individuals individually. You had to call the healthcare orgs. Now it's done real time. Really excited about that.
All right, what's next? We've got some of the foundations in place. I've taken those 400 applications down to about 350, so I've got 350 more to do. If you kind of watch the slides build to my left and to my right, what we really see ourselves as being is the healthcare data router of the country, right? Why would we come to your hospital, your healthcare organization, spend three to five days? If you've lived through one, it's entirely disruptive, right? You're, you know, you're sending people in partnership with us and the teams. Why not do this real-time? You saw it build.
We can get that data and give you an accreditation score before we get there. How about you know real-time where you stand in accreditation? We're still gonna come crawl through certain things that we have to look at the physical plant, HVACs, surgical sterilization, et cetera. What will go from three to five days, an immediate burden, an immense burden on the healthcare organization, is gonna be drastically reduced while simultaneously giving you real-time insight into where you stand from a healthcare accreditation perspective. Excited for the partnership with Palantir. Again, those five forward deployed engineers I mentioned were essential to this.
They don't get enough credit, hence me calling them out, and I know they're super giddy right now in the New York City office in having heard their names. Really appreciate the partnership. Thank you all for helping me make healthcare in America great. Thank you.
Please welcome Senior Vice President, Field Operations at Centrus Energy and President at American Centrifuge Operating, LLC, Patrick Brown.
Thank you. All the hype around AI, it's great to be in a room full of leaders actually deploying real use cases, delivering real value. Got a lot of work to do. We're in a race to lead in artificial intelligence, right? In every race, there's gonna be winners, and there's gonna be losers. I'm here to talk about what it takes to win. That's energy. Not just any energy. It has to be 24/7, reliable, safe, scalable, and secure. Sometimes clean, right? If you look at available energy resources, not a lot can meet those requirements. Nuclear can. It excels in all of those categories, right?
If you want to lead, if you wanna restore American energy dominance, and you wanna lead and win the AI race, U.S. must lead in nuclear power. That asks a question. Can the U.S. truly lead in AI and nuclear power if we don't lead in nuclear fuel? 'Cause today we do not. Almost all global enrichment capacity sits outside of the United States. How did that happen? We invented uranium enrichment, Manhattan Project. We won the war. We scaled the technology. We powered the Cold War. For decades, we led the world. But then, like many industries that you may work in, supply chains globalized, capacity migrated.
For us, in 2013, we shut down our large commercial enrichment plant. With that shutdown, we handed the leadership to Russia. That loss of fuel security matters. It matters for the country because fuel security is energy security, and energy security is national security, right? If we wanna power the AI economy, if we wanna repower our industrial base, if we wanna refuel the fleets of aircraft carriers and submarines sailing around the world, we must rebuild American capacity to enrich uranium at scale. That's the mission of Centrus, to restore this lost capability for both commercial nuclear power and national security. We're in the early innings of our mobilization.
You can see here, this is in Ohio. This is our 16 centrifuges producing HALEU fuel. That's the fuel for advanced reactors. You hear a lot about advanced reactors. This is the only HALEU enrichment in the Western world. Just 16 machines. Our facility's massive. It's the size of the Pentagon. It wasn't built for 16 machines. It was built for thousands. You can see here, this is only 1/4 of the building, but we have two buildings. It fills 11,520. Last October, we announced our expansion plans to fill these buildings to capacity. We began ramping our operations, but as we got started, we faced a hard truth. Our paper-based processes were too slow.
Our data couldn't keep up the speed of the mission. We needed to think different. We needed a partner. I'm not sure if you saw the press release, but at 6:30 A.M., we announced our partnership with Palantir. We're happy to take that forward to make this happen. That gets us to where we are today. This is what you see here is the centrifuge mission control. We are digitally retooling our entire operations here. This picture here is a live operational picture across every centrifuge we're building. When we used to try to get a progress report on the status of this, it would take. Sometimes the data would be eight weeks late, right?
We began working with Palantir. Now we have a real-time view. This view here is showing real-time status of what's the build status, what's the supplier risk, what's the crew constraints. Now we don't have that eight-week lag. We know instantly. If something changes in Oak Ridge, where we're building the centrifuges, we know. If something changes in Piketon, Ohio, where we're operating them, we don't have to wait eight weeks. To get going, the first thing we built was a project controls command center. This is a real-time operational picture of the program. This gave us situational awareness for the first time.
This will act as a foundation for growth because we used to just spend maybe thousands of dollars a day. It's gonna be millions of dollars a day, every day, for years. We needed to get some sort of awareness of what's happening on the program, and this use case does that. We were getting it started. It wasn't smooth. We had, I think our four deployed engineers told us it's the worst data they've ever seen. I think they say that to everybody, though. I'm not sure. I didn't question it 'cause I believed it. 'Cause that was the same thing I was saying, right? Yes, I mean, a common thing that you probably have seen, the data was fragmented across disconnected systems.
You have schedules in one system, cost in another, workforce data somewhere else. We had no single source of truth. It was really just silos and armies of people stitching it together so we could make sense of it and make a decision. With Palantir, we now have that unified operational model, the shared Ontology where every site, every centrifuge, every task is now connected, modeled, continuously updated. Of course, visibility alone isn't enough. Usually, when you're seeing it's usually too late. We're trying to build on the capability, look around corners, right? Take action before it's impacting us, right? Here, agents detected a scheduled delay from a material casing defect, automatically generates a root cause analysis, identifies the cost and schedule impacts.
Because that Ontology is connected across the entire program, employees, suppliers, inventory, variances, every recommendation's made with full operational context. It's not making decision on one piece of data or silo of data. It knows what's happening across the business. For us in the nuclear field, you have to audit everything. You know, the NRC isn't very comfortable with agentic autonomous control. Having every action logged and traceable is critical for us, whether it's human or AI decisions. Here, agent model is modeling that staffing reallocation option, trying to get around that schedule impact. It analyzes the situation, presents a solution.
It hands it off to a human. Human then reviews it, adjust it as needed, and approves. Because of this, what used to take multiple phone calls, maybe emails, manual analyses can happen within minutes. The system then, and once it's approved, the system executes, writes back to our systems, staffing schedules are updated, payroll's updated, site supervisors are notified automatically. This is an example of AI human teaming, where these agents aren't replacing our people. They're extending their capabilities, unlocking new capacities, so then the experts can then put their attention on, you know, our most challenging problems. Here's another example of a quality failure of a bearing shipment.
Without immediate action, this could cause weeks of delays, maybe a month delay. In this case, agent is looking at cost versus schedule trade-offs, analyzes alternatives, finds a solution, and it presents a recommendation. It recommends that we switch suppliers, fast-track a replacement order. Human then reviews, accepts. This cuts shipment time from three weeks to one week, reducing the delay, keeping us on the critical path. What we saw here was a defect detected, a root cause analyzed, staffing reallocated, supplier rerouted, all automated, resolved, logged, and auditable. This isn't a static system. Every action taken here compounds. The system learns and improves, enables faster, more accurate decisions tomorrow.
This is how we're gonna scale from those 16 machines you saw in the beginning to 11,520 in our first phase. Not with spreadsheets, not with paper, but with a true industrial operating system. Where are we going? Project controls was the starting place. Obviously, the real transformation occurs when we extend this across the entire value chain. Every centrifuge becomes a living object when it comes in as a supply chain, engineering, manufacturing, quality, and when it gets put into operation in Ohio. Everything's connected, live, learning, and improving with each decision. This will empower entire organizations to all work on the mission together, whether they're in Oak Ridge manufacturing, in Ohio operating, or even at headquarters.
We sometimes like to talk to those guys too. Together with Palantir, we are building the operating system for American uranium enrichment. This will allow us to restore a lost capability to our country, and it will carry our nation into the next era of nuclear power. Thank you.
Please welcome Chief Digital and Artificial Intelligence Officer from the U.S. Department of Defense, Cameron Stanley.
Good morning. I wanna take you back to the year 2016, because that's really where the story of Project Maven began, even before the project stood up formally as the Algorithmic Warfare Cross-Functional Team, because it began with a hypothesis. What does the third offset look like? If anybody knows about the offset strategies of the late 20th century, the first offset, or basically it was the function of how do we overcome differences in mass and scale that our adversaries have on top of us, in order to wield a more effective military enterprise. What we saw from the first offset, which was how do you use nuclear weapons to overcome that?
That was the first technology offset. Second one was stealth and precision-guided munitions. Secretary Carter, when he was trying to come up with a third offset, said, "The real advantage that we're gonna have in the 21st century is decision advantage. How do you get better decisions faster than your adversary? That's what wins wars." The problem was in 2016, there were a lot of problems that we could look at for different types of technology, and AI was the one that he wanted to focus on particularly for Project Maven. The Algorithmic Warfare Cross-Functional Team was established to basically say, "Get AI in the hands of the war fighter, focusing on UAV PED."
What do I mean by that? What's happening for our airborne surveillance assets so that humans don't have to stare at screens, they get tired, they get distracted, they miss things. What we ran into was, let's use AI to try and tip and cue the right systems so that humans didn't just have to stare at a screen 16 hours a day and get tired. With that, this is where we started, and the hypothesis was if we got AI in the hands of the war fighter, it would work. We developed some of the best computer vision models possible.
They were built on our data, deployed on our systems, everywhere in the world that we could possibly, you know, integrate it, we were integrated. You could detect cars and people. That was our primary challenges back in the Global War on Terror days. We actually fielded these systems quite robustly across the entire landscape. The problem wasn't ISR PED per se. That was one problem. The bigger problem was the fact that our processes were not set up and our technology was not set up to use data-centric techniques in order to make better decisions faster. That just solved the problem of one individual in the entire chain. The real problem we were trying to solve was this.
This is an image from a theater operations center in Scarlet Dragon. Scarlet Dragon is an exercise that Project Maven ran with the XVIII Airborne Corps. What you can see there is a bunch of static pictures, whiteboards. Don't be distracted by the screen. That doesn't have any automated things on it. That's just PowerPoint slides. We literally were trying to push AI detections into workflows that humans were limiting on our outputs and our outcomes that we were trying to achieve. We rejected the hypothesis that getting the AI in the hands of the war fighter was the right answer.
What we really needed to do was take three steps back and say, "The real issue isn't AI, the real issue is workflow. How are we making decisions?" That's when we came up with what I call the decision-centric approach. The decision-centric approach, I came up with nine questions. There are other approaches out there. This is my easy one. Mainly because DARPA has their nine questions, I felt like I should have my nine. So when you're thinking about trying to improve decision-making processes, you always have to start with the decision. You also have to look at how the decision's made today. What part of the process are you accelerating? What data's required to make the decision?
How is the data going to arrive? How will the user interact with the data? What's the reduction in human input? How are you measuring success, and what's your iteration plan? By looking at these, it's pretty straightforward. Those of us in the data world, we do this intuitively every single day. The challenge is getting senior leaders, especially those who are very competent, very professional, to understand that your job as a data professional is not to try and replace them. Your job is to make them better, and you make them better by giving them the data that they need and the time and the format and in the capacity that they require in order to make decisions.
There's another problem. That's one decision, and in complex workflows, as we know, there are dozens of decisions. The challenge isn't coming up with the right approach. We've got that. It's now how do we get the user community to buy into that approach in order to have them completely digitally transform their entire workflows, so that they can see how data-centric techniques, allow them to solve all of those decisions simultaneously and get to real advantage, in my case, on the battlefield. What we did was—everybody in technology knows the left flywheel quite well. Technology development, that's pretty straightforward, right? Standard spiral development process.
No one should be surprised by that. What we fail to recognize usually in the technology development space, at least in the department, is that there's a process improvement flywheel that's happening at the same time, ideally. The question isn't how are you improving the process with the technology, it's a question of how do you interlink and couple those two flywheels together so that you're delivering technology at the right phase in their process, and they're giving you the operational feedback in the right phase of your process, so that you can have mutual synergistic improvements in how we approach things.
It's not just the technology, it's also the process. The process is clear. That's what we need to be focusing on. What we found when we looked at the AI into different types of workflows is that the system was wrong. This is Maven Smart System, Palantir's software-as-a-service product that we are deploying across the entire department. As you can see, it's not just one data feed, it's multiple. Instead of having eight or nine systems for those decision makers to look at every single day in order for them to make decisions, you then fuse it into a single visualization tool.
The single visualization tool allows you to select, deselect different types of data, look at different approaches to data, but more importantly, action from the same system that you're trying to develop your workflows around. Once you have a detection that you wanna actually move into a targeting workflow, for example, this is what we do. Left click, right click, left click, magically it becomes a detection. That detection then gets moved into a workflow. This is standard digitized workflow, but I wanna walk you through it quickly. You have different types of targets that are identified on the left there. Every single column produces a different type of decision-making process.
Once you have that decision and you're trying to actually action that process, we now move into COA generation, course of action generation, where we are automatically, via a number of factors, trying to identify what the best asset to prosecute a target looks like. Once we've got the different approaches and we select one, we then can move directly into how do we action that target? We've gone from identifying the target to now coming up with a course of action to now actioning that target all from one system. This is revolutionary.
We were having this done in about eight or nine systems where humans were literally moving detections left and right in order to get to our desired end state, in this case, actually closing a kill chain. When we started this, it literally took hours to do what you just saw there, and through a number of different deployments, we've been able to reduce that time significantly all because of two things. One, our ability to actually integrate with our customer base directly and have them work with us in order to understand their process and us developing the right technology.
More importantly, connect those disparate systems in a way that's never been done before using an abstraction layer called Maven Smart System that connects and interconnects all of those things with the right data approach, the right data Ontology, and the right data formatting to connect these systems. This is not something that happened overnight. This took seven years to get here, not only from a data connections perspective, but also to connect each of those systems together. Where are we going from here? Obviously, there's many things that we have to accomplish. We're not done yet, and this is the thing that gets me up in the morning as the CDAO of the department.
Every single tool that I get from my vendors, Palantir especially, updates with time. It's the first time in my career that I have a system that gets better day after day after day , because we are integrated with our customers, we are listening to their feedback, and we're giving them the tools that they ask for, not what some requirements manager asked them to build five years ago. It's literally happening today.
I do this every single day because I care about one thing and one thing only, that 18, 19, or 20-year-old kid who had no choice in where he went or what threat he's facing because I want him to win and come home. That's why we do it. Palantir is very helpful in delivering this. Maven Smart System is an incredible system. Yeah, I live my life no fair fights if I can avoid it. Let's not have fair fights. Our guys win, and we come home. Thank you.
Please welcome Global Head of Commercial at Palantir, Ted Mabrey.
As Dr. Karp mentioned in his opening remarks, we hold these conferences so that you all can learn from each other and inspire each other, and we tend to have a policy of allowing no Palantir person on stage. For any of you who have tried to get an extra ticket for one of your colleagues and tried to wrestle with Sasha around this, Sasha enforces that quite militantly. I thought it was extremely important, it was worth having that fight to get on stage today because I'm observing a dramatic step change in what you all are doing with our products, and I think that has really material knock-on effects for how you organize around deploying this technology such that you win.
I wanted to take the time to take a step back and share what we're learning together with you all that I think is going to position us to win together this year and into the future. I'm gonna hit three concrete things very quickly. First, what I think has changed. Two, why I think that is so important when we think about how we organize together. Three, how we're evolving our engagement model to enable you to use this technology to win in the maximalist way. The first thing that has changed that I think almost everything else I'm gonna hit sits downstream of is the pace of deployment is accelerating really dramatically.
I've done nothing but deploy Palantir for 16 years. Literally all I've done. As I observe what our deployment teams are doing with you all, it looks honestly like a little bit like an alien process to me. Over the last really three months, three to six months, the pace, the scale, the breadth of what you all are doing with our technology looks like something that I do not recognize, and it made me start to wonder, how do we need to adapt and make sure that given this change, you're able to evolve and use this technology in anger? As examples of that, we had a team that did a factory tour last week at a battery pack manufacturer.
We love to do the actual concrete, see how the thing actually works, and they recorded the factory tour, understanding how the factory actually worked, having discursive discussion about what works well, what breaks, what happens when the product leaves the factory, what are the types of constraints that are imposed on the factory that make their job difficult. We'll then take that recording, feed it into AIP, integrate it with the technology and the trade craft that is being coded in AIP, and then the team over a 24-hour window was able to develop four applications.
A customer-facing application, an operator-facing application to manage the line, a technician assistant to be able to increase uptime, and a corporate quality workflow across plants that are addressing upstream issues around design, engineering, procurement quality, and be able to deliver those four applications in a 24-hour window. Second, we had another deployment that was on-site with a global logistics provider, and while on-site, cartel violence broke out in Mexico. This completely disrupted their operation. They had to reroute a lot of different things. Through the course of the afternoon, the team built a global disruption manager to track what was happening in Mexico, but also extend it for all of the different geographies that they operate in.
And then critically take that and link it into as they observed, what do you actually do when one of these disruptions happen? How do I manage product rerouting, material holds, changing to the coding process in terms of influencing what's actually hitting the logistics network? Able to do that in an afternoon to now have something that is available for the next disruption, like a disruption in the Strait of Hormuz. At a brand new customer, we're just starting with a customer that manages a very large fleet. The initial challenge statement was, "Can you build a better predictive model to manage fleet availability?" In the first 24 hours, the team was able to build four models that outperformed the existing state-of-the-art.
In the next 24 hours, they built agents that were able to consistently monitor all of the inputs that were describing the actual state of the network and challenge assumptions that the model were built on in order to fix those forecasts live. The next day, they integrated that with inventory data and sales forecasting in an optimization engine. By the fourth day, had an automated routing engine that was suggesting ways to reallocate tens of thousands of assets for maximum throughput and quality. That happened in 4 days. Lest you think that this type of pacing is only happening in clean, pristine, new environments.
You already heard from SAP and Accenture today about how this technology is allowing you to accelerate the rate at which you can get to the technology you want to run by migrating away from the old technology that you have, and we're seeing that across legacy systems everywhere. There's an automotive financing company that was intending to migrate their finance backbone, projected to be a two-year plus project, was able to execute that in two weeks. A major global retailer that was trying to onboard stores, 2,000 stores onto a new system, expecting that to be a two-year problem, was able to do that to 99.4% accuracy for 200 stores in 2 weeks.
Only limited by the data availability and the presumption that you could do all this quickly. All of that pacing is really changing the perception of what you should be demanding from software. If you want something, you want it to be valuable, it should be ready immediately. These applications and the things that are happening quickly are not vibe-coded peripheral applications that are sort of toys on the periphery of your business. The speed of scaling is also dramatically accelerating. As another example, we have a customer that is a software provider that has a legacy software stack built on old proprietary technology, but serves a mission-critical function for its customers.
They're undergoing a replatforming effort that they expected to take from a planning perspective, five years, with of course attendant risk in actually executing that plan. By going and working together, taking a code base, there was millions of lines of code, requirements, documents across both regulatory requirements and the user journeys that need to be supported in hundreds of thousands of documents, translating that into a granular technical Ontology, every code file, every potential column that was going to be added to the database. They were able to build an Ontology to one, not only track and understand the interdependencies across that project, but also have a foundation to run long-running agents.
To start to suggest and build new products so that they could not only replatform but transform in place. That project that was expected to take five plus years is now expected to be done in less than a year, and of course, our Impl team is accountable to continuing to accelerate that every single day. With a medical device customer, they started with, "Hey, can we use AI in order to proactively monitor our adherence to the regulatory and compliance requirements that we have in a heavily required industry for our manufacturing and process controls?"
Through the course of building that application very quickly, it became clear that there was another issue, which is just understanding the full opportunity base of existing products and thousands of different SKUs that have been accumulated over time through acquisition in different MES systems, PLMs, ERPs. How do I actually understand the current base of what I have in the feature set in order to serve my aggregate customers? Another R&D team that is constantly surveilling the market landscape. Reading literature about innovations, studying what competitors are releasing, understanding evolutions and standard of care that create market opportunities and focus points for their R&D.
How can I then collapse all of those things into something where I have the equivalent of CI/CD for new product development instead of iterative handoffs of a new idea that doesn't meet regulatory controls, months and months of handoffs, and dramatically moving that process to something where as they can innovate, they can hit all of the compliance checks that are required to push new products out? For a drone manufacturer, the first problem initially in the first day was, "Can we just get a better handle on what is happening with our in-service fleet?" The availability, where they're having mechanical issues, where they have service issues.
It became very clear that it's valuable to get a handle on that within the first 24 hours, but it really wasn't particularly valuable if you couldn't integrate that back with the existing operation of what you do about it. How can I automate purchase orders? How can I replan production? How can I make sure that my supply chain is coordinated with that? How do I automate customer interactions so that I can actually get a technician out there oftentimes in austere environments? If you're not solving all of those problems together at pace, you're not actually doing something, you're just pushing the squishy balloon and moving efficiency around.
The last thing that is really dramatically changing is by doing the work, we're discovering the work that is actually valuable to do. Instead of the targeted insertion of incremental efficiencies, we have a Tier 1 automotive supplier that does value engineering to reimagine, reengineer every single one of their components and every single one of their parts. Think about, can I use a cheaper material? Can I move from shrink wrap around a wire to a plastic covering? How do I reengineer each one of those components? It's a very lucrative exercise for them to do that, but it's also very time-consuming and costly. It's cross-functional. There's a lot of controls associated with it.
By doing it once, you then are able to create a foundation where then you can create an agent that is essentially operating as an advocate for every single component in every single product on every single customer every single day, trolling and understanding every change in the dependent variables, supplier costs, new regulations, new requirements from the end customer to do that value engineering at scale across the entirety of the portfolio. Similarly, and perhaps in a way that only a Palantirian could see, that same dynamic is playing out with a fashion retailer that it needs to understand, do I have the right product with the right material in the right place at the right time?
When you realize that's not just a product management perspective because the market is not the market, each store is different, each sizing is different, and can I have that coordinated, essentially agentic advocate for each address in the same way that I do for each, or each wiring harness? We're seeing this with a medical provider that is able to automate and understand their revenue cycle management, where the automation of that is valuable but incremental, but the real value is giving them more confidence to understand what new clients they can take on to increase their practice and increase the utilization of their actual providers.
Seeing this within a manufacturer in the deep in the industrial base that generates or builds hatch covers for submarines, whereby automating the quoting workflow and getting engineers back on the floor, they can create more certainty in terms of what they can provide, which means the government could order more parts, which means they were able to move from two shifts to three shifts and do more of what they do in order to win, not just do the same thing more efficiently. Taking a step back, why that all matters, why I think it matters for you, and the thing that I think it is changing in sort of a paradigm shift with how you think about deploying and integrating with this technology.
What our best customers are doing is they are holding this technology accountable to enabling them to win. In Palantir terms, to dominate. I think what I mean by that perhaps is best expressed by saying what it's not. What they are doing is essentially rejecting what I would pejoratively label the private equity buyout mentality of how to deploy and use technology. I wanna do the same thing a little bit more efficiently with a SaaS software provider from the software industrial complex that provides the lowest common denominator solution of what I do elsewhere, so I can do the exact same thing with fewer heads.
Then have maybe a short-term margin bump until the rest of the market adapts. This is not interesting, and it is not the highest and best use of this technology. It needs. Because of the pace at which it can operate, technology is in the fight for these customers, whether that is literally in the fight supporting something like Operation Epic Fury, or as you heard from Admiral O'Connor, what it is accountable to with ShipOS is not some technical requirement, it's to ships at sea and subs in the water. I love that line, and thinking about what ships at sea and subs in the water for your business in terms of what you're trying to accomplish.
One of the reasons that I think the technology and that pace allows it to be operating in the fight is for the first time, it actually makes economic sense for your most valuable substantive operators to be engaged in defining what the software is that runs your business. Because the software is now immediate, adaptable, and to the extent it is embodied as our Silicon Valley brethren would have you believe, it is yearning to unlock itself to do what you want it to do. When it can be immediate and you can have that turnaround, it becomes the highest and best use of your best people. Everybody knows these people, whether they're in the CEO or they're at the edge.
That person who the place doesn't run without. But they just wanna do the work. Maybe they don't wanna be the manager. It certainly never made sense for them to engage with systems that were designed to raise a tide or drive standardization when they are the thing that is not standard. They have the intuition. They know how the plant runs. They know how to underwrite the policy, whatever it might be. You can now build Iron Man suits around those people, so that they can do 1000x what they could do before and set the standard for what is unique about your organization that you wanna do.
Or that can come from the CEO, the most motivating customer. "This is what I want my company. This is what winning in my industry is gonna look like in 2030. I understand it." Then it's my job in order to manifest the initial MVP of an implementation of that company by Easter. Then the final thing that we're seeing our customers do is that they're using the technology to reintegrate their companies. Instead of having the sequential, conceptual or literal manufacturing process of managing complexity by isolating teams that are served by specific isolated systems with specific boundaries, they're organizing around the actual problem that they wanna solve.
Where you now get actual, quick, and empirical returns from having those people cross-functionally in the same room operating together like the X-Men. Finally, I'm gonna wrap up with how we're evolving our engagement model and how we would like to engage with you. I like to think of this almost as like the social contract that we wanna have. The first thing that we need from you is we need you to tell us what your business is gonna look like if it wins. And that can come from the CEO, of course. It can also come from the edge.
It can come from that substantive operator where we can build together and then build momentum to give them the thing that they know , actually is required to make things work for your company. What you're going to see happen, is you're going to expect that the minute you start working on that, you're gonna create an initial solution, and then you're gonna realize that the actual value of the solution isn't gonna be captured , unless I can start to integrate workflows upstream and downstream from the specific place that you started.
What we do in building that initial solution for the first day, we're gonna need to pull in another team for the second day, for the third day, as you kind of expand fractally to capture that actual value chain. In the course of a week, you should see this come to life. If it comes to life, then what we wanna do is replicate that for each and every one of your critical value chains over the course of a very intensive 30-day period. In that 30-day period, we're going to be very demanding of your time, attention, and requirements in order to move at pace.
If we can do that together over that 30-day period, what we'd like to exit with is a one-year plan that is gonna take this system and integrate it across the entirety of your business, so at the end of 12 months, what you have in 2027 is financials that look like ours. This type of engagement is what I'm spending all of my time on. We understand that it requires a lot of engagement from you all. Of course, that requires trust and understanding in our relationship. We'll meet you where you are if you aren't there yet, to the extent that we have capacity.
I wanna be very clear about everything that we're earning as we build your trust is to get into an operating model that is this aggressive because this technology is begging to be deployed this aggressively, such that you all win. Thank you.
Please welcome Head of Corporate Development at Palantir, Sasha Spivak.
One more round of applause for our keynote speakers. We've spent a lot of today unveiling and sharing the big magic, the big wins, the big outcomes, the transformational visions. I recently learned about an internal chat channel that one of our commercial teams has called Magic Moments, and this is a channel that's meant to highlight the moments on site or with customers that just feel a little magical. These are moments like a Delta completely redesigns a workflow, and a user walks up to it first, for the first time and knows exactly what to do.
These are moments like a customer shares, not a professional anecdote, but a personal anecdote about how Palantir and how working with Palantir has affected their life.n These magical moments can seem little, but they're innately human, and it's our great honor to create both the big magic and the little but human magic alongside y'all. Thank you for being here. Have an incredible rest of the day. Please be honest with each other. Share what's hard, what's working, what's not working, exchange stories, and let's get to work.
For those of you that are joining us on the live stream, thank you as always for being here, and for those of you that are here with us in person, we're gonna head to lunch.