Good morning. Thank you very much. I know, Wednesday morning, Vegas, 8:00 A.M., 8:30 A.M., it's kind of brutal. We do appreciate you showing up. I have something, unfortunately, a little bit awkward to announce about the scheduling today, so please forgive us. Last night, I was invited to participate on the CEO panel that's occurring this morning at the plenary session. Kris and I wrestled with this. This occurred at 3:00 A.M. last night. Kris and I wrestled with this because we had been preparing and rehearsing for today. It was very important for us. This is our first big analyst, more of an hour, but analyst day. Really came to the conclusion that it's probably in the best interest of both the holders of Enphase as well as Enphase for me to attend that.
Which means that originally what I was going to do is spend the first 15 minutes with you and then turn it over to the rest of the team, and then come back for the Q&A. I'll still do the 15 minutes, I'll still have my session, but I won't be here. Right after that, I will leave and I won't be here for the Q&A. If any questions come up during this presentation, either for myself or for the team, don't hesitate to ask as we go along. We will carve out some time at the end for questions specifically. Obviously, all these discussions are under the safe harbor provision. I'm going to start off at a higher level, talking about our general vision and where we're going.
After that, Raghu Belur, our Vice President of Strategic Initiatives and Products, also co-founder, will be talking about our technology. Jeff Loebbaka, VP of Sales and Marketing Worldwide, is going to talk about both the markets and our go-to-market strategy. Kris Sennesael, our CFO, will be talking about financials, both the financial strategy as well as the specifics. Let's go ahead and start. As we start, I think it's very important to recognize the scale and the scope of the opportunity that we're talking about. I think in a very real sense, we're looking at an unprecedented opportunity, both in terms of growth of a company as well as wealth creation. Today, solar is less than 1% of the energy mix in the U.S. It's less than 3% of the energy mix in all of Europe.
If there are at most 450,000, 500,000 solar homes in aggregate to date in the U.S., there are 900,000 new homes being built this year. The scale and scope of what we're talking about and how we are thinking about Enphase as we build Enphase is not about getting from 500 to 1 million or 1 million to 2 million. It's how do we do 20 million roofs in the U.S.? How do we duplicate that scale in other countries? How do we move into commercial? How do we move into utility-scale solar? There are going to be many opportunities in front of us. Our job is to be able to navigate these waters and to grow in a profitable way as we take advantage of our technology position and our leadership.
In order for solar to realize its potential, in order for solar to do what it needs to do, there are certain key initiatives that have to be met. Solar has to evolve. The way solar has existed, it cannot scale to the heights that we're talking about. We have to improve the return on investment for both the owner and the installer. We do that for the installer by decreasing the complexity of design and installation. I don't care what your ROI is if your solar installation is challenging or if it's faulty, you're not going to buy it. Operations and maintenance have to be developed to be able to be done in scale. Utilities need to be able to control and manage the grid and do that with a bunch of distributed resources. Finally, safety. Solar has a great safety record so far.
Scaling to 20 million roofs, we need to make sure that that continues. How do we achieve that? At Enphase, we have fundamentally a very deep philosophical conviction that the future of energy is about a distributed networked architecture. All of us come from the tech industry, and we've seen exactly this occur many times in many markets. If you go to a Google or a Facebook data center, you don't see a large server anymore. What you see are many small servers, thousands of small servers that are replicatable, that are inexpensive, that are hot-pluggable, that are simple. If you want to build a supercomputer today, you build it with 10,000 Intel chips on CPU blades for exactly the same reason.
For that exact same reason, we're seeing solar do the transition in the same way, from a big, large monolithic inverter to lots of small, simple devices. How does that benefit you? Well, clearly, for the owner, we harvest more energy and therefore produce a better return on investment. In fact, our energy harvest increase is embedded into most calculators today. We also produce a better return on investment for the installer because we simplify all the design and installation. With a bunch of microinverters, you don't have string sizing, you don't have high voltage, you don't have special certification and qualifications necessary. We've reduced those barriers and enabled tradespeople now to become qualified solar installers. Our reliability is far and away the best in the industry, but you need to look at reliability in two ways.
The unit reliability of an Enphase inverter is the highest reliability of any inverter. More importantly, the system reliability is yet again higher because even if an inverter does go down, the system is effectively performing, in fact, in many cases, still outperforming the original assumptions because of the extra energy that we produce. That doesn't mean that an inverter will never go down. That doesn't mean that a module will never go down. When that occurs, with a traditional inverter, you have a sort of break-fix model that says that when it breaks, I've got to run and fix it. I've got to drop everything I'm doing because now the entire system is down. This is true in a residential system, a commercial system, even utility-grade systems. With an Enphase system, an inverter goes down, the response is do nothing. Wait. It's one inverter.
We've had many of our installers tell us that that's exactly what happens. They'll wait till there is, in aggregate, once every six months, once every year, in a particular area, maybe there's a module issue, maybe there's a wiring issue here, there's an inverter issue there. They send one guy out. The guy going out is the least expensive person, and that person can go ahead and manage his entire area. Because of our communications technology, we know in advance if there's going to be an issue and where that issue is. That is scalable to 20 million roofs. That is uniquely scalable to 20 million roofs. We have a microprocessor built into each and every inverter. The inverter, as Raghu will talk to you about, is very software definable, very programmable.
We can adjust to the evolving needs of utilities, and there are multiple examples of this already, where utilities have requested us to make changes to a group of inverters to help stabilize the grid. Back in Petaluma, at our headquarters, we were able to push a button and make that happen. Then finally, safety. We are the same voltage as in a building. We solve the safety problem in a way that is impossible for anyone else to solve. We don't need big disconnects. We don't need the safety switches and all the issues that are being developed in order to solve the safety problem are solved inherently in a low voltage design. We've seen, as veterans of the tech industry, we've seen industries move in this direction.
For the exact same reason, and for the reasons I've mentioned, we're seeing the solar industry move in this direction. The question is, why aren't other people doing this? What does it take to actually make this happen? Answer is very simple. There are four buckets of technology, each one of which is complex, each one of which requires its own unique set of engineers, and all of them have to work together seamlessly to make the solution work. It starts with a semiconductor-based, low power, microelectronic solution. That takes a lot of the most advanced power electronic designs in the world, coupled with, as I said, semiconductor design, firmware, software, everything to make that device work. On top of that, now you've got these devices that are distributed. How do you communicate to them?
We had to develop our own communication technology that runs on the power line, that allows us to communicate to and from each and every inverter. It's a bidirectional communication technology that is proprietary to Enphase and, in fact, embedded in our ASIC. We're collecting data. Today, we're collecting 600 gigabytes of data every single day from over 250,000 sites in 70 different countries. With all of that data, we're analyzing it, we're reporting on it. We're alerting people when there are issues. Many of our customers are using that data for billing. This is mission-critical data that's being managed on a scale that the solar industry has never seen before, and the rate of growth of that data is stunning. You can imagine, with how Enphase is growing, that data rate is growing.
Managing that data is a complexity that very few companies in the world, forget solar, in the world know how to do today. Finally, high-volume manufacturing. There are very few industries in the world that have the requirements that Enphase has. That have the kind of volume we are talking about, millions of units a year, moving to millions of units a quarter, under the cost challenges that the solar industry has, coupled with our reliability. I can only think of two, frankly. There is the automotive industry and the disk drive industry. Outside that, nobody knows how to do this. Nobody has to do this. The manufacturing IP, the intellectual property that defines how we build these units is owned by Enphase. We have our own manufacturing process engineers. We define the entire manufacturing line.
We will dictate what equipment, where the equipment sits, what is the solder composition, what are the belt speeds, what are the temperatures. All of that is controlled and defined by Enphase. In fact, much of the manufacturing line is automated. That automation is designed and owned by Enphase. It has to occur. What I am describing to you, by the way, is, as I mentioned, very unique. There are only a couple of companies in the world that take this kind of responsibility. It works great because, again, it fits the general model where we own the IP, but Flextronics, our contract manufacturer, owns the equipment and owns the people. The next topic is on competition. I am just going to take a minute here because I think there is a confusion here. There are no reports that show microinverter market share because they would look stupid.
You would see Enphase at 97%, 98%, and you would see 50 other companies at a fraction of 1%. You do not see that report. I think as a result of that, some people feel there may be a confusion that, well, therefore, there is no competition, and we are going to wait till there is micro competition. The reality is we have had fierce competition. We have had incredibly aggressive competition. The issue is that we have beat the competition. SMA and Power-One, the two largest inverter manufacturers in the world, have had multiple generations of microinverters for years to no success. In fact, SMA is no longer even showing their micro in their booth. I cannot speak for them, but They have hunted, they have tried to compete, and were completely unsuccessful.
We have had multiple, when I say multiple, I say anywhere between one and two dozens, very well-funded startups in this space that are today, all of them, dead or dying. There is probably 30 or 40 Asian companies that have been around for four or five years trying to compete with Enphase, all to no avail. The competition has been incredibly strenuous, incredibly harsh. The difference is that we have overcome that competition, and now we are in bigger and bigger volumes every day. Now, with the next generation technology, we have taken yet another leap. Remember that they are competing with us, or have been competing with us against the inverter we are shipping today. What Raghu is going to talk to you about, our next generation inverter, is many times more complex than what we are shipping today.
We've had well over 100 engineers working on this thing for the better part of three years because of that level of complexity, because of what we need to do in order to accomplish what we need to do. The wrong takeaway is that we don't have competition. We certainly do, and there could be something that comes out that we're not aware of, which is why we're always looking over our shoulder, why we're investing in R&D to increase that lead. It's important to note that there has been tremendous aggressive competition. Today, Enphase is the last person standing. Again, all that is great, what makes Enphase investable fundamentally is everything I said, coupled with the right business model. We're really the only company in the tech industry that has in the solar industry that has a tech business model.
Very standard for all of us, very prosaic in Silicon Valley, wildly unique in solar. By that I mean we focus our OPEX on the development of intellectual property, semiconductor design, software, services, so that we can provide more features, functions, and products, as you see here, to our customers to help elevate our ASP while working with Flextronics, working with a contract manufacturer to continuously reduce our price and build products at the highest possible quality. That combination is where you see the gross margin expansion. That combination is the definition of a Silicon Valley hardware company. That's who Enphase is. Again, this is a business model with decades of proven success. Again, uniquely in the solar space. What does the future for Enphase hold? It's unbelievably rich.
We have never, at Enphase, been as excited about what's going to happen in the future as we are now. The residential space is going to turn from just solar into energy management with the advent of storage and load management. You're going to see storage and solar coupled together in the utility space. Again, Raghu will talk to you about the products that we have that are going to support that. Utility scale solar systems. These bins, the 250, 500 MW inverters are dinosaurs. They are dinosaurs. As we enter the utility space, we've said many times that we are entering the utility space, people are going to understand that a distributed network architecture is the only real architecture that scales. The microgrid market. We're doing a lot in the microgrid. We're doing a lot with the U.S. military. We have several microgrids in Nigeria.
The microgrid market is yet again uniquely suited to an Enphase solution, in part because of its location and the challenges with O&M. Utility services. Again, you're going to hear more about this, we know more about the distribution and feeder networks that we're in than the utilities does. We are already in discussions with multiple utilities, talking to them about how we can supply services to them that not only help them monitor the grid, but help them control the grid through our inverters. There's a host of other new markets that we're not even ready to talk about yet. Finally, in conclusion, I think we all probably agree that solar represents an unprecedented market growth opportunity for those companies who are in the right place to take advantage of it.
Enphase is an undisputed technology leader. As of today, with what we're announcing, we've taken yet another leap forward in technology. We have the right investable business model. Finally, the opportunities in front of us, both through market and segment expansion as well as geographic expansion, are simply stunning. If there are any questions, I'll take them right now. If not, I'm going to pass things over to Raghu Belur, our Vice President of Strategic Initiatives. Again, I apologize, but I'm going to have to run. I will hopefully catch up with you back at the booth. Thank you very much.
Good morning. Thank you very much. I want to start the video, but before I get there, what a first day of SPI. Our booth was just nuts. It was absolutely crazy. I don't know if you guys got a chance to come by or not, but it was absolutely wall-to-wall people. It was just fantastic. I want to start with a quick video here.
In the solar industry, quality and reliability are extremely vital. Your solar system should be up, it should be running, it should be effortless. The technical challenges of achieving that at volume are incredible. We need to build a product that can support 20 million roofs. How do you do that? Well, that's a completely different level of reliability that hadn't been seen in the inverter industry before.
Bhuṣan and I worked in the semiconductor industry. It's ingrained in us to actually think about system reliability first and foremost.
Quality is built into the entire product development life cycle. It's not just quality of a product or the quality of a system, but it's quality of the whole solution and the quality of the company that stands behind it
All of this came out of the high-tech business, we've applied really high-tech principles to solar. Before we even go into the planning phase of the product, we outline all of the standards that have to be met. Since our system has to perform at really wide temperature ranges, we choose components that not only do that, but will operate well outside of it. We have certain components, like our capacitors, that we actually do years worth of testing before we put them in the design. The system is constantly communicating with our Envoy device that we can pull real-time data from the inverters to know exactly what they're doing and push data out to them to either upgrade them or fix problems if we have them.
When we do a product development, we go through a series of stages with successive higher stages of validation testing, culminating in 1 million power on hour tests under stress conditions. The performance data from the field is critical to our process. We have upwards of 5 million units in the field that does about 400 GB of data a day. We test the units under stress conditions, simulating lifetime. We test to very specific temperature conditions. We do thermal cycles, and these tests themselves last months. Those chambers are validating the reliability of the unit under very high stress conditions. We do ongoing reliability testing in production so that we can assure that what we produce in production is equivalent or better to what we produce in development.
Flextronics is our manufacturing partner, they have a lot of experience with products that have to work for beyond 25 years, which is a different mindset. We have Enphase employees embedded with the Flextronics team. They oversee day-to-day production. We build 15,000 units a day. With automation, we minimize variation and improve quality and repeatability of the product itself. We developed our own manufacturing monitoring system from the U.S. We have real-time views of all of the different assembly operations and test steps in the process. The quality inside of Enphase is in the organization around the product. Build the product well, test the product well, that's extremely key. The fact that the company is willing to take the steps and make sure that the customer is satisfied, that's the difference between I have a quality product and I have a quality system.
Quality is not something that we do in addition to our job. Quality is what we do. It's embedded into everything that we produce.
It's great. As you can see clearly, we take quality and reliability very seriously, and this is not just at the manufacturing step, but starting all the way from conceiving the product, the design, all the way through customer support and service. As we jump in, I want to get into kind of the technical details here, if I may. When we started the company, we never thought about the inverter as a piece of hardware. At Cisco and my background is in telecom, we always thought about it as a system. The system includes the power electronics, the communications, the software, and all of them working seamlessly together. While we have done a lot of innovation on the power electronics side, we have built this really extensive communication and software infrastructure.
Just to give you a sense of scale, as Paul mentioned, 600 gigabytes of data gets managed on 250,000 installations in 70 countries. Right? In that sense, what we have done is we have built this operating system where solar is just one application that's running on this operating system. Now you can think about how that operating system is extensible in that now you can think about adding storage, load management, utility services, et cetera. We'll talk more about that. As you can imagine, it's very IP rich. We have filed 125 patent families. Each family has multiple patents behind it, out of which we've been granted already 48 patents, and these are very rich patents in all areas, power electronics, communication, software, et cetera. Now, again, when you think about this system, think about the inverter as the kind of physical layer.
Inverter communicates with the gateway. Gateway, the software is the platform layer, then the gateway then communicates with the cloud, and that's the application layer. All of this combination of the physical layer, the platform, and the cloud all working seamlessly together as a system. The inverter itself is a digital architecture, the heart of which is our own custom semiconductor. Right? It's an incredibly complex device. It's not complexity for the sake of complexity. It's what's needed as we go further and the grid itself becomes more complex. I'll give you an example of what I mean. Right? I worked at a startup, and Martin and I both, we worked at a startup many years ago. We were doing cutting edge optical communication technology, absolutely cutting edge. When the product came out, we blew the doors off the market.
Cisco bought us for $7 billion at that time. Right? This product, this microinverter, is an order of magnitude more complex than what we did back then. It's because you're trying to optimize performance, reliability, cost, all for high volume and in high volume manufacturing as well. Now with the 5th generation of product that we just announced, arguably the most complex, technologically complex device we have designed. It took us about four years. It's not complexity for the sake of it, but it's what's needed as we go further, I'll talk a little bit about it on the next slide. Right? This inverter itself, Paul alluded to this, is software defined, meaning, as I said, the digital architecture with the ASIC is running a microcontroller in it. We can completely alter the behavior of that inverter through software.
Why is that important? There are a couple of key advantages to that. PV or solar is a 20-year asset, if you look, the grid conditions are changing dramatically, load conditions are changing dramatically, the weather conditions are changing dramatically. You cannot expect that you can have a static PV system and it's going to last 20 years. It's just not going to happen. You need a system that's fully adaptive. You need a system that you can push remote software upgrades to and alter their behavior. You're seeing this play out in places like Hawaii, where the utility called us and said, in a high penetration solar environment, they said, "Can you alter the behavior?" We were able to do that.
First, they said, "Hey, we think there are 1,000 sites." We said, "No, actually, there are 5,000." They didn't know how many sites they had on their feeder network. We knew, obviously, because we have that information. We were able to go in there, do the software upgrade. Of course, we had all the audits in place. We did the upgrade and changed the behavior. I think that's going to continue. That's not new. Anytime solar is high penetration, you'll have to change its behavior and alter its behavior. Another example of this is when you go into a new geography, we don't develop a microinverter from the ground up to go to a new geography. In most cases, we just do a software change. Just an example of that is the requirements for France, Belgium, U.K., Australia, all different. Regulatory requirements are completely different.
We make software changes, there's a lot of leverage that you get from that model. Talking specifically about our S-Series or 5th generation microinverter. 97% average efficiency, 275 watts. Again, when I look back when we first started, they said 94% could be done. That's where we started. Here we are at 97%. What's unique about this is the following, right? Solar anymore cannot be simply attached to the grid. It's just not going to happen. As solar penetration continues to increase, you have to integrate into the grid, which means you have to have the capability of providing advanced grid function, things such as power factor correction or VARs, both VAR types functions. This gen 5 inverter that I have here does all of those things.
There's one key feature that it provides that I think is an enabling technology to other solutions as well, and that is, it's a fully bidirectional device, which means it's really not a microinverter, it's an ideal converter. It can go DC in, AC out, AC in, DC out. We're joking about it. You can even do DC in, DC out. Why is that so important? We'll come to that in a second. Before we get there, I want to talk about kind of what our broader vision is. Our broader vision is, here's a residential example of that is, how do you provide to the system owner the lowest cost, highest reliability, and seamlessly deliver that energy to the system owner? It's in the context of a total energy management solution where you have PV on the roof. Again, it's all distributed network architecture.
As Paul mentioned, we don't have any centralized power conversion. You have PV on the roof, you have storage in your garage, you have control, and you have a smart appliance or a smart thermostat, for example. All of them being managed seamlessly through our energy management solution. This is a manifestation or a realization of that operating system model I was talking about, where solar, storage, smart appliance, et cetera, all comes together, and it is what we are leveraging for our solutions. Here is a quick video of how that plays out, how the energy management solution plays out. Again, it's PV on the roof, storage in your garage, control, and then in this case, then some management through, you have a smart thermostat, again, to manage your loads, for example. Solar in the morning, comes up, solar starts producing.
It's not producing a whole lot because it's 30 in the morning, you're drawing some from the grid. You go off to work, and now what happens is you have too much generation of solar. When you have excess generation of solar, you start charging the batteries. Once the batteries are full, you start feeding power back into the grid. You come back, well, 6:30 is kind of early to come back, but nevertheless, you come back home, and then now what you do is you start, there's still a little bit of solar left, as the evening comes along, you start discharging from your batteries and start feeding all the loads. This is an example of how that whole control system or that whole energy management solution works, and you rinse and repeat the next day. Now let's talk specifically about storage.
Pardon this kind of engineering line diagram, again, philosophically, we fundamentally believe a distributed network architecture is better. That's what we have today. That's the Enphase solution in that you have a PV system, you have panels on the roof attached to a microinverter, and that microinverter converts from DC to AC on a per panel basis, and it's producing, say, in the home, 240 volts, standard AC voltage, standard homemade voltage, and it's connected through our Engage Cable or standard AC wiring into a 20 amp breaker, right? The reason why we did this, there are four reasons. One is you produce more because you don't have a lowest common denominator problem within the panel. They're all connected in parallel to an AC module. The second advantage is no single point of failure.
It's a very safe solution because you don't have high voltage DC, and it all gets managed through our gateway. All of them communicate over the power lines to our gateway. We applied that exact same principle to storage, and we said, just like you have an AC module on the roof, you have an AC Battery in your garage. Unlike every storage solution that's out there, regardless of what they say, unlike every storage solution that's out there, they're all centralized topologies, which means you have a battery bank and a big central inverter that does the conversion. It suffers from the exact same problems. You are as good as the worst-performing battery, worst-performing cell in that battery bank. You have a production problem.
You have a single point of failure. That inverter fails, in some cases, both you lose PV and storage, and in other cases, you may lose just storage. The third problem is it's extremely complex to design and install. You need three people. You need special equipment. These battery banks and inverters weigh 100 kilos, 150 kilos. You can go look at them. They're refrigerator-sized stuff. Finally, it was bad enough that people were routing high voltage DC all over the place. Now, couple high voltage DC and high energy chemistry, I'm not sure that's very smart. What we said was, "Look, we're going to do an AC Battery, break everything down into very small, manageable chunks." You address the production problem because you don't have, or you completely minimize the lowest common denominator problem here. Very small number of cells in a cell module.
You don't have any single point of failure. The module fails or the battery fails or the inverter fails, only that's isolated. It's completely plug and play. AC coming in and AC coming out of it. By the way, AC coming out of it, you just plug it into a standard AC bus and goes to a 20 amp breaker. Very simple to design, very simple to install. We can drive that cost down. Finally, no high voltage DC. From a safety standpoint, arguably much safer solution. Very modular, very scalable. You can start with four kilowatt hours and go to five, six, seven. You can grow it as you want. By the way, this exact same architecture, this exact same topology works in commercial as well. This whole system gets managed.
If you think about the cost of storage system today, it's kind of there are four costs that drive storage systems. The chemistry cost, the electronics or the power conversion cost, the installation cost, and the integration cost. Let's talk about it for a second. Chemistry cost, we know that's headed towards commoditization. I see numbers like $150, $100 a kilowatt hour. I say yeah, absolutely. No question about it. It's heading towards commoditization. No question. The second cost is power electronics. We use the same, it's the bidirectional microinverters, the S275 that's in that AC Battery. This is leveraging the scale of PV, millions. We can drive the cost of power electronics down substantially. Installation, like I said, it's actually a one-person install. It's all plug and play. The AC Battery weighs about 40 pounds.
You literally hang it on the wall and run it into a 20 amp breaker. It's a one-person install. The fourth element is, unlike PV, unlike solar, you put panels on the roof, when the sun shines, it converts energy. It's not good enough to have storage in your garage. It's just not going to work. It's not going to do anything. You need an expert system that's there that's managing your solar and PV and making the decisions on when to charge, when to install, when to discharge. Because it uses the same communications and software infrastructure, it's just another application, it's seamless. Plug it in, and it just starts pretty much working at that point. To give you some specifications, it's 1.2 kilowatt hours. We use lithium iron phosphate. Safety was a very big driver. Having said that, we are chemistry agnostic.
Just like we are panel agnostic, you can put a Trina next to a Yingli without micros, you can put any chemistry in there. We picked LFP to begin with, particularly this particular partner that we chose, ELIIY Power, fantastic performance, designed for stationary application, and very ultra-safe batteries. We looked at all the testing, was very impressive. Like I said, very modular, plug and play, and fully integrated into the Enphase solution. This is what this guy looks like. That's the AC connector that comes out of it, just plugs into a standard trunk and drop cable that we use for PV as well.
Prismatic cell?
Yes. We looked at both 18650 and 26650 because of the packaging, thermals. That ELIIY cell will discharge down to -20 degrees Celsius. You can charge to -10 degrees Celsius. It was a very big driver for us because we want these to sit in the garage, we want to do passive cooling on it. We know that has a big impact on RTE and cycle life. We took all of those into consideration. We actually designed our own, what I call an effective energy model, that takes into account cost, RTE, cycle life, DoD, at different temperatures convolved to our application. It was quite obvious. That's why when somebody says, gives you a spec, it should be followed up with half a dozen questions, because not everything is that obvious.
I want to be very clear that when we talk about energy management, it's not just applicable to residential. It applies just as much to commercial. Just the application, the core elements of it are the same. You've got PV on the roof, you've got storage, you've got load management, it integrates into BMS. Your optimization function is different. You may be doing demand charge reduction, or you may be optimizing for highest availability. The underpinnings of it are the same. That's the beauty of that operating system model that we were talking about. I want to make one point here. When you think about commercial, we don't need to make it this pretty, and it doesn't have to be aesthetically so nice. We did it for a residential application. For commercial, it's exactly the same concept of a distributed network architecture.
It's just repackaged in a different form factor, which we are working on right now. What's next? Because as part of my responsibility of strategic initiatives, I'm looking at things like microgrids. I'm looking at things like utility-scale deployments and utility services. I'll talk about utility services in particular, but let me just touch on microgrids. It's absolutely the future of energy, both in developed countries as well as in developing countries. In developed countries, it's about resiliency, and in developing countries, it's going to be about energy access. In our opinion, microinverters are a perfect fit for those applications. One, because of the software-defined nature of it, because of energy management, because of ease of O&M. It will be the biggest driver, and I think it's absolutely the future. We have a number of projects that we are involved with right now.
You'll see some announcements about them. Utility scale solar, we have always believed that microinverters are a great fit for utility scale solar. We haven't done it until now in the interest of focus. The value proposition, particularly around operations and maintenance, holds very strong in utility scale services. In addition to yield, in addition to CapEx, not just on the inverter, I'm saying the system level CapEx, plus O&M. All driving towards, we know it's an LCOE sale. Again, stay tuned. Let me talk briefly about utility services. One of the challenges with solar is the technical burden it places on when there's very high penetration of PV, and the utility is not ready for all that reverse power flow, because now you have power going into the distribution network.
With the infrastructure that we have, with all the deployments that we have, for the first time, the utilities have rich information deep in their own network. In fact, with all due respect, we know sometimes more about what's going on their feeder networks than they do. That's because part of the monitoring that we do is not just on how the inverter is doing. We monitor how the panel is doing, and we also monitor how the grid is doing, because they are directly sitting on the grid. I'll show you an example. I can't disclose where the location is, but I'll show you an example of high penetration of solar. It's a time lapse on a day where grid events took place, where the grid was out of specification, that caused solar to go offline. Not uncommon, but we have that information.
Once you have that information, couple that with the fact that we have bidirectional comm, and we can make adjustments to that inverter. The utilities are seeing that as very valuable. Having said that, we don't have any business models built around it as yet. I'm just stating what the utilities are telling us. This is a heat map of all the installs that we have. Actually, you can go to our public site, and you can actually see them. As you can see, all this number of systems installed worldwide, residential and commercial. Here is that I was talking. You'll see these red flashes. Now, that may be a site that has an issue, but you see that big event? That was something happening on their network. You see these big flashes. We have all that information.
We know exactly what happened and what was the cause of it. In conclusion. Yes.
You have a complete question with respect to inverters. Can you just give us a little bit more detail what an event is and how you actually help?
Yeah. In that particular case, it was what's called an AC voltage out of range, meaning the utility is expected to maintain a certain voltage. When you go measure the voltage in that socket, there are very specific federally governed rules that say, "City must stay within that range." Because of high penetration of solar, they're having a hard time maintaining that range.
There could be spikes in production, clouds going.
Yeah, clouds go by. Your frequency may be drifting because it's not fast enough to catch up. All that PV comes on and off. It's like a massive amount of loads coming on and off, inverters and solar. It's causing that challenge. We detect all of those. By law today, by certification, they say, "When you have an out of voltage or an out of frequency event, you have to jump offline." It's the law today. What we are saying is, "Hang on. What we can do is ride through some of these transient events or widen the window." In some instances, we did widen the window. Now, we are proposing advanced grid functions where we can ride through some of these events.
What you can do is kind of get around it, lower the amount of voltage coming from certain locations, or is it that you can go and add to that?
What we can do is we can say, "Hey, this is a transient voltage event." We can say, "Do not go offline." That way, the primary generation source doesn't see a massive change in load. It sees right through that transient event. You can say, we're going to change the operating parameter that says, do not trip offline between, today it's set at, say, 59.7 hertz to 60.3 hertz, widen that window. Even if that event occurs, it rides through it. That's just one simple example of what can be done. We have done that. Again, I can't disclose where it is, this happens a lot.
You talk a little bit about bidirectionality.
Yes.
My understanding is that you guys have a resonant topology that you're working on.
Yeah.
There's only a couple of companies working with this approach. Obviously, the form factor is complicated, so to be able to influence that. How complicated is it for someone else to take that approach and design something and start to catch up? Because that seems to be the key to demand energy storage effectively.
Yeah. I think the evidence has been that it's been very difficult to catch up with our previous quasi-resonant topology. I think this is even much more complex. Like I said, arguably the most technologically advanced work that we have done. I think it's just that much more complex to provide full bidirectionality, all the advanced grid functions, and still be at 97% and 275 watts, because efficiency also drives reliability. In terms of magnitude, I can say this is much more complicated than what we have ever done. That itself, people are finding it hard to catch up on. As Paul talked about competition, we've had an incredible amount of competition right from the beginning. I mean, ABB and SMA, two big behemoths in the solar industry, have been trying to do micros, and this is using the older topology.
The advanced grid technologies, actually string inverters have been able to do that for a good period of time. It's mandated in Germany. What do you think this does in terms of your product positioning in Europe, and how soon does that hit?
Clearly we can do those, but it's also degree to which you can do it. That's my point. With so many changes taking place on the grid, I mean, having been in this industry for eight years, the grid is more different than ever before. It's because there is more PV, solar on there, I keep saying PV for photovoltaics, but more solar on there, EV coming online, et cetera. The grid and the weather conditions and load conditions are changing so dramatically, you want a completely software-defined system. Which means it's not sufficient to have, hey, here's a list of eight advanced grid functions like volt-VAR, Frequency-Watt, ramp rate control, ride through, et cetera. You need something a lot more flexible than that. When you have a hardware-based solution, you're going to be limited by how much you can do.
I'm arguing that four years from now, you'll need maybe two slopes on your ramp rate control. Anything is possible.
What has to happen between now and when you start shipping that product in order for it to be corporate and able to further solar penetration in Hawaii?
There are a couple of things, right? One is, it means doing what Queensland is doing, where they're saying go to zero export and then I'll allow interconnection. I think storage, discussions have been going on. There are RFPs out there to provide storage. We have had those discussions, and the utilities are saying, "Yeah, if there is storage, we can foresee allowing that." There's got to be policy also has to catch up, who pays for it, et cetera. I think there are a few things that need to happen. Some rules have to be set up, and they're already working on Rule 14H, which is the Rule 21 equivalent in Hawaii. Policy, technology, I think, will be ready. Policy needs to get in place, and the PUC needs to approve it. All the regulatory stuff needs to occur.
It's absolutely a fertile ground for where this thing's going to happen first. What's really interesting for us is that it's the canary in the coal mine. It's going to happen everywhere. San Diego, I think, is going to be next. I mean, San Diego's got very high penetration of solar on a pretty tight distribution network.
You actually have the product now. I'm just wondering, does it threaten that at all to
I would say one thing, though I'm going to talk about markets and TAM and the model for our success in entering new markets. To answer your question, it's not just about entering new markets with our existing solutions, say, which is, as Raghu showed, solar generation, but then having those same go-to-market engines available for other products and solutions that are sold in addition to solar generation, storage, home energy solutions. That becomes very large, very quick. The answer is, we can leverage what we're building today throughout the market and the relationships and bring more products and more forward solutions, which will help grow business.
In order not to be confused, storage is not going to drive any meaningful revenue this year or 2015. We need to have the product out there to start introducing it to do alpha and beta sites to share with the rest of the solar industry what we are doing and get ready for the ramp probably in 2016, 2017.
Quick conclusion, and then I'm going to hand it off to Jeff. I think we are pretty tight on time. Again, I want to re-emphasize what Paul said. We are absolutely a technology company, and our goal is to bring innovative products and solutions. I keep saying this is about relentless innovation. That's what we are about. With that, Jeff.
Thank you, Raghu. Good morning. My name is Jeff Loebbaka. I'm the Global Sales, Marketing, and Service SVP. I want to spend about 15 minutes going over a little bit about the solar market, growth trends, sizing, and then more specifically, how does that relate to Enphase and our growth drivers over the next several years. If you look at the data, it's a large and fast-growing market. It's an attractive market, expected to be 60 gigawatts in 2016, a 20% compounded annual growth rate. If you go back to some of the things that Paul mentioned in terms of some of the data points from the International Energy Agency, this could be very conservative if you believe some of that data.
It's a large and attractive market, and what's important is today, even as much growth as we've realized over the last several years as we've grown, we're just playing in a small segment of this marketplace. How do we turn that into even bigger growth opportunities? What are the primary growth drivers that we see at Enphase over the next several years? The first is we've been very successful, as you know, in our core market, U.S. residential. We now have over 40% share. The question is, well, that's pretty high. How do you grow from there? Where do you go? Yes. When you're at 40% share, you're getting incremental share growth. It's tough to make quantum leaps from there. That market is going to grow from one gigawatt or so this year to three gigawatts in the next few years.
That's going to drive a lot of volume, number one. Number two, major growth driver is U.S. commercial. We have just scratched the surface in U.S. commercial. With primarily a residential product, we've been very focused on small commercial. Small commercial, you can see, is just a small piece of the U.S. residential marketplace, and we've been able to be very successful there, getting close to, we think, at the end of this year, about a 40% share in exit run rate. When you look at our new C250 and what we believe we can do in medium and large, getting into those segments, that's a 30x growth in the market availability from 250 MW to almost a 9 GW TAM in just two years. Now, we're playing in a much bigger pool. Thirdly, we're going to continue our geographic expansion.
New markets, new countries, as we've done to date, we've successfully grown in a number of new countries, but we're still only playing in about 2 GW. If you look at our overall combined share, we're at about 5%, and we're growing. As we look at the next couple of years, we talked a little bit about this new S275, for example, in Germany, countries like that, opening up that market gets us access to another 20 GW of market opportunity. Now, we're going from roughly playing in a very small 3 GW to 4 GW space to 30 GW, which at that point will represent over 50% of the market opportunity in 2016. A tremendous growth in access to market and opportunity. How are we going to be successful? How are we going to win? Well, we've got a proven model.
We've learned a lot over the last few years in terms of what it takes to enter, build, and grow new markets. We have an incredible brand, a brand that stands for the best quality reliability, the leader in technology innovation, and best-in-class support. We have incredible distribution and reach. Global partnerships with companies on the distribution side, like Rexel, players like SunEdison, where we have a very large and growing piece of their business in the residential small commercial business unit today. We're expanding that into Europe and Australia and APAC. Major PV module manufacturers where we have global relationships, like SolarWorld. We can take that and leverage that, and that helps us go big and go fast and get access to those markets. We have very deep and strong partnerships with market leaders.
As an example, in the U.S., seven of the top 10 U.S. residential installers use Enphase for a significant piece of their business. As a matter of fact, more than half of those use Enphase almost exclusively. They've made the conversion, they see the benefits of the proposition, and they're doing Enphase all the time. As I mentioned, we've learned a lot in the last couple of years in terms of what are the right strategies, what's the playbook, what go-to-market strategies do we have to execute, what investments do we need to make in the market? We're able to take that now and apply it to new countries and new segments as we enter them. A lot of years of blood, sweat, and tears to build that model.
The Enphase proposition is very compelling for both the business proposition for EPC and installers and as a value proposition to project owners, whether they're homeowners or commercial owners. What does that mean? The Enphase system totally changes the way that the economics and the overall system performance is done. While CapEx is important, it's not the only thing. This focus on $ per watt of CapEx, while important, is really not the most important thing. It's about the system IRR. It's about the levelized cost of energy. The Enphase system drives lower levelized cost of energy, which drives higher IRR. Whether you're an individual homeowner, whether you're Credit Suisse with 20,000 homes in a portfolio or whether you're a project commercial owner, you're getting the best IRR System and make more gross profit. It's that simple.
Not only that, but as you adopt Enphase, you get operational efficiencies. You're able to reduce costs because of simplicity in terms of design, labor, and you actually wind up saving money. You carry one piece of inventory, right? All these things are huge in terms of business processes. That's how we win. That's what's enabled us to do this in our first important growth driver, which is the U.S. residential, and get this kind of share growth consistently year-over-year with a leading share position of below 40%. If you think about this category leadership in terms of the position, you think about the fact that it's going to be a 3-gigawatt market in 2016, and a market that's growing over 50% in terms of MW year-on-year. That is going to drive a lot of growth.
In other words, excited we're not expecting huge share growth, but our leading share position in a huge market that's growing very fast is a good place to be. Let's talk about commercial. As I mentioned, even with primarily more of a residential product, we've had a lot of success in the small commercial space. We got up through the summer over 150 MW of commercial installations installed in the U.S. This year, as you can see, we expect to continue that growth and install 50 MW. That represents over 7,000 systems of anywhere from smaller systems like 10, 20, 30 kW, up to our largest system, which is 2.3 MW. Enphase is being used, it is successful, but it's not even yet with a product that is optimized and designed for commercial.
That gets us roughly to about, at the end of the year, about a 40% share of the space. Because the medium large commercial space is much bigger, it's only about 6% of the overall commercial. Lots of upside opportunity. Again, what's driving that is the industry-leading best business proposition for the EPC and installer and the best value prop, the best IRR for the project owner. Why are we well-positioned to ramp in medium large commercial in the U.S. and drive growth? You've seen our announcement, and hopefully, we'll show you some of the new products that we've just launched. The C250 is designed and built just for the commercial market. 480 volt three-phase. We haven't had that before. More importantly, though, we recognize that it's not just about the Enphase system and the hardware and that piece.
It's also got to be combined with a complete end-to-end solution. Design and quoting tools, providing access to financing partners, and most importantly, we think as a new and very unique differentiator, O&M. I'll talk about our O&M offering that we're going to have as part of our overall commercial offering. We, again, are going to work our playbook and have dedicated sales, marketing, technical resources focused on the commercial market, commercial EPCs, commercial developers, just like as we've done in other countries as we ramp and scale. Again, we believe that the levelized cost of energy in Enphase systems in commercial is going to be consistently about 10% lower than string and central systems, which means a higher IRR. Let's talk for a second, though, just about O&M in general. We've talked a lot about our technology platform, a distributed approach. It dramatically changes O&M.
I don't know how familiar you are, but O&M is an ongoing cost that is pretty substantial in terms of the cost of maintaining and optimizing the performance of an asset, the solar system. We're pretty confident that because of Enphase system, it's roughly half the cost on a yearly basis to spend on O&M than you would have to spend on a traditional string or central. Why is that? Because we have a distributed architecture. There is no single point of failure. It's not a break-fix model. It's never emergency. It's not the system is down, it's out for a week, you've lost energy, it's an emergency, you've got very expensive resources out there trying to fix it.
We've got big data with our cloud-based Enlighten system that provides incredible visibility, analytics, and information to the project owner to know is their asset being optimized, and how can they improve it. It allows proactive and predictive O&M, and that in turn drives a much more cost-effective model for O&M. It's a totally different paradigm from the way things are done today and a big enabler. That's why what we're doing as part of our commercial offering is actually offering an Enphase service offering. Because not only do we have the technology capability, but we actually have built up an incredible infrastructure that's a competitive advantage. We've got 50 technical people on the phones. We've built tools that help us analyze what's going on and react, so we can act as the service agent for the project owner, guaranteeing 98.5% uptime, which is huge.
Which is huge for a commercial owner. Handling all the monitoring and analytics, providing all the proactive system maintenance, and on-site corrective and post-warranty work, make sure that that asset is up and running, and do it and offer it to a project owner at a lower cost, which is better economics and gives them peace of mind. We will leverage this infrastructure in a very big way that, again, combined with the overall system and economics, makes it even more compelling, we believe, in commercial. The third driver. That was the second driver. The third and final driver is about geographic expansion. This is our footprint today, and you can see there's still a lot of markets and there's still a lot of countries that we aren't- To enable us to. We've got the model, we know how to scale it.
Let's just talk about an example about how that works and what it's represented. In the U.K., we entered about 2 years ago. We started shipping product about 2 years ago. The U.K. is going to be about a 400-megawatt market next year. It's one of the largest, if not the largest, residential market in Europe. You can see in the chart, consistently, we've grown share to the point where we believe this quarter we're going to be about 11%-12%. In other words, the same chart that I showed you for the U.S. growth, we're starting to see the U.K. Even though we entered at a much later date, when the market was more mature, when it was a much more competitive priced, competitive marketplace. We've done that because we've got a scalable model. We know how to build it. We know how to enter markets.
We've dedicated the resources. We've hired the resources for a dedicated team, and we've been able to be very successful. We can take that into other markets as well. Yeah.
Is your pricing and market profile different in the U.K. than it is in the U.S.?
I'll let Chris talk about that.
Yeah. We have been very clear about that in some of those international markets, it's a much tougher competitive environment, pricing is much tougher there as well.
Price points in the European market are a little bit lower than the U.S. Our relative premium is pretty consistent, we're able to still get that. We're still winning over, just as we do in the U.S., new customers with that. In conclusion, we are just at the beginning in terms of Enphase and where we are. We've got significant growth drivers and expansion into markets that will help drive significant growth for us in the future. As I mentioned earlier, it's not even just today about solar generation, but if you look at it over the next two, three years with this sales and marketing and go-to-market engine, putting storage in there and other home energy solutions, as well as traditional solar generation, which is what we're doing today, opens up a very large opportunity.
With that, I'd like to turn it over to Kris Sennesael to talk to you all about the. Yes, sorry.
One question there on the commercial, you already showed the 50-megawatt number. It's not driven by the new product, it's all just the existing.
It's all existing.
Why has it jumped from 13 to 14 in the market? Is it something?
The market is growing well, We've continued, we think, to fully gain share. It is reflective of good growth in the market, as you've seen overall in some of the markets primarily, and some good execution and continuing to grow share.
Any new customers that you have managed or anything like that?
Yes, there's definitely some new customers. The interesting thing is in commercial, it's a pretty fragmented market, whether it's small or medium large. It's not like it's easy to point to and say there's this one guy that you've successfully won and we've got a portfolio of 20 MW. It's a very fragmented marketplace. There are definitely some new partners that we've won and some new business, as well as existing customers. Yes.
On that issue of, on that earlier question about the competitive aspect of the U.K., maybe you could explain a little bit more about what the nature of that is in terms of technology and solar. What's making that market more.
Europe in general has been more competitive simply because they've been a little bit more mature markets. You had some dynamics where because of the incentives, like FITs, for example, you had a country like the U.K. where the FITs were very high. It literally was a land rush. Everybody got in and made a lot of money, and they changed it to make it more sustainable and viable long term. In there was a great fallout where the market got much smaller, and all these guys were competing against each other and driving prices down. You had that dynamic in terms of maturity as well as dynamic and the change in the market that made it more competitive. Just in terms of the overall price points relative to the U.S., they're not like a magnitude of difference, but they are definitely lower.
You see that in some other markets like Germany and places like that. A lot of that happens sometimes when there's major changes in policy and incentives. What happens is the whole ecosystem and models have to change, and sometimes that change is very challenging and it causes some bad behaviors. Right. In this case, you got a lot of guys trying to stay in business and just giving away on price. You go into that space, you're saying, "Well, here's this totally different proposition." You got to really work it. Yes.
You talk a lot about the benefits of your product versus string inverter and how effectively you're saving on the installation side of things and how that. Can you quantify a little bit about the savings on those aspects and also discuss the payback period for the higher upfront investment of your product versus the string inverter?
Yeah. If you were to ask some of our installers, we get a range of what they tell us they feel they're saving. By the way, none of this is scientific because a lot of our customers aren't necessarily in, they're not running Oracle and that kind of system capability. If they look at their business and they truly factor in all the different pieces, for example, they have less design people. They can do it a lot faster. If you look at it, inventory carrying costs, installation labor, those types of things, when they really modify their model for Enphase. A good example of that is the old model used to be is you have a high-cost DC electrician on the site for the whole day. That's really expensive.
A lot of guys who do Enphase will just have an electrician. He comes, and he's there for one hour, and he connects it, and it's done. Things like that, we hear anywhere from at least $0.05 a watt to maybe up to $0.08, $0.10 a watt. It depends. That's an installer saying, "Overall, if I look at all my businesses and all the different resources and different functional areas, I'm saving that," and that's meaningful to them. In terms of payback and return, I would just put it more in terms of, we tend to look at IRR. The net of it is we look at the IRR, and it's better. I can't quote necessarily exactly what it is, but we believe it's enough to be meaningful. Typically, what happens is that's because you're factoring in your IRR, higher energy.
If you're looking at a 20-year system and you buy a string, guess what? You're going to buy a new string in year 11 or 12. You don't have to do that with Enphase. All those things are factored in. The IRR is better. The payback isn't much different at all, if it is at all. We tend not to look at payback. We tend to look at IRR, so I can't quote exactly. Great. Let me turn it over to Chris. Thank you.
All right. Thanks, Jeff. Good morning, everybody. I'm Kris Sennesael, CFO of Enphase. After Paul shared with us the vision of Enphase, Raghu talked about the key technology building blocks that is really helping Enphase to transition from a microinverter systems company to an energy management company, Jeff talked about the success in our key markets, as well as some of the growth markets that we play in, I'm going to wrap it up by talking about our balanced, profitable growth strategy. If there's one key takeaway I want you to take away from this, it's that accelerated growth drives EPS expansion. I repeat, accelerated growth drives EPS expansion. I'll go back to this at the end of my presentation. First of all, let's make sure we all understand what it really means, driving a balanced, profitable growth strategy.
Maybe I'll start with saying what it is not. Driving a balanced, profitable growth strategy is not just about focusing on the top line and trying to grow the top line as fast as we can, throwing a ton of more resources in R&D until the marketing added, and never made a $1 of profit. There's too many companies who have done that. That's not what Enphase wants to do. It's also not just about trying to maximize the profit here in the next 12-24 months, and stop making any investments in new initiatives, developing new products, making that transition towards energy management, and maximizing the profit in the next 12-24 months. That's not what we're trying to do, because that will slow down our growth 3-5 years down the road.
We really want to try to strike the right balance between focusing on top-line growth under constrained conditions, while at the same time improving the bottom line. There's basically four key elements to that. It's about all of our revenue growth, gross margin expansion, driving leverage through operating expense management, and that results in EPS expansion and EPS growth. First of all, let's have a look at how have we executed on those four key metrics in the last six quarters. As you can see on this chart here, on all four of them, they're all trending in the right direction. If you look at revenue, you can see nice sequential growth in 2013, quarter by quarter. Of course, followed by a normal seasonal decline.
In this case, it was actually better than normal seasonal decline in the first quarter of 2014, followed by an acceleration of the growth in the second quarter of 2014, with revenue being up 41% year-over-year or up 42% sequentially. That revenue growth is mainly driven by the U.S. residential market, as well as our execution into our growth markets, as explained by Jeff. If you look at the gross margin, you can see that the trend of gross margin improvement, which started many years ago, we had 10% gross margin in 2010. We doubled to 20% in 2011, go to 25% in 2012, to 30% in 2013. We've continued to further improve the gross margin all the way up to 33% in the second quarter of 2014. Over a span here of six quarters, that's up from 27%-33%. That's up 600 basis points.
Great execution there. In terms of driving leverage in the model and being financially disciplined in our operating expense management, you can see we also made continuous improvement. The operating expense as a percent of revenue in Q1 2013 was 45%. We drove it down to 33% as a percent of revenue right now. Of course, the combination of all of that has improved our bottom line. The EPS, non-GAAP EPS, in Q1 2013 was -$0.21, and we got it close to breakeven in Q2 2014 at -$0.01. On all four key metrics, give great execution and trending in the right direction. The first one, of course, is revenue growth and accelerating those growth drivers. Accelerating the growth that you saw already happening in the second quarter of 2014. Jeff already talked about that.
This is mainly three growth drivers, and actually, there is a fourth one now. All this exciting stuff here, the question was already asked. This is not going to drive meaningful revenue in the short or medium term, but this is expected to drive very meaningful revenue in the mid to long term. In the short to medium term, it's all about those three growth drivers that Jeff already talked about it. The U.S. residential market, which is 75% of our revenue, that market is on fire. We have great market share there, and we need to continue to focus on that market, execute on that market, continue to slightly improve our market share, and that will fuel a lot of growth. The other two growth drivers are very important as well.
It's segment expansion, not only playing in the residential market, but expanding into the commercial market with our new commercial product offering, and eventually as well in the utility scale market and the microgrid markets. Thirdly, our continuous geographic expansion by growing share in the countries we already are outside of the U.S., and continue to enter other countries, new countries we have not entered yet. Just to put that in a different perspective here, I have this slide up here, where today, actually, we have 85% of our revenue is in residential, and 15% is outside of residential, mainly commercial, mainly small commercial. From a geographic point of view, 85% of our revenue is in the U.S., and 15% is outside of U.S., and that's why I have this 85/15 model. It's our internal target, and I'm willing to share that with the investment community here.
It's our target to drive that 85/15 to a 50/50, meaning 50% coming from resi and 50% coming from non-resi: commercial, utility scale, and microgrids. From a geographic point of view, having 50% of our revenue in the U.S. and 50% outside of U.S. It will probably take us three to five years to get to that 50/50, and all of that will depend on how hot and how strong the U.S. residential market continue to grow. It will be difficult to outgrow that market, but that will be our target here, making progress towards that 50/50 target model. That also will give us a more diversified revenue base and customer base, which is a very important element in our strategy here as well. That's in terms of revenue growth and revenue and driving those revenue growth accelerators.
Let's have a look at gross margin and gross margin expansion here. As I said before, we've done a great job at expanding the gross margin up to 33%. By the way, at 33%, I believe we are probably by far the highest gross margin inverter company. Most of our competitors have margins in the teens. We have margins in the low 30s, making good progress towards our long-term target model of 35%-40%. Despite the fact that a large part of our ongoing cost reductions, we continuously share with our customers. You can see here that in the first half of 2014 versus the first half of 2013, revenue per watt decreased by approximately 8.5%.
I say revenue per watt, that means revenue not only coming from the microinverters, but also includes revenue coming from accessories like our proprietary cabling system or the gateway, some services and Enlighten, and so on. We have 8.5% price or revenue per watt decrease in the first half, which is pretty consistent with what we have seen in the last couple of years. We have been able to outpace those revenue per watt decline with a more cost per watt decrease of approximately 15% here in the first half of 2014 versus the first half of 2013. Again, that 15% is pretty much representative of what we have been able to do the last couple of years. The big question I always get is 15%, that's great, that's good, but will you be able to continue to do so?
The answer there is yes, not necessarily 15%. I mean, somewhat in the ballpark there. The reason why we can continue to drive down the cost is summarized here. There's three key elements that we continue to work on in driving down the cost. First one is innovation. Innovation, innovation. We are really only in the first inning of a nine-inning innovation game. There is so much more that we can do, partially based, of course, on our architecture, semiconductor-based design and architecture, where we continue to integrate components into the semiconductor, reduce the printed circuit board, reduce the size of the microinverter, and make it cheaper. There's many other things as well in terms of innovation, material science, mechanical engineering, and so on. That's one big bucket, and there is still a lot that we can do.
Second, scale. We did 1.6 million units. This year, we are doing more than 2.5 million units. Eventually, we are going to do 5 or 10 million units per year. Of course, when you have those base volumes, you can really drive down the cost drastically. Last but not least, power. As higher power modules become available, we pair those higher power modules with higher power microinverters. First generation was 175. Second generation, 190. Third generation, 215. Fourth generation, 250. Now our fifth generation, 275. While the unit cost of each generation has continued to come down, that combined with higher power, the cost per watt comes down drastically. Of course, it takes time because it's not because those higher power panels become available, that everybody immediately jumps on those higher power panels.
It will take multiple quarters and years till those higher power panels get adopted in the market. That is for us a continuous tailwind for a gross margin improvement. By the way, we as a microinverter company, uniquely benefit from that. Central or string inverters don't benefit from that at all. It doesn't matter what the power per panel is for central or string inverters. We have a lot more here that will continue to drive down the cost in the next couple of quarters and years. Talking about leveraging the operating expense here for a second. You can see here, again, it's very important because as I said in our balanced profitable growth strategy, we can really try to accelerate the growth even more than we do so by throwing in a ton more resources at it.
We would never make a $1 of profit, and that's not the right thing to do. What we are doing here is driving growth, but at the same time being financially disciplined. In order to do so, we have an internal rule that I'm willing to share with you here, guys. The rule is that we allow our operating expense to grow at half the speed of revenue growth. Just as an example, if I pick a random number, if revenue grows at 40% year-over-year, we allow operating expense to grow at 20% year-over-year. It's not necessarily going to be like that every quarter, but it's more kind of like over a year-over-year trend and guardrail that we use to manage our operating expense.
To illustrate what that does in terms of leveraging the model, I start here with the Q2 results, which was $82 million quarter with 33% gross margin and 33% operating expense to revenue, resulting in a breakeven operating margin. If I just, for simplicity here, assume in my what I call the 2x model, 100% revenue growth from $82 million to $164 million, I assume my gross margin is at my long-term target of about 35%-40%, and I allow my operating expense to grow at 50%, which is half of the 100% revenue growth here, I actually come to an operating expense of 25% to revenue, resulting in an operating margin of 10%-15%, which is a good start. 10%-15% operating margin is a really good start. It's obviously not our end goal or target.
We stick to our long-term target model that we have introduced by now almost three years ago, where we will continue to drive the margins to the 35%-40% and where we want operating margins in the 15%-25%. These are world-class, high-tech operating margins, and that's what we aspire to be, and that's what we target, and we'll continue to drive so. Obviously, we need more than 2x revenue increase per quarter to hit those long-term target models. To wrap it up here, again, accelerated growth is all about driving EPS growth and EPS expansion. Again, to illustrate that here with my model, you can see that there is a tremendous opportunity here to go and drive EPS expansion. Again, starting from my Q2 2014 model with $82 million revenue, margins at 33%, operating margin at breakeven with $0.01 of negative EPS.
If I plug that in into my 2X model and assume that my revenue doubles to $164 million per quarter at 35%-40% gross margins and those 10%-15% operating margins, that drives $0.35-$0.50 of non-GAAP pre-tax EPS. Which when you annualize those EPS numbers, you get to $1.40-$2.00 of yearly EPS, which is a tremendous improvement from, of course, where we are today, and which clearly illustrates the earnings power that this company with this business model has. I said pre-tax EPS. Currently, we still have approximately $100 million of NOLs. First couple of years, we're not going to pay taxes. At a certain point, as we continue to be more profitable, we will have to start paying taxes.
Trust me, we will do everything as much as we can, and we'll address that at the appropriate moment. Again, the key takeaway, accelerated revenue growth here really drives EPS expansion. With that, I'm going to open up the floor here for questions. Pavel?
Thanks. On the doubling of revenue from $82 million to $164 million, you mentioned that you want to be more international over time and more commercial versus resi over time. Given that pricing is lower in the international market and probably in commercial as well, what kind of volume assumption does that revenue forecast assume?
I can't really express it in volume. The volume will have to more than double, obviously, because of this continuous price erosion. Maybe I'll answer your question differently. It is true, as Jeff also indicated, that expanding in some of those international markets, not all of them, but some of those international markets, as well as expanding in commercial and eventually utility scale, it will put some pressure in the headwind from a growth margin point of view. That is why we definitely will continue, as hard as we've done in the past, we will continue in the future to drive down the cost of our products. We spend a ton of money on R&D, and a lot of that goes through innovation, new product introduction as well.
We continue to spend a lot of money on R&D and technology involvement as well to go and continuously drive down the product cost. We will need that to be successful in those markets and to keep our margins making progress towards our long-term target model.
I also want to clarify that pricing in other European international markets, while a little bit lower, is not substantially lower. Okay? I don't want you guys thinking that there's a huge difference in terms of price point. It is very low single-digit difference of six. It's small. Okay? It's not huge.
Ruth.
Could you talk a little bit about the pricing of the battery system?
We haven't disclosed any pricing yet of the battery system, my intelligence tells me that competitive products are currently being priced anywhere from $1,500-$2,500 per kWh. We feel very comfortable to have a very competitive pricing compared to those price levels and still make some good margin on it.
As a follow-up, how do you think about what are you doing for warranty there? One of the pushbacks that I've seen from integrating module level with power electronics, and I'm sure you see this too, is that module players want to be responsible for a long lifetime of something they don't control. You are potentially subject to warranty issues of the battery chemistry, which you don't control. How are you going about that?
We have not fully disclosed and communicated what our commercial offering is going to be, and so warranty, we will communicate that later on. What I can say, and Raghu, feel free to jump in, is that the battery is a lifetime expectation, is expected to be more than 10 years.
I think a lot of that, so obviously, warranty is a kind of a business decision, effectively. Some warranty decision will get driven by incentive programs that are, as Jeff, for example, in California, has an incentive program that requires a certain warranty level. Again, it's really going to be at the time of launch, the product and business decision on what warranty we will offer.
Sorry, just another follow-up. Remind me. If the battery lifetime, so if there is something equivalent to the S270 in there.
Yes, the S270.
Okay, let's just assume that's 25 years.
Yeah.
The battery is 10 years. I assume that the system is easy enough to sort of retrofit batteries after 10 years.
Yeah, it's not field serviceable. Whether we do that, whether we retrofit it or not, those are all configuration decisions that get made.
Yeah.
Right now, it's literally two connectors that are going to connect the cell module and the power connector.
I'm trying to understand if you separate the battery piece. Like if that plastic housing comes off.
Very likely we'll offer one warranty for the whole AC Battery.
Chris, just to follow up, how does the target model change when you start going into storage device? How did it do with different margin impact than ?
Yeah. As I said before, we do not expect any meaningful revenue in 2014 or 2015. Revenue from storage will start ramping in 2016 and 2017. As I also said before, we do expect a decent, good margin on that product as well, given where we believe the cost point is and where we see pricing going into that market. From an operating expense point of view, we will have to manage those new initiatives within the guidelines that I have explained before. All incremental revenue that will come from those new initiatives at good, decent margins will mostly flow through to the bottom line and be accretive to the bottom line.
If I can talk a little bit about pricing again on that. Not unlike what we have done with our solar offering, we will price their solution. There are four elements that drive pricing or cost of a storage solution, it's the chemistry, it's the power electronics, it's the installation, it's actually a pretty big element of it's tough to compress, and integration. We are going to price to say, and we are affecting three of those four. Chemistry is going to be commoditized, we know that. The other thing we affect quite significantly, we're going to price to the value of the solution rather than bare box. That's how we're going to maximize our margins.
So-
Sorry, follow-up question.
On the box there, is that something that ELIIY is making or you're assembling it and who controls the IP and the patents around that?
There is a lot of patents. The heart of this and the highest patents in this business is this guy there. There is 275. This is where all the core IP is going to reside. The current planning is that the battery vendor is going to provide the cell module, which has the state of charge, temperature, all of the safety features for monitoring built into it. Most likely, we are going to rely on Flextronics. They are the best partner to assemble the pieces together, which is take the inverter, take the battery module, connect it, and put it into a box.
Chris, just wanted to follow up on the next-generation microinverter and the margin profile there. In the past, you guys have reduced the number of components, which definitely had a good margin for you. Can you talk to that and also the timeline for how that is going to be rolled out?
Right. First, on the timeline, the fifth generation microinverter in the U.S. will be available approximately mid-2015. Shortly after that, followed by availability for the European or outside of the U.S. version. From a margin point of view, again, we are managing the whole margin on a company-wide level and trying to make good progress towards our 35%, 40%. Obviously, taking the benefit of the higher power microinverter here will help us to drive down the cost per watt. It will help us to move, including the new microinverter, towards that long-term target goal.
Chris, as a follow-up to that question, when you introduced the Gen4, you guys talked about a year-long transition. As you introduce the Gen5, what kind of transition or when should we expect between the Gen4 and Gen5 in terms of components?
My expectation is that this transition from Gen4 to Gen5 will be even longer than the transition from Gen3 to Gen4. Can I answer one more question on pricing and value? While we continue to drive the cost of the microinverter down itself, we're also paying very close attention to the entire balance of system. Let me give you a few examples of that. This new gateway that you're seeing, the new Envoy-S, that'll be released at the same time as the S275, it integrates, for example, the revenue-grade meter. A lot of our partners have to install today a separate meter. Today, it will be integrated into that itself. Cellular connection is integrated into that. The whole thing fits into a box, an external box, and you'll see that when you go to the booth. AC in, AC out. Again, you're driving the system-level labor costs down.
Yet another example. That's on the ground. On the roof, for instance, when we released the M250, it's integrated ground, which means they no longer needed to run what's called a grounding electrode conductor and a copper wire through all the inverters. That had a huge savings in not just material costs, but labor costs as well. Finally, we just announced that the inverter, when paired with a specific racking vendor, I think it's Unirac, you don't need to ground the modules anymore. Panels don't need to be grounded anymore. We ground it through the inverter and through our cabling system. Again, another example of where we are continuing to drive the system-level cost down. We influence that quite a bit.
Maybe, Phil, let me answer there as well, that because we expect a very long transition spread over multiple years, and some of that will depend on when some of those advanced grid functions become mandatory in certain areas as well. As a result of that, we will continue to drive down the cost of our fourth generation. Again, leveraging those three same buckets, if you don't mind. We are still making some innovative changes to our fourth generation that helps drive down the cost of it. Of course, the volume ramp into the multi-millions on the fourth generation helps drive down the cost of the fourth generation, and we could still continue to see a shift from 215 to 250 in that fourth generation as well, which also helps to further improve the gross margins.
Can I ask you-
Oh, I'm sorry.
Go ahead.
In the Gen4, you guys have two versions of microinverters, actually, a 250 and a 250. Do you plan to have similar? Similar question on storage, do you plan to have a bigger size storage? Obviously, commercially, you might want more and you planned a bigger size storage.
Yeah. Again, it's what the market drives it. There are panels that today, if you look at it, there are panels that are down. Still people are installing 235, 240-watt panels, and the 215 is a better pairing for it. Now I just saw an LG panel, 60 cell, that's 310 watt. That are pairing it at 250 watts. Yes, we're going to do what the market needs. The nice thing about it is, do it in software. Again, I want to say that, reinforce what that software defined concept should mean. Similar to that here as well, from an application standpoint, we are looking at offering this both as a C over two, what's called 2 overcharge discharge rate, or a C over four charge discharge rate.
It's 1.2 kWh of energy, and it's either 275 W or two of them, which gives you 550 W, gives you C over two or C over four, and it's going to be very application dependent. In a commercial application where it's demand charge driven, most likely it'll be C over two than C over four.
Right. Maybe to add to that, for residential solution, this is it, and again, it's a modular approach. If you want six kWh, you need five of those boxes. Right? For commercial, we will, of course, think this through and think through the whole packaging, to offer more bigger solutions.
It's going to be a repackaging of the exact same concept. Most likely stay at 1.2 kWh. There's some certain advantages.
Just a quick follow-up question on commercial in terms of the new product you guys have launched, right. Not a lot of information on the presentation. Can you talk how that's coming along and what progress you guys have made and et cetera?
You mean just in terms of volume? The market right now, we're in beta and we're doing pilots. We expect to have full availability and volume in late December, early January. Reception that we've received and as we're talking with customers has been very positive. Very positive. Many customers are telling us they're just waiting for it. So far we're getting a lot of good momentum. Again, we're just starting that process, right? We're just starting that process.
Yeah. Again, Bala, just let me make one point on the C250. People have strings, right? They have a certain string, then they go to a combiner box and a string system and a central system. On average, they put anywhere from 14, now 1,000 volts, maybe putting 20 on a string before they go to a combiner box. You can put 48 units, 48 modules or 48 panels on a single 20 amp breaker for that C250.
That's an example where you have a higher inverter CapEx, it's offset quite a bit. It could be, we think, as much as $0.06-$0.07 a watt on lower BOS CapEx costs. Pretty substantial.
Yeah. I joke about this, when I was in Tesla, that fight was settled 125 years ago. AC three-phase is just always better.
Certainly, you guys have seen 30%-35% cost reduction on successive generation. Can you talk a little bit about your expectation on moving from gen 4 to gen 5 in terms of two things, one, where your full cost reduction is, and then what the component count looks like from gen 4 to gen 5?
I'm not sure we're divulging any specific numbers on what the cost reduction is. Again, the focus on that, we were talking about in terms of further integration. We do have our, I think it doesn't line up perfectly with the generation, but gen 6 ASIC in here, mixed signal ASIC, multi-million gate ASIC. Further integration has taken place. Parts count has continued to come down. Even though the complexity and the requirements have gone up, actually parts count has come down. I can't give you an exact number or quantify a specific number. I don't think we are releasing precise numbers on that sharing with the market. You'll see it on a cumulative system level basis, what our gross margin. Yes, we continue on that exact same path. Of course, this is another 10% higher in power.
On a dollar per watt basis, we get a disproportionate benefit as well.
You guys have kind of evolved the business model slightly, incrementally moving into some financing options now with the energy storage. Can you talk about potential to monetize some of the grid services? Obviously, voltage management along certain distribution lines can be very valuable for utilities in terms of managing generation. Where do you see in terms of real possibilities to monetize that or is that just something that stand-alone, building out deeper penetration with the solar?
It's absolutely something that we are thinking about. We are talking to utilities about for monetizing those kind of grid services. Fleets have been receptive, not in they are saying, "Hey, we give you $X for it," but in recognizing the value of what we're doing. We are actually in discussions with utilities. Typically the way they work is, they usually have a pretty long cycle. They start with pilots. That's what we're in the discussion with them, with about half a dozen utilities on entering and doing some pilots that do what's called distributed energy resource management services or DERMS. Yeah, there's real value in it and we just won't give away stuff just yet.
Two questions. I think you probably answered the first one already, but I missed it. On the commercial systems, how much lower do you anticipate the gross margin to be next year when you're shipping volume?
It's going to be below average. A lot of that commercial business as well is really deal by deal. For sure, for the larger deals, there's going to be some competitive pricing there, and it will be below average.
The managed portfolio, which you continue to see head towards the target.
Absolutely. Volume is our friend.
Second question. What's the maximum charge and discharge rate of that battery unit?
The cell module itself is capable of doing actually one C or more.
Okay.
In this particular case.
One inverters.
Oh, yeah, one or two inverters. It's either 275 or 550 give you C over two, C over four. That's the limiting variable. I don't really call it limiting. I think that's what the application needs is.
Is there any reason someone would want to have two inverters in there to have that?
If you want a C over two in some. Yeah. In some applications, I expect that to happen. For example, in demand charge, in some instances, I expect that to happen. Of course, our ambitions are to be utility scale as well, because I absolutely believe the same architecture applies for a two-megawatt hour storage solution. I think we may need for frequency reg, we'll need higher power.
Like the Matrix, but walls with absolute power from the universe.
Yeah, exactly. They did it there, so too.
That's right.
One more for you, Chris. If you guys achieve anything close to the 2X model, one of the corollaries is you'll have fair amount of free cash flow that the business will be throwing off. Can you talk about the priorities for how that would be allocated, if in fact that you materialize? Thanks.
Right. I think it's a great question, and I'm always pleased when I get this because I think six months ago or 12 months ago, I got a lot of questions like, "You guys are going to run out of cash. How are you going to survive?" Right? How quickly it now changes to, "Wow, you guys making a lot of cash. What you going to do with the cash?" Right? Yes, keep in mind, we have a very low CapEx model because most of the manufacturing is all outsourced. We do spend some CapEx dollars there on some proprietary test equipment and in some cases, some factory automation and so on. The combination of driving healthy profits with a low CapEx model results in some really strong cash generation. Last quarter, we ended with $38 million of cash on the bank.
We still have $6 million of historical debt that will be repaid over the next couple of years. In addition to that, we have access to a $50 million working capital facility that has not been used. Yes, I feel really good about my balance sheet. I feel really good about the cash generation capability of this business model. Having said that, I don't really have an answer about what I'm going to do with all the cash.
Big party.
On the storage front, can you guys speak a little bit to the strategic decision to pursue something in residential as opposed to maybe commercial where you might have utilities and charges that drive a little bit better economics on storage front?
Actually, we're going to do both residential and commercial. I think, if you look at it from, clearly, there's a business model that exists for commercial and demand charge. I think the more valuable business model there is a combination of solar and storage. Absolutely, we are thinking commercial is absolutely a valuable place. Plus, outside of the U.S. as well, places like U.K., places like Australia. The really nice thing about this architecture is it's literally swapping out this guy. Since you're already making an S275 for the non-U.S. market, it will be, not to trivialize it, a matter of putting this in and getting it certified. No, we are thinking about commercial, we are thinking about outside U.S. as well. Same architecture, though, that's the key, it's going to be distributed network architecture.
Any more questions? Well, with that, I thank everybody for participating in our first mini Analyst Day. I think it was a great discussion and great questions. We're going to wrap it up here. We're going to now walk over to the booth at SPI. It's number 812. Whatever the number is, once you walk in, you will see our nice, beautiful booth there. By now, I think it's probably already busy there with a lot of other people. Raghu, Jan, and myself will be there. I hope Paul will catch up with us as well. There will be plenty of other Enphase people out there, so feel free to grab them and ask your questions. Thank you again, and have a nice day in Vegas.