Good afternoon. My name is Mike, and I will be your conference operator today. At this time, I would like to welcome everyone to the Bloom Energy conference call to discuss Bloom's services business. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there'll be a question-and-answer session. If you would like to ask a question during this time, simply press star, followed by the number one on your telephone keypad. If you would like to withdraw your question, press the pound key. Thank you. You may begin your conference.
Thank you for joining us on this call to discuss Bloom Energy's services business. To supplement this call, we will be referring to a presentation on the Investor Relations website. The matters that we will be discussing today may include forward-looking statements regarding future events and the future financial performance of the company. These statements are subject to risks and uncertainties that we discuss in detail in our documents filed with the SEC, specifically the most recent reports on Forms 10-K and 10-Q, which identify important risk factors, including those related to the COVID-19 pandemic that could cause actual results to differ materially from those contained in the forward-looking statements. These include statements about the effects of COVID-19 on the company's business results, financial position, liquidity, demand for our energy server and new applications, timing of new applications, and the supporting market ecosystem and outlook.
We assume no obligation to revise any forward-looking statements made on today's call. During this call, we may refer to GAAP and non-GAAP financial measures. These non-GAAP financial measures are not prepared in accordance with U.S. generally accepted accounting principles and are in addition to, and not a substitute for, or superior to measures of financial performance prepared in accordance with GAAP. A reconciliation between GAAP and non-GAAP financial measures is included in our latest quarterly earnings press release. After the prepared remarks, we will take your questions. I would like to note that we are all dialed into this call remotely, so we apologize in advance for any audio issues that may occur. With that, I'll turn the call to Greg Cameron, CFO, and we'll begin.
Hey, thanks, Suzanne. And thank you all for joining us today. Listen, this is our continued series on investor events on the company. And our goal here is to continue to provide insights in not only how the business is operating and how we are running it, but as well give you a sense of the leaders that we have in the group and just expand who you get to hear from on that. So I'm really excited about today's really focused on our service business. We've got Glenn Griffiths with us today. He's been with Bloom over six years. He leads our quality, reliability, and EHS efforts. In addition to that, Glenn is our P&L leader for our service platform. So he's going to share his perspective on his business. Glenn's background: he was with Hewlett-Packard as their Chief Quality Officer.
He had several roles while he was at HP, focused on both product and service quality. And prior to that, Glenn served 22 years with the U.K. Royal Air Force. So really excited to have Glenn with us here today to share his views. He's got some slides that he's prepared around the business that he will walk us through. And after that, I'll come back on and take your questions and go through a little bit of Q&A. So with that, Glenn, I'll turn it over to you.
Thank you, Greg. From the outset of the first product we sold in 2007, we have invested in our technology, infrastructure, tools, processes, and people to build a technology solution and long-term service offering that is a core part of the differentiated, resilient power solution that Bloom brings to the energy industry. Although still in the early stages of our growth curve, we already have a global footprint of over 650 unique sites and a service backlog of $3.4 billion. This backlog represents revenue that will be recognized rapidly over the next 15 years and annual payments or cash receipts from our customers over that period. With a 100% attach rate of O&M contracts to each product set, our services business is growing at the same fast pace as our product business.
Given that our average service contract award is either of equivalent or greater value than our average associated product revenue, we expect to see our service backlog value grow commensurate with our annual product revenue. Moving forward, as we continue to rapidly scale, this will provide us a predictable, profitable, and growing annual revenue stream. We have spoken previously about our target of achieving 20% profitability in our services business. I will spend time today showing progress on our plans to achieve that target profitability level. Before I jump into the details, I'd like to spend some time talking through what we do in our services organization and how it all works from a contracting and commitments perspective. I'm showing here our service offering overview. We operate, manage, and maintain all Bloom systems worldwide. We have established a 24/7 remote monitoring and operation centers in the U.S. and India.
The remote centers start up, operate, optimize, and shut down systems as needed. They also remotely update firmware and dispatch service technicians when required to resolve on-site issues. They utilize the embedded instrumentation and automated controls in every Bloom server to monitor performance and, if necessary, adjust parameters to ensure we maintain optimum outputs. We use advanced analytics to proactively identify potential issues and make changes remotely to address them before they impact performance. I cannot overemphasize the benefits that this capability provides us. Before joining Bloom, I was chief quality officer of Hewlett-Packard, and in that role, I was never fortunate enough to have the level of instrumentation and real-time access to operating data, along with the ability to tune performance in the field that Bloom has put in place.
Using this capability, we are able to track real-time performance of every system across the world with traceability through to performance of individual parts, and we can even identify variations in different suppliers' performance for the same part. This is a powerful capability that we fully capitalize on to provide a closed-loop feedback path to engineering, quality, and operations teams to identify areas for reliability and performance improvements and rapidly cutting changes to our existing products. It also provides a comprehensive data set on long-term Power Module parameters and operation, which provides learnings that are incorporated into our next-generation products. Now we have this investment in automation and instrumentation in place. As our fleet grows, this cost fully scales to the extent that we can support a fleet 10 times or greater than our current size with a shrinking marginal additional cost the more we scale.
Another benefit of the leverage of this overhead is that as we bring more systems into service, whether it be in one country or location or in another region, not only is there a lower overhead cost on the new contract, we also gain additional margin improvement on older contracts in other locations and regions. Our level one service primarily deals with preventive maintenance items such as replenishment of filters, absorbents, as well as minor repairs such as fans and blowers. The parts cost reductions scale with volume, and we also leverage this footprint when we add additional systems in adjacent areas. For new countries and regions, we do add some incremental cost. However, as we add further systems in region, we enjoy the benefits of scale. We do see aspects of level one work as non-core and already outsource some elements of this work.
As we continue to grow, we intend to move towards an authorized service provider model and seek the right partnerships to ensure we maintain our high standards of quality and customer focus while further bringing down costs over and above the leverage we already enjoy. We consider the level two service activities as core to Bloom's business. Power modules deplete over time, and we have built our services model both around driving technology innovation and cost reductions combined with fast and efficient removal, repair, and refurbishing of our power modules. We are able to remove and replace a power module at a site within an hour without interruption of customer power.
We have also invested in a dedicated repair and overhaul factory, which is able to take a used power module and quickly integrate new, lower-cost technologies and longer-life parts combined with reworked and/or recycled parts to produce a longer-lasting and significantly lower-cost replacement than the original. I will go into the cost reductions we are achieving later. So to summarize how this all works in practice, when we sell a product, we attach the services contract to it 100% of the time. For the service contract, we provide the remote monitoring and diagnostic services and level one support that I mentioned earlier to ensure ongoing operations of the power modules for our customers. When we reach contracted power efficiency, our level two service replaces the power module and implements our latest technology.
We refurbish the removed power modules at a significantly reduced cost versus a new one and then return them to the field. It is worth noting that in 2021, over 75% of our power module replacements will be refurbished units, which have been redeployed to customers at a lower cost to us and provide for a longer life and higher power than when they were first installed and also than the power modules they are replacing. This provides us with a significant cost advantage that I will go into more detail shortly. By embedding O&M support as a core element of Bloom's value proposition, it enables us to stand firmly behind our customer commitments to provide clean, affordable, reliable, and resilient power and ensures that we remain close to our customers as we help them manage their evolving power needs and deliver value throughout the lifetime of our long-term contracts.
Our customers benefit from our continuous improvement cycle as we introduce better, more efficient technology with every refresh cycle. It also positions Bloom well for future product and associated services revenue as we gain an in-depth understanding of our customers' power needs, and we can more effectively bring our future clean technologies such as hydrogen fuel cells and electrolyzers to the broader market and introduce these technology upgrades into a large installed customer base both quickly and at low cost. Our contracts are based around the individual business needs of our customers, primarily availability, resiliency, power output, and efficiency. Contract conditions can vary greatly depending on the priorities of the customer and location of the site.
For example, capacity factor commitments can range from 95% down to 80% or even less, depending on whether the customer's priorities are based on maximizing annual and lifetime savings versus the grid or ensuring a critical load is always covered or potentially both. We also provide contracts where we have variable capacity factors depending on the customer load profiles. When it comes to efficiency, we have contracts beginning in the range of 54%-50% lower heating value, cumulative lifetime efficiency through to other contracts that have no efficiency commitment and are based solely on capacity factor or resiliency criteria. As mentioned earlier, when we contract for efficiency, we take into account site design characteristics such as transformer loads, meter locations, and even gas composition to ensure we balance our customer contractual commitment with our ability to deliver over the life of the contract.
When we price our contracts, we take all of this into account and price accordingly. We assume continued technological advancement based on our product roadmap at the time that enables improvements in power output and efficiency at lower costs while also taking into account some level of margin slippage due to unforeseen events. We view the current long-term contracts and our future ones as a predictable annuity stream with the exciting potential to continually improve margins based on our proven ability to drive innovation in power density and aggressively take costs out. I will spend some time later going into detail on the progress we have made to get us to the break-even point of our services business and how that will drive future profitability. There are two main drivers of our service performance and associated profitability.
By far, the biggest driver for performance is the duration of the power module's efficiency and power output, as well as its installed useful economic life. The longer that a power module can provide the requisite power and efficiency without the need for replacement, the lower the associated costs will be to meet the customer contract. Despite Bloom continually increasing power module life, longer-term contracts will require one or more replacements. Therefore, the second driver of profitability is the cost to replace the module with either a new or refurbished module. I will go into more details on each of these areas. However, it is the combination of significant improvements in both of these areas that is enabling Bloom to pull forward its expectations on service profitability. What we are showing here is a typical range of installed life we have experienced with each successive generation of Bloom power module.
Simplistically, a power module life is driven by two things: what efficiencies did it start at, and what is the rate of degradation? Let me address each of these in turn. We already published the maximum beginning of life efficiency that our systems have demonstrated, which is verified by independent engineers. For example, Bloom 5.0 beginning of life maximum efficiency is 65% lower heating value. However, when measured at a site level, we expect to see variation of measured efficiency based on items such as gas pressure, gas composition, local site design attributes such as transformer efficiency, and so on. Therefore, there will always be different levels of starting efficiency with exactly the same power module installed in a different location under different conditions. What I'm showing here are typical ranges of starting efficiencies for each generation of Bloom power module.
The rate of efficiency degradation can be dependent on the duty cycle, environmental factors such as water and gas quality, as well as controllable factors such as electrochemical degradation mechanisms, component quality, and manufacturing variation. When we look at power module life, we expect to see a distribution of performance over time based on these factors. Hence, the spread of degradation curves we show here for each generation. In addition, within each generation, we drive continuous improvement in materials, manufacturing, part quality, and power management controls to continue to improve the slope of degradation over time so that even within each generation, we see improving performance each and every year that we manufacture that vintage.
When we put that data together, as you can see here, each generation of Bloom power module begins life with a higher demonstrated field efficiency and a lower degradation rate over time than the previous generation. With our current shipping 2021 Bloom 5.0 modules, we are expecting to achieve over 5.5 years of installed power module life, which is a two and a half times improvement on our Bloom 1.0 power modules. Let us now look at what that means financially. In order to provide a compelling customer value proposition for our 1.0 and early 2.0 contracts, the expected low power module life led us to contract at a negative lifetime gross margin with the expectation that improving module life would lead to margin recovery in the latter stages of the contract.
As we continue to develop Bloom 2.0 and 2.5, we have been able to gradually move contract margin upwards. As we transition to Bloom 5.0 with a much longer power module life, we have been able to significantly reduce the number of power module changes required to support all contract durations. This chart provides a simplified representation of the net effect of improving power module life, showing an increase in the expected margin for each generation of contract from initial net loss to an expected gross margin of 20% for more recent contracts. We are able to further hasten this shift towards profitability with backward compatibility of our power modules, which enables us to put 5.0 power modules into a 2.5 power module enclosure to achieve longer life and higher power, thereby further reducing the number of power module changes required against the original contract.
This capability, combined with the rapid growth of our 5.0 installed fleet, means that the combined services margin is improving at a faster pace. In fact, we are now at a stage where the large majority of our total contracts are for customers who started on our 5.0 technology, meaning the lifetime expected margins for those contracts are targeted at 20%. In addition to having significantly reduced the number of power module replacements needed to support all existing and future contracts, we have also aggressively improved our service cost run rate. First of all, the impressive material cost reductions and lean engineering initiatives driven by our engineering, operations, and procurement teams have led to a reduction in both the cost of a new unit as well as a proportionate reduction in the cost of a refurbished unit that utilizes a smaller percentage of new parts.
In addition, the dedicated repair and overhaul team have reduced the cost and increased the percentage utilization of refurbished parts, which provides us even greater percentage cost reduction than those of a new unit. In totality, these combined improvements have led to a 40% cost per kilowatt reduction over the past five years alone. To put this in context, in 2021, more than 75% of our power module replacements will be refurbished power modules that are being redeployed into the fleet. The cost advantages that refurbishment provides us, plus the increase in life and power relative to the power modules they are replacing, is a critical component of our services strategy. It is important to understand the significance of this attribute and bear this in mind when evaluating the current and expected future costs to support the Bloom service contract.
Anyone overlooking this Bloom advantage could incorrectly conclude that an older technology performance replacement rate and cost will remain the same for the whole duration of the service contract. Hence, they would significantly underestimate the margin benefits we are experiencing with each improvement we introduce. We will show this data on the next slide, but to summarize, while 75% of our fleet have Bloom 5.0 power modules installed, a significant portion of which are refurbished power modules operating at greater efficiency and power and with longer expected lives than when first manufactured. A strong focus on quality and reliability has led to the second area of unit cost reduction, with a steep drop in the consumption of other parts and the associated replacement labor costs. Not only are we consuming less parts, we are seeing reduced cost per part as our product volume continues to scale.
These two factors combined have led to a 71% part and labor reduction over the past five years. Coming up are some examples that will illustrate how these improvements come together to impact our bottom line. Before I do that, I wanted to pull together a summary of the relevant improvements we have made over the past 10 years plus that have brought us to the current inflection point for service profitability. The top half of this table shows each successive power module generation, the deployment window, power output, and footprint. The next section shows our deployment mix for each generation, and it's worthy of note that 95% of our deployed fleet is generation 2.5 or higher, with an average expected life of 4.9 years or greater.
You should conclude from this slide alone that any analysis of external fleet performance data with installations earlier than 2016 will not be much use in predicting current fleet and future fleet performance and cost. The bottom half of the table shows the actual average life performance of our fleet prior to 2016 and the projected installed life and future life of our 5.0 and next generation 7.5 product. Naturally, there is a distribution with these averages. For example, our oldest installed Bloom 2.0 power module is 7.7 years old and still going strong. The final row shows an indexed unit cost, which demonstrates the extent of the cost reduction between generations. Comparing Bloom 1.0 with Bloom 5.0, you can see that we have a two and a half times life, a 2x power, and 80% plus lower cost, all of which individually is beneficial to services.
Now we'll show the dramatic impact on services performance and margin in combination. What I'm showing here is an efficiency against time plot for a Bloom installation that would represent a Bloom 1.0 installation commissioned sometime in 2009 with a typical 15-year service contract. I am picking this age of site specifically as an illustrative example of against our contracts and the associated future cost implications that are incorrectly being made based on the partial and limited data that is available on older sites. Hopefully, this slide gives folks a better understanding of how it all works in practice. What you can see here on the left-hand side of the chart is that we started this site in 2009 with our Bloom 1.0 power module delivering 25 kilowatts, with a starting efficiency typically of this vintage around the 50% plus area.
If this were a typical one megawatt site, there would be 40 Power Modules in place at this time. Bear in mind, 10 plus years ago, this was a breakthrough efficiency accomplishment, and the early adopter customer contract would have been negotiated on a lifetime efficiency based on the capabilities of the equipment at that time and taken into account the site efficiency losses mentioned earlier. What we show next is the actual degradation of those first-generation Power Modules, which resulted in us having to replace them approximately two years later and incurred a significant cost to do so. We expected to have to do this, and we constantly sold our earlier service contracts at a loss in order to grow our business. In 2011, we replaced the 25 kilowatt 1.0 Power Modules with a smaller number of our replacement 30 kilowatt 2.0 Power Modules.
And what you can see here is that the efficiency jumped. This means that as we refreshed our technology, we operated significantly above the contracted efficiency levels, and the customer economics improved dramatically with a higher efficiency being delivered. With a higher average life of approximately three years for the 2.0 systems, we didn't need to put in the next replacements until 2014. Please note that for the majority of this term, we operated above the initial startup efficiency at this site. At this time, we put in a smaller number of our 42-kilowatt 2.5 system with an improved starting efficiency of over 60%. So once again, our improvements continue to further increase the value proposition to this customer. And for Bloom, once again, we are putting in less power modules at a much lower cost per unit.
With a much longer life of our 2.5 power modules, we can now fast forward to 2019 before we have to go to the next and also last module replacement against this contract. During this period, our engineering, operations, quality, and reliability teams have developed a process where we can take Bloom 2.5 power modules and fully refurbish them with new coatings to achieve the same power output and efficiency as the 5.0 system. Therefore, I also have the choice of retrofitting a 2.5 system with a 5.0 power module, or I can upgrade the site again and install brand new Bloom 5.0 power modules. In this example, I chose to use refurbished 2.5 modules as they are my lowest cost option to support the remaining five years of this contract. For other contracts, we may make a different decision.
Once again, I can put in less units at 50 kilowatt again at a lower cost, and again, the customer gets another boost in efficiency. The customer now has a site operating at between 12%-15% higher efficiency than when we commissioned in 2009. For Bloom, from 2019 onwards, other than routine maintenance and minor parts replacement, I have little remaining service cost until the end of the contract. Now let's look at how this plays out on a financial basis. What I'm showing here is an illustrative view of the service revenue and cost over time for the same contract. The green line here shows the steady annual revenue that we receive for this contract, and the blue line shows the annual service costs in the year they occur.
What we can see is the large spike in costs we incurred with the first power module replacement for the two-year stage and the losses associated with that activity. With the second replacement, which were the 2.5 power modules, the impact of installing higher power and lower cost replacements can clearly be seen. Despite these cost improvements, you may have noted that our cumulative gross profit continued to decline. By the time we got to the third and final replacement, where we are using low-cost refurbished units, we are finally moving towards the break-even point on this contract. As I mentioned earlier, the only remaining costs expected to be incurred are for minor part replacements and maintenance. Therefore, we can expect the remaining five years of annual service revenue to be profitable on an annual basis.
This illustration is just one example of how the improvements we have implemented are driving services profitability and how even our earliest contracts, which were intentionally sold at initial loss, can become profitable over time. As we mentioned at the start of this presentation, we have over 650 sites with multiple contracts, all at different stages of their life cycle. What I'm showing here is what the profitability of a current contract looks like on a cumulative basis. As you can see, we now need only two full power module replacements instead of the earlier three. Combine this with our much lower current cost per power module and demonstrated cost capabilities. We now expect to remain margin positive throughout the whole contract duration and achieve our overall services target margin of 20%.
When we roll up all the improvements we have made on our existing contracts and overlay the contracted revenue and expected cost profiles of our backlog, we are seeing that we have finally reached the inflection point for services profitability. This, combined with the incremental investments in services Greg referenced in the last earnings call, has led us to have confidence that our services organization will now reach sustainable profitability in 2021. And furthermore, we expect to deliver year-on-year margin growth towards our 20% margin goal. To conclude, before we move on to questions, Bloom has a strong and healthy services business that is growing revenue in line with our product sales. Our company focus on technology and process innovation, coupled with an almost maniacal focus on cost down, has enabled us to transition our services business to become sustainably profitable this year.
We continue to have aggressive company-level power module life improvements and product cost down plans, both in 2021 and beyond, and these will continue to significantly benefit the services P&L. In addition, in 2020, we created a dedicated services organization to further accelerate refurbishment program cost reductions. These plans, combined with the economies of scale we achieve as we continue to rapidly grow, have us confidently on the path to achieve our 20% margin goal. Thank you. I'll now hand you back to Greg for Q&As.
Thanks, Glenn. I think we're getting ready to take some questions here. Mike, can you just give us instructions again so we build out the list of questions?
At this time, I would like to remind everyone in order to ask a question, press star, then the number one on your telephone keypad. We'll pause for just a moment to compile the Q&A roster.
Great. Hey, Glenn, while we wait for questions, let me give a couple that I get when I talk about the company to potential investors and analysts. One is, you mentioned this during your presentation, but I just want to go back to it. We've talked about profitability for the service business going forward. You've got us here at break-even this year with a profit in 2021. Does that sustain as we move forward? Will you think we'll be profitable going forward as we leave 2021? And will we grow from there?
Yeah, as I showed in the last slide of our compound portfolio, the majority of our projects now are the 5.0, which we led at a good, healthy margin. So what we're seeing this year is really, as we retired our older contracts and we got 75% of our modules of the newer stuff, I'm expecting us now on a good, healthy year-on-year growth, hopefully somewhere in the 5% area, Greg.
That's great. That's great. That's wonderful. You mentioned this in the call, and I've got a few follow-up calls because we had your slide in the earnings deck with KR and I a week or so ago with the concept of backward compatibility, and you mentioned it a couple of times. So are the machines designed? Is the server designed that it can take different power modules from different generations? How much work needs to happen at the customer site?
It's actually pretty much zero at the customer site, Greg. When I talked about the one hour for a Power Module replacement, that includes the next generation. What we do is we take something like a 2.5 module. The choices we have, we can use the latest generation of coatings on the same internal, the actual fuel cell stacks themselves. We can actually take those units, recoat them with the latest electrolytes, example, and return them into the same module and put it straight back into the same enclosure at the customer site within an hour. We also have the option to actually take completely new stacks and columns from the next generation and put those into the previous Hot box as well. Sorry, the previous Power Module as well. So therefore, we have a variety of different options we have when we drive backward compatibility.
And over and above that, Greg, what we also do is, as the team, the engineering and operations and procurement teams identify better-performing components at lower costs, those are available for us to retrofit. So our overall strategy is to try and make everything backward compatible. So it really is a very simple thing. Where it becomes more difficult for us is if we're going up in a significant amount of power, we may have some additional cost in replacing the electronics, the power electronics. But there's a crossover point for us on the economics. Where actually, if you're getting that much more power, it's worth spending the extra money on a longer-term contract to upgrade the electronics.
So that way, and again, we can do that without the customer site going down because of our redundancy, meaning we can do power module and electronics in isolation, basically a hot change.
Yeah. You've mentioned data analytics. And at the beginning, you talked about just the amount of information coming off of our server. And how do we manage all that? I mean, how much data do we get from the machine? Where does that go? How do we use that data to differentiate a Bloom server?
Yeah. And it goes in. We've got a very distributed server farm. It's naturally run by our IT department, but it's actually hosted within our engineering team as well. And interestingly enough, we have a dedicated analyst team embedded in the engineering who do the analysis for both the engineering, the quality and reliability, and the services organization. So that enables us to have a very consistent view of performance data. And we've got the complete life of every single power module that we've put out in the history of the company. So it's a significant data set. We're able to sort that data based on vintage, based on supplier. Based on actually, in some ways, the more recent data, we can actually go down to who assembled it on the factory floor.
We've got that level of transparency and visibility of both our operating data all the way through to our supply chain and manufacturing data.
And are we pulling information right off the machines as they're operating? I know as I've walked through in San Jose, you see the output and the temperature. And you can model from what we're pulling off today based on the curves as you think about managing the fleet?
Yes, absolutely. So what we do is we got telemetry with every unit. It transmits back into our secure service system and network system, and we pull that data in. We can create thermal models of the unit itself. It goes into our RMCC. So if you walk into our RMCC, which you've done, obviously the team's done, you can actually see the sensors for every single power module, how it's performing, how it's operating. There are automated alerts against that data that highlight issues for us or adjustments that need doing as well.
That's amazing. This is one question that actually popped up as we were preparing for this presentation over the last couple of days. And I don't think I've ever asked you this, but I'll ask you here: is the overlap between your role as quality leader and service leader? I think there's some great economies to that and great perspectives that you bring to that role. And I just wanted to get your perspectives on that.
Yeah, actually, it's a very interesting position for me because you could argue that the role of a quality and reliability leader is positioned to act on behalf of the company. Sorry, on behalf of the customer. And also to drive down service costs is one of the roles of the quality and reliability leader. So for me, it's good to say that I spend the first four and a half to five years in the company staring at reliability data or quality data in detail every single day. Understanding how our systems perform, understanding what the issues are, understanding how we do trade-offs and how we look at things in the company, and so on.
And so when I took on the services role, I had a really good understanding of our data capability and then also a really good understanding of our engineering capability as well as our service organization. So when I took over the services organization, I was able to work across all parts of the company and really understand what are the decisions that we need to make, how do we drive changes, etc. And let me give you an example of that, Greg, where every week we sit down with the service organization, myself and Venkat, our CTO, and we go through in rotation changes being driven in the company. And so an example of this will be last week alone. In fact, it was last Friday.
So Venkat and I sat down with my senior director of reliability, who I consider is one of the best reliability engineers in the world. So he sat down with him and his team and the engineering team, and we went through the roadmap for every single power module life improvement that is in the roadmap that engineering are working on. So in turn, we went through those, we reviewed with the engineering team the projects, we reviewed the data for the trials that are being done, we looked at the expected time for implementation. And then by working closely with that team, we're able to identify when the improvements will cut in. And we're looking at things, Greg, five years down the road for some of these changes because we put an improvement in on power module life. I'll give you an example. Look at a coating thickness.
We're going to get a thicker coating. It's going to extend the life by one and a half months. That one and a half months is only going to appear at the end of a five-year, at the end of five and a half years. So it's down the road, it's going to drive service improvements. But what it enables us to do is it's going to add increase to our product cost slightly, but the overall benefits in the next service contract far outweigh that cost. So we're actually able to save in our current service fleet with the next unit that goes out, but also reprice our value proposition to the existing contract and offer a lower price to the customer.
Me being able to move back and forth between the engineering and the services organization and look at data in conjunction with a CTO is invaluable for me in this role.
That's amazing. It really is. I learned something each time we do this. I think, Glenn, there was a couple of comments that I've received as I came in to be CFO and some of the diligence items that I looked at in the company previously and whether it was in the financials or in the documents that were available to me, and one of the things that I saw is just the questions. And I think you addressed this during when you were talking about the life extending, the cost coming down, but as I looked at early days of having contracts where you're at negative margin and people having the extrapolating that out and saying that is something that's going to continue in perpetuity.
I think you did a very nice job today around dispelling some of those questions and talking about the contracts that we're doing today are all priced at a targeted profitability to them. And maybe even some of the contracts that we might have had issues with before, and you saw that in our service P&L where we were losing money before. We've now either moved those contracts. Those contracts are either expired or they've moved to profitability. So that's one way in which it's allowing us to get to profitability, not only for the business, but takes away any liabilities that existed for individual contracts before. I thought that was worth just discussing and talking a little bit about kind of the financial model and how it works.
So as you see that the business is moving towards profitability, you see that the contracts that you have to manage as a service leader, you can move each of those contracts to profitability as well?
Yes, I do. In fact, the analytics we got behind that now, we're able to look at each and every profit and loss for every site that we've got installed. So we're looking there about how do we, for the ones that are the most unprofitable, how can we accelerate those units, those sites to profitability the fastest and so on. And then that then moves the needle on the profitability the most. And as you know, from the dialogue we include over the year, we spend a lot of time managing that in order for us to ensure that we get to that transition point as quick as we can and stay there.
Yeah. I knew before, and I just wanted to get your sense for the size of it now. I know in the past, as we have performed on some of our contracts, while we were above any warranty levels we were contracted to be at, sometimes through the friction of just managing the fleet, sometimes from an output standpoint, we will not be to where we expect it to be. And is that still an issue financially for us, or is there a large liability associated with any performance?
It's actually very minor. If we look at it at a company level, Greg, it's less than 0.5% of our overall revenue. And we're managing that down every year. So for me, it really is something that's behind us in the rearview mirror.
Yeah. Great. Great. Listen, I think you've done a very nice job with the business. Glenn, I think you've done a nice job of positioning us to go forward. I think you've dealt with the rest of the team, have dealt with anything as Bloom went through its growing pains in making sure that we were building a profitable business here going forward and one that will be a key part of how we think about growing revenue, how we think about expanding our margins and meeting some of the financial perspectives that we laid out as part of analyst day in December. So I'm very appreciative of you and the team for doing that. I'm also very, very appreciative of the amount of effort that you and the team went through today.
I think you've done a very nice job of taking what is a very complicated business and making it very easy to understand and not only what we've accomplished, but what we have left to do. So thank you very much. Listen, we're going to end the call here. I appreciate everybody participating today. We'll make, for those that won't be able to see the presentation or want to see it again and go through the materials, we'll make those available on our website. So Glenn, thank you very much. And with that, operator, we will end the call. Thank you.
Thank you.
This concludes today's conference call. You may now disconnect.