Good afternoon, everyone, and welcome to First Solar's third quarter 2018 earnings call. This call is being webcast live on the Investors section of First Solar's website at investor.firstsolar.com. At this time, all participants are in a listen-only mode. As a reminder, today's call is being recorded. I would now like to turn the conference over to Steve Haymore from First Solar Investor Relations. Mr. Haymore, you may begin.
Thank you, Holly. Good afternoon, everyone, and thank you for joining us. Today, the company issued a press release announcing its third quarter financial results. A copy of the press release and associated presentation are available on First Solar's website at investor.firstsolar.com. With me today are Mark Widmar, Chief Executive Officer, and Alex Bradley, Chief Financial Officer. Mark will begin by providing a business and technology update. Alex will discuss our financial results for the quarter and provide updated guidance for 2018. Following their remarks, we will have time for questions. Please note this call will include forward-looking statements that involve risks and uncertainties that could cause actual results to differ materially from management's current expectations. We encourage you to review the safe harbor statements contained in today's press release and presentation for a more complete description. It's now my pleasure to introduce Mark Widmar, Chief Executive Officer.
Mark?
Thanks, Steve. Good afternoon, and thank you for joining us today. Our financial results for the third quarter were solid, with net sales of $676 million and earnings of $0.54 per share, driven by the sale of development projects. From an operations standpoint, we have started the first commercial shipments of Series 6 from our factory in Vietnam, and progress to date on the initial ramp has been good. Commercially, we continue to be very pleased with the strong demand for our technology, as demonstrated by our net bookings of 1.1 gigawatts since our prior call. It is important to note we have booked over 1.6 gigawatts since the May 31st solar policy change in China.
Before delving into the specifics of the most recent bookings, I think it's important to highlight some of the trends that we are seeing which support the near-term and long-term growth of utility-scale solar globally. As has been the case for some time, the low cost of solar power continues to be the primary driver of demand. Beginning with the U.S., solar procurement from both utilities and corporate customers is strong and growing. According to a leading third-party market research firm, 8.5 gigawatts of utility-scale solar was procured in the first six months of 2018 alone. Looking forward, we expect this procurement trend will continue to be robust.
Based on our tracking of utility-integrated resource plans, as well as other public announcements, we expect utilities outside of California to procure more than 15 gigawatts of solar in the coming three years, a number that has increased by several gigawatts over the past year. Much of the growth is coming from regions such as the Midwest and the Mid-Atlantic, which are still in the early stage of utility-scale solar adoption. Several announcements over the course of this year highlight the trend of U.S. utility-scale solar growth. For example, AEP announced a plan to add more than three gigawatts of solar as it targets a 60% reduction in CO2 emissions. Similarly, in its 2018 integrated resource plan, Consumers Energy in Michigan proposed five gigawatts of solar and a plan to retire all coal generation by 2040.
Most recently, NIPSCO, the second-largest utility in Indiana, decided to retire all of its remaining coal power plants, in some cases as much as 15 years earlier than previously expected, and replace them with over one gigawatt of solar. NIPSCO cited the low cost of renewable energy as a key factor in this decision. Notably, all three of these utilities are in regions that have not historically been associated with solar, but where the low cost of solar power now provides a compelling economic alternative to thermal generation. When you factor in additional procurement from California, which has passed a mandate for more than 60% of renewable power by 2030 and has decided to close its last nuclear power plant by 2025, the potential for solar growth in the U.S. over the next several years is even greater.
It is important to note the nuclear power plant closure was not made solely on the basis of solar's favorable economics. The limited flexibility of the nuclear power plant to adjust to market price signals relative to solar was another key factor in the decision. Within the renewable sector itself, another factor supporting the growth of utility-scale solar in the U.S. is its increasing competitiveness relative to wind. As the production tax credit continues to step down, and with the recent IRS guidance establishing ITC safe harbor requirements, it is expected that solar deployments will outpace wind in the U.S. through 2023. The transition is expected to be particularly noticeable among corporate buyers of renewable energy, who in recent years have tended to procure more wind than solar.
With multiple gigawatts in our development portfolio, we expect to be able to take advantage of the recent IRS commence construction guidelines and enhance the value of these development opportunities. Outside of the U.S., economics also continue to drive solar growth in many markets. For example, in India, the LCOE of solar power is around $20 a megawatt hour less than coal. Similar economic benefits, as well as growing concerns about carbon emissions, have led to a resurgence in solar demand across many parts of Europe. Last year, Spain awarded four gigawatts through a tender process, and many more gigawatts are being planned in Europe, in some cases on a merchant basis as a result of the low cost of solar energy. In France, EDF recently announced its solar power plan, with intentions to develop and build 30 gigawatts of solar from 2020 to 2035.
Another factor that may have a significant impact on solar procurement in coming years is the potential for increased electrification of transportation. While it is uncertain how quickly the transition to EVs will occur, momentum is building, and there is potential for significant solar deployments driven by this shift. With this context around what we're seeing in the macro environment, I'll now turn to slide four to discuss new bookings since our last earnings call. In total, our net bookings since the prior earnings call were 1.1 gigawatts, which brings our total year-to-date net bookings to 5.2 gigawatts. Continuing the trend in the first half of the year, systems bookings were especially strong this quarter, with more than 350 megawatts of new development projects.
After accounting for shipments of approximately 700 megawatts during the third quarter, our future expected shipments, which stretch into 2021, are now 11.3 gigawatts. We are largely sold out to the end of 2020 when taking into account our new bookings in combination with opportunities that are signed but not yet counted as bookings. Our systems bookings are comprised of two PPAs that we signed with leading utilities in the U.S. The first of these projects was a 50-megawatt AC project that we will construct for PacifiCorp in order to provide affordable solar power to a Facebook data center in Oregon. We're excited about the opportunity to partner with PacifiCorp and contribute to powering Facebook's operations with 100% renewable energy. The project is expected to achieve COD in 2020.
The second PPA for the quarter was a 227-megawatt DC agreement signed with a major utility in the southeastern United States. This project is intended to support a collaboration between the utility and its corporate buyer to meet renewable energy objectives. It is a tremendous opportunity to be part of this project, which is expected to reach COD in 2021. Additional details will be available in the future. Both of these projects are prime examples of our differentiated capabilities that enable us to address the renewable energy goals of corporate buyers in partnership with utilities by leveraging efficient and reliable large-scale off-site generation. Since our first C&I PPA with Apple, we have now contracted nearly 1 gigawatt DC with corporate buyers to support their renewable energy goals.
With the increasing number of companies joining the RE100, we expect C&I demand will continue to grow, and we believe that we are strongly positioned to serve the needs of this segment. In addition to these project development bookings, excuse me. We also signed an EPC agreement with Tampa Electric to construct a 50-megawatt AC project in Florida. This marks the fifth EPC agreement that we have signed with Tampa Electric, and we continue to look for ways that we can partner together. Note this is not included in an incremental booking as we have signed the module agreement for this project last year. The remainder of our bookings were module sales in the U.S. and various international markets. New international bookings of nearly 300 megawatts are particularly notable in light of the challenging price environment in certain markets.
In contracting this volume, we remain disciplined with regard to pricing and continue to focus on our points of differentiation. This differentiation includes focusing on regions where our technology has an energy advantage, deep customer relationships, and opportunities that favor our leading eco-efficient technology. As we move forward, we will continue to leverage these strengths to capture the best opportunities across all of our markets. Lastly, before moving on to other topics, I'll briefly highlight the progress we are making in our O&M business. Year to date, we have booked 2.3 gigawatts of new projects, bringing the total fleet under contract to nearly 11 gigawatts. With nearly 80% of this year's bookings coming from projects where we are not the developer, we continue to capture additional value with our O&M offerings.
The single largest booking this year is a 530-megawatt agreement with Tampa Electric to provide O&M services across several project sites. We began initially with Tampa Electric as a module agreement arrangement. It's grown into the scope to include EPC services and now O&M services. Continuing on to slide five. I'll next discuss our mid to late-stage bookings opportunities, which now total 7.9 gigawatts DC, a decrease of approximately 400 megawatts from Q2 due to the strong bookings reported today. On a geographical basis, North America remains the region with the largest number of opportunities at 6.7 gigawatts DC. However, we also continue to have a significant number of opportunities in both Europe and the Asia Pacific region.
Similar to prior quarters, the total potential opportunities includes deals that are signed but not yet counted as bookings until certain conditions precedent are closed. The total is now more than 550 megawatts, a slight decrease from the prior quarter due to the projects that were booked. Included in this total is 150-megawatt DC PPA opportunity in the western U.S. In addition, today, which is not reflected on the slide, we signed another PPA that is over 140 megawatts DC in size, which when included, brings this total to approximately 700 megawatts. Our Q3 systems bookings of approximately 450 megawatts DC, we now have 2.5 gigawatts DC of future contracted systems shipments.
Combined with the 2.6 gigawatts of mid to late-stage opportunities, which now includes 300 megawatts DC of projects that are signed but subject to CPs, we feel we have good line of sight to our one gigawatt yearly systems goal. Continuing on to slide six, I'll next provide an update on our Series 6 production ramp. The most significant development to highlight from the past quarter was the start of our production at our first Vietnam factory in September. With this milestone, we now have three facilities in three separate locations producing Series 6 modules. Commercial shipments began from this location earlier this month, and the factory is demonstrating throughput levels at 35% of full capability in only two months since the start of production. The factory ramp was accelerated relative to Ohio and Malaysia by applying prior learnings, including starting production with an improved module framing tool.
With each successive factory ramp, we expect the length of time needed to reach full production levels will be progressively shorter as we implement best known methods captured from the prior factory ramps. Our second factory in Vietnam is also making good progress towards the start of production. Construction on the facility is now complete and more than 80% of the tools have been delivered. We originally anticipated that the factory would start production in the second quarter of 2019, but based on the tremendous effort of our manufacturing and engineering teams, we now expect production will commence in the middle of Q1 2019. Finally, construction of our second U.S. factory in Ohio remains on track. Relative to our Ohio and Malaysia Series 6 factories, since our last earnings call, we have made good progress ramping each of these plants.
Our Ohio factory is now demonstrating throughput levels near 90% of its full capability as compared to 60% at the time of our last call. Our Malaysian factory is demonstrating throughput levels at 75% of full throughput capability versus 40% previously. In Ohio, we are finishing the installation of the last inventory accumulator to adequately buffer the line. In Malaysia, we are just starting the installation of the accumulators and expect completion by year-end. We also continue to make good progress upgrading the tool set, primarily focused on improving availability for both factories. For example, in Ohio, we have completed approximately 70% of the tool upgrades identified. The combination of the accumulators and tool set upgrades are expected to enable both of these factories to demonstrate full throughput capabilities by the end of the year.
Module wattage in our lead factory in Ohio continues to improve steadily, with approximately 50% of recent production at 420 watts per module or higher when running at full process entitlement. This compares to wattage of around 415 watts at the time of our Q2 earnings call. We are now seeing top bins reaching 430 watts per module versus 425 as of the last call. Module wattage for our Malaysian and Vietnam factories is lower than Ohio as we are still in the process of matching the efficiency between the factories. Note, during the initial factory ramp, we do not run our anti-reflective coated product. We ramp our non-ARC product first to focus on throughput and performance. A key driver of matching efficiency to our lead factory will be the transition of our anti-reflective coated product.
In addition, because these factories are still in the early ramp phases, bin distributions are much wider than our lead factory. Over time, as we match the process between the factories, we expect the fleet average efficiency to increase and bin distribution to narrow. Lastly, before I hand the call off to Alex, I'd like to briefly highlight some of the remarkable work that First Solar, as a thought leader in the industry, is helping to facilitate related to grid flexible solar. At our Analyst Day last December, we introduced the concept of Solar 2.0, which moves beyond the traditional view of solar as an energy-only contract to valuing solar as a flexible and dispatchable resource. As we highlighted at the time, flexible solar can enable solar penetration rates without the need to add storage.
As part of our comprehensive engagement with stakeholders on this topic, we recently sponsored a study conducted by E3, which simulated the impact of flexible solar on an actual Florida utility system. The results of the study are impressive and indicate that operating solar flexibly as a scheduled resource provides significant additional value to the utility in the form of expected reduced fuel and maintenance costs for conventional generation, reduced curtailment of solar output, and reduced air emissions. The study simulated utility-scale solar deployments at a level up to 28% annual solar energy penetration, and the benefits increased as the level of solar penetration grew. While we'll discuss more about this topic on future calls, we invite you to view our recent press release on the topic and visit the E3 website to access the study.
I'll now turn the call over to Alex, who will provide more details on our third quarter financial results and discuss updated guidance for 2018.
Thanks, Mark. Before turning to our financial results for the quarter, I'd like to highlight that we'll be hosting a call in late Q4 to discuss our outlook and financial guidance for 2019. A press release with the date and details of the event will be issued approximately two weeks in advance of the call. As Mark mentioned, we had solid third quarter financial results driven by the sale of several key development projects. I'll provide some more context beginning on slide eight. Third quarter net sales were $676 million, an increase of $367 million compared to the previous quarter. The higher net sales were primarily a result of closing the sale of the Willow Springs project in the U.S., the Manildra project in Australia, and selling some smaller Japan assets. Note that each of these projects achieved initial revenue recognition in Q3 based on the respective project percentage of completion.
In addition, we recognized higher revenue from the California Flats Project as compared to the prior quarter. Systems revenue as a percentage of total quarterly net sales increased to 82% in Q3 versus 66% in Q2 as a result of the project sales mentioned. Third quarter gross margin improved to 19% as compared to a -3% in the prior quarter. The improvement was due to the mix of higher gross profit projects recognized and a $25 million reduction to our module collection and recycling or EOL liability, partially offset by higher third quarter Series 6 ramp charges of $48 million. Our systems segment margin was 24% in the third quarter, and the module segment margin was a -5%.
As it relates to the module segment gross margin, bear in mind that sales were still comprised entirely of Series 4 volume, as early Series 6 volume was allocated entirely to our systems business. However, the module segment COGS is comprised of both Series 4 COGS and Series 6 ramp-related costs, as these are allocated to the module segment. The net reduction to the module segment gross margin from ramp-related costs, partially offset by the EOL adjustment, was $23 million. Third quarter operating expenses were $71 million, a decrease of $25 million compared to Q2. Plant startup expenses decreased by $10 million as a result of lower Series 6 pre-production activities in Malaysia, partially offset by increases for our initial Vietnam factory. In part, the lower startup is due to accumulated learnings from the prior two Series 6 factory startups, which we've been able to apply to Vietnam.
SG&A was lower versus the prior quarter, primarily due to a benefit from the reduction to our module collection and recycling liability and lower variable compensation. Q3 operating income was $59 million compared to an operating loss of $104 million in the second quarter. The quarter-over-quarter improvement in operating income was primarily due to higher net sales, improved gross margin, and lower operating expenses. Income tax expense was $2 million in Q3 as compared to a tax benefit of $6 million in Q2. As it relates to U.S. tax reform enacted last December, we have not recorded any adjustment through Q3 related to our original estimates. We expect to finalize our accounting related to tax reform in Q4 based upon finalization of currently proposed tax regulations and the filing of our federal and state income tax returns.
The combination of the above-mentioned items resulted in earnings of $0.54 per share in Q3 compared to a net loss in the second quarter of $0.46 per share. Moving to slide nine. I'll next discuss select balance sheet items and summary cash flow information. Our cash and marketable securities balance ended the quarter at $2.7 billion, a decrease of $405 million from the prior quarter, primarily as a result of capital expenditures to support our ongoing Series 6 capacity expansion and the timing of receipts from system projects. Third quarter net working capital, which includes the change in non-current project assets and excludes cash and marketable securities, increased by approximately $215 million. The change was primarily due to an increase in unbilled accounts receivable. Total debt at the end of the third quarter was $466 million, an increase of $10 million from the prior quarter.
The increase is primarily associated with incurring project-level debt in Japan and Australia, partially offset by debt assumed by the purchaser of the Manildra project. As a reminder, essentially all of our outstanding debt is project-related and will come off our balance sheet when the projects are sold. Cash used in operations was $225 million, primarily due to the timing of receipts from systems projects. Keep in mind that when we sell an asset with project-level debt that is assumed by the buyer, the operating cash flow associated with the sale is less than if the buyer had not assumed the debt. In Q3, buyers of our projects assumed $56 million of liabilities related to these transactions. In year-to-date, that total is $116 million. Capital expenditures were $238 million in the third quarter compared to $195 million in the prior quarter.
The cumulative spend on Series 6 capacity now exceeds $1 billion out of a total expected spend of around $1.9 billion, so 6.6 gigawatts of capacity. Lastly, depreciation and amortization expense was $34 million in Q3 versus $30 million last quarter. Turning to slide 10 on next to our updated 2018 guidance. Before covering the detailed updates to our guidance range, there are some key business updates to discuss. Firstly, as we highlighted on last quarter's call, there was the potential for our guidance to be lowered based on the sale timing of our Ishikawa project in Japan. Despite the weather-related issue experienced early this year, all modules have been installed and construction on the project is nearing completion. However, based on the timing of the sale process, we now expect to complete the sale of the project in 2019.
While this does impact our 2018 outlook, the sale in 2019 allows us to optimize the transaction and realize the full expected value of the project rather than potentially sacrificing project value to ensure a closing this year. The impact of the delay in the sale of the Ishikawa project is expected to be partially offset by the sale of two other smaller Japan projects. Selling these projects in Q4 has been part of our opportunity portfolio, and with the revised timing of the Ishikawa sale in 2019, it now makes sense to aim to complete the sale process for these projects in 2018.
Secondly, we expect a reduction in our module sales for the year as a result of some changes in shipment timing, as well as a reduction in certain international shipments. Thirdly, as we progress through the year, we've determined that while the cost for our initial startup activities for Series 6 is lower than originally expected, this savings is expected to be more than offset by higher ramp costs. As a result of the distribution of Series 6 related expense, we expect cost of sales to be higher due to the increase in ramp costs, while operating expenses will be lower due to reduced startup. With these in mind, we're revising our 2018 outlook as follows.
Starting with net sales, we're lowering the range to a revised forecast of $2.3 billion-$2.4 billion in order to reflect lower module sales and a revised timing of the Ishikawa project sale, which is expected to be partially offset by other Japan projects mentioned. Our expected gross margin has been lowered by approximately 200 basis points to a revised range of 18.5%-19.5%. The reduction accounts for the increase in ramp and related costs from $60 million-$100 million, lower margin from module sales, and the change in mix of systems projects to be sold. The operating expense forecast has been lowered by $45 million to a revised range of $345 million-$355 million. Plant startup expense is decreasing by $30 million to $90 million for the full year. The remaining reduction in OpEx is primarily due to careful management of core operating expenses.
Our outlook for operating income has been revised down by $40 million at the midpoint to a new range of $90 million to $110 million as a result of the lower revenue and gross margin, partially offset by the reduction in operating expenses. Below operating income, the most significant update is the forecast full-year tax expense, which is now expected to be approximately $15 million. The decrease from our prior expectation of approximately $35 million is a result of reduced operating income, as well as a change in the jurisdictional mix of income. Our guidance continues to assume minimal additional equity and earnings for the balance of the year. Putting these revisions together, our earnings per share guidance is now $1.40 to $1.60. The operating cash flow range has been lowered by $200 million as a result of the revised timing of project accounts receivable collections and lower module sales.
With some projects receiving Series 6 modules later than initially planned, this has concentrated work later in the year and has resulted in some cash collection timing moving from 2018 into early 2019. Capital expenses are unchanged at $800 million to $900 million for the full year. As a result of the decrease in operating cash flow, we're lowering our net cash guidance by $200 million to a range of $2 billion to $2.2 billion. Lastly, we're lowering our shipment guidance for the year by 200 MW to a revised range of 2.6 GW to 2.7 GW, that change is due to the lower module sales mentioned previously. Finally, turning to slide 11, I'll summarize the key messages from our call today. Firstly, we had very good execution and solid financial results in the third quarter as we closed several project sales and managed our OpEx effectively.
As we've seen over the course of the year, the timing of project sales can produce uneven quarterly results. As we move forward, this trend is likely to continue, particularly as it relates to international project sales, which may be sold nearer to completion of construction. Secondly, as we continue to make progress with our Series 6 manufacturing ramp, we now have a third factory shipping Series 6 modules and demonstrated throughput continues to improve across all sites. We've also accelerated the start of production of our fourth Series 6 factory. Lastly, the strength of demand for our Series 6 product remains a highlight. With bookings of 1.1 GW since our previous call and total future contracted shipments at 11.3 GW, we have strong visibility to future demand.
In particular, with over 350 MW of PPA awards included in our new bookings, more than 1.6 GW of systems bookings year-to-date, we continue to make excellent progress in building a systems portfolio that we expect to average approximately 1 GW per year over the next few years. With that, we conclude our prepared remarks and open the call for questions. Operator?
Thank you. Ladies and gentlemen, if you'd like to ask a question, please signal by pressing *1 on your telephone keypad. If you're using a speakerphone, please make sure your mute function is turned off to allow your signal to reach our equipment. Again, that is *1 to ask a question. Our first question today will come from Philip Shen with Roth Capital Partners.
Hey, guys. Thanks for the questions. Since the end of May, module pricing continues to fall lower. We're getting to very low levels now, and I know your cost structure can compete with it. Wanted to ask a few questions around this topic. Is the current environment putting any economic pressure at all on your 2019 or 2020 bookings for either Series 6 or Series 4? I know your contracts are legally binding, but what kind of pressure, if any, are you receiving from your customers? We hear the pressure may only be on Series 4, which is a much smaller % of your overall bookings. Additionally, how much of your current bookings are for Series 4, and how much Series 4 capacity is left to book? Finally, would you contemplate winding down Series 4 capacity earlier than expected?
I recall back in December, you guys extended that capacity, but would you consider ramping that down earlier than expected and then in turn accelerating the Series 6 expansion there? Thanks.
All right. Phil, there's quite a few there, so hopefully I'll get them all. Let's start with the economics on the 1.1 gigawatts, which 700 or so megawatts was module, and the 350 or so that we highlighted was systems business. When you look at the module business-- And again, of that 700, about 300 of it was outside of the U.S. Slightly more than half of that volume was actually Series 4, of that 700. There was a big chunk of Series 4 in there.
Most of the international number that we recited was international, and that was a Series 6. We tried to highlight in the script, in the prepared remarks, that we have the very good luxury of focusing on where we can capture the highest value, where we have point differentiation, where we can capture the energy yield that we have in certain markets, where we can capture the eco-efficient or eco-friendly attributes of our module and focus on those markets.
When you look at that and you look at the economics and what we saw for both Series 4 new bookings and Series 6 new bookings, you got to remember, the Series 6, in particular, bookings are starting to go out further into because we're sold out through 2020, so we're starting to see more latter half of 2020 volume bookings that we saw this quarter. We saw a nominal reduction to the ASPs in this quarter. Our sales team has done a fabulous job of positioning the product, capturing the best value from the customer. I'm talking nominal being $0.001 deltas on ASPs. From what we booked last quarter versus what we booked this quarter on both Series 4 and Series 6. I'm extremely happy with that. The team has done a great job of getting the best value for that product.
The way I would look at that is where we're pricing Series 4, even with the new bookings, will be mid-teens type of margins, which I'm happy with. I'm getting about If I look at my last quarter bookings, I got about a 10% premium of Series 6 over my Series 4 price. That's obviously very positive. From that standpoint, the volume was great and the actual underlying ASP and the economics relative to the headlines. I understand what's going on in the market, and I hear it as well sometimes from our customers. I understand pricing in some markets is very aggressive. We're not seeing anything today. If we do, we're not going to book it because I don't need to. We're booking good economics with good customers, and we have great relationships with. That's the new bookings aspect.
As it relates to the contracts, as I said before, the contracts were negotiated with customers with a risk-sharing approach and creating binding obligations between both parties to perform. To the extent the parties do not perform, then there's implications around that to that party. We have LDs if we don't deliver per requirements. Our customers will have costs associated with termination of and forfeiture of deposits and those types of things. As it relates to Series 6, I haven't had one customer come to me in the United States and have any discussion relative to that. We haven't seen that. The customer understands if they want to do that, then it's very clear what the implications are. We'll enforce our rights underneath the contract.
We had a couple of examples internationally that we, for example, in a couple markets that have turned soft, we had one example with a customer that we were selling modules into China. Given what's happened with the policy changes in China, they didn't move forward with the contract, and we took their 20% security. Again, we will enforce the rights underneath our contracts. There was a small number of megawatts that we had in India because of what's happened with the new tariffs. I'm talking small megawatts. It was less than 10 that we shipped the vast majority of the contract, and there was a small residual amount that didn't ship based off of the timing of the cutoffs of when the tariffs would be applied. We didn't fulfill that contract, the 10 megawatts, that customer.
We've had a little bit of noise because of what's happened with economic turmoil in Turkey. That customer is going to redeploy those modules to opportunities outside of Turkey. That's the backdrop of what we've seen relative to our contracts. As it relates to Series 4 with the current bookings that we just had, we're sold out of Series 4. We've got more than enough demand around Series 4. There's no more Series 4 volume that we need to worry about contracting. Now, I said this before, as it relates to some of our contracts, we have optionality of delivering Series 4 versus Series 6 for some of our customers. We need to focus on what creates the best position of strength for solar in 2021. A higher mix of Series 6 is going to be better for me.
Scale is important, drives contribution margin, I want as much capacity of Series 6 in 2021. Now, it may mean that I may negotiate with some customers to move some volume that's contracted in Series 4 over into Series 6 and potentially push some of that volume into 2021 in order to enable me to scale and to capture the highest Series 6 production profile that I can in 2021. Those discussions could be happening with customers, and we'll continue to look at that. As it relates to our decision around that, we may provide a little bit more color on our guidance call in December, but some of those conversations we are having with our customers in the background.
Our next question comes from Sophie Karp with Guggenheim Securities.
Please go ahead.
Hi. Thank you for taking my question. I was wondering as far as systems demand in the U.S., are there any particular areas, states in particular that you see that are stronger than others? I know some of your competitors have been highlighting Texas and some other Southeast states. I'm kind of curious where you see demand coming from in the next few years. Thank you.
I think for us in particular, the Southeast is a really strong market. As we highlighted, one of the larger PPAs that we announced as part of our booking this quarter was with the utility in the Southeastern U.S. Last quarter, we also announced we had an acquisition of a portfolio of projects in the Southeast, which also came with a PPA, and we're real happy with having more development sites in the South Carolina region in particular. We're doing a lot with TECO, you saw that. The Southeast region is a very strong region for us, across Texas and then obviously into the Southwest, continues to be good markets.
The one thing I would say, and we kind of highlighted this, especially as you look across the horizon, to your question, over the next couple of years, solar is only going to be increasingly more and more competitive through 2023. If you look at some analysts' reports, especially when you go beyond 2019 into 2020, over the next several years, you're going to see the vast majority of utility-scale renewables being solar. I'll use an analogy that one of my customers used, who does both solar and wind. Their view over that horizon is if you look at the map of the U.S. today and red being solar and blue being wind, there'll be a convergence of red over the blue.
The red will continue to grow around most of the U.S., except for maybe the Midwestern states where there's a very strong central states where there's a very strong wind resource. As you look at it over the next several years, you're going to start seeing competitiveness of solar across many different geographies. I was talking with a customer recently, and even in the Northeast, where wind is a good resource, the economics are penciling out better for solar right now. I think we're strong in a couple of traditional markets today, but that's only going to expand and grow, and we highlight what we're seeing with AEP and what's going on in Michigan and Indiana.
Across Michigan, Ohio, and Indiana, those are all states that are in early innings, and we're seeing a tremendous amount of opportunity for solar deployments over the next several years.
Our next question will come from Paul Coster with J.P. Morgan.
Yeah. Thanks for taking my question, Mark. It's a little difficult at the moment to get a handle on what the normalized gross margins are and what your earnings power through the cycle is here. It looks like the cycle itself is somewhat kind of attenuated by this visibility you have. I'm hoping that you're getting to the point now where you're able to sort of forecast the earnings kind of power within the band for a couple of years at least. Is that starting to come into focus, and when can we understand what your gross margin kind of structure is?
Yeah, Paul, certainly internally, we obviously forecast it out. I don't see us providing guidance there multiple years. We'll provide guidance for 2019 specifically later this year. If you go back to the Analyst Day that we had in December of last year, we gave an outlook at that point, trying to break down the various components of the value chain that we have through the module development piece, through EPC and through O&M, and we gave indicative gross margin numbers around that. We said at the time that that was being used for guidance for the year, but you could think about that as an indicative view of how we looked at the business in the long term. You're right that the current visibility we have over the next few years, based on the contracted pipeline, is very helpful for us.
I think the message we gave in the Analyst Day still holds, which is those are good indicative ranges to use when you think about gross margin across the various segments. We've talked about the amount of the development business that we look to have around that gigawatt a year, a mix of self-developed project assets plus EPC business. You can use that. Then internally, clearly, as Mark said, we are focusing on putting ourselves in the strongest position we can for 2021, such that when we're through this period of current contracted backlog, hopefully fully over to Series 6, to our most advantaged product at that point, we still maintain a competitive position, even in a market where we will naturally see ASPs come down, and we're seeing that competitive environment be stronger at the moment.
For the long-term view, I would still guide you back to what we talked about last year, and obviously there's considerable strength over the next couple of years based on the contracted pipeline we have today.
Yeah, the only thing I'll just add to that is that we do try to hopefully get people to continue to think about not only at the gross margin level, but what is the op margin expansion that we can see as we see higher contribution margin coming through from growth, right, as we continue to expand the platform. One of the challenges we have right now is we're only producing a little bit less than three gigawatts when you combine the Series 4 and the Series 6. Series 4 is two-thirds of that number, and that's obviously not our most advantaged product.
Just as that mix shifts from all the Series 6 and then we start growing from three into four to five and six into seven, and with a relatively flat OpEx profile, there's an opportunity for a meaningful op margin expansion, which I think everyone needs to take into consideration.
Our next question will come from Brian Lee with Goldman Sachs.
Hey, guys. Thanks for taking the questions. I had two of them. I guess, first off, last quarter, if I recall correctly, you called out lower margins due to some higher ramp costs on the back end of Series 6. Wondering what's incremental here in 3Q, since it sounds like you're calling that out again as one reason for the outlook change here. Then I had a follow-up.
I think on the ramp, I think actually last quarter, we were at $60 million for the full year, that was the number for the prior quarter. I think we started the year right around $60 million or so of ramp. We hadn't changed the ramp last quarter, so the guidance didn't reflect any change to the ramp. What happened this quarter, partly because we ended up starting our first Vietnam factory, we'll start our second Vietnam factory sooner than we had anticipated, is that the profile has shifted from start-up into ramp. Start-up's come down $30 million. The full year number now for ramp is $100 million. It went from $60 million to $100 million. There's about a $40 million increase in ramp. 30 of that is just the geography shift between start-up into ramp. There is about a $10 million of incremental ramp cost.
Some of that is partly, we're still working in both Malaysia and Ohio. Our framing cell is basically a manual back-end process. The throughput through there is still not where we want it to be. As a result, we also had to hire some incremental labor in order to deal with the revisions that we've made to the frame. There's some of that cost that's in there. The way I would look at it, Brian, outlook to outlook ramp, what we guided to last call was $60 million. That was consistent with Q2, excuse me, Q1, and now it's going from 60 to 100, but $30 million of it's a geography shift, $10 million of it relates to some incremental ramp that we are seeing mainly associated with Malaysia and Ohio.
Okay. No, that's helpful color . I appreciate that. Then just my second question was around free cash flow. I know there's a bunch of moving pieces here, the end result this quarter was pretty negative. Then if I look back a little bit further, at the overall cash flow profile since your Analyst Day in December, there's been like a $350 million-$400 million swing in free cash flow to the negative, as cash flow from ops is down and CapEx is up. Wondering how you're thinking about the profile into 2019 and if you're confident that free cash flow can get back to positive next year or if that's maybe still too aggressive of you right now to assume, just given the recent trajectory. Thank you.
Yeah. Brian, I'm not going to comment on 2019, obviously we'll give you numbers later in the year. I think when it comes to the op cash piece, there's a lot of noise based on how we've been selling assets, I wouldn't focus particularly on that. What you're seeing a lot on the free cash and on the cash position overall is a significant decrease at the moment, just based on timing. As we've been shipping Series 6 product later to certain sites, that means under the EPC agreements, we hit milestones later, then with the billing cycle, we're receiving cash later. You've got a dip in the quarter here, which may actually go through the end of the year as well, just based on the timing of construction.
We're actually seeing receipts from some of the larger projects we have potentially spilling out over the end of the year into 2019. You're going to see a reflection of that in terms of the unbilled amounts on the balance sheet as well. We've also added in Perrysburg too. If you go back to the beginning of the year on the CapEx side or the initial guidance we had in December, that CapEx has gone up significantly with adding in the new Perrysburg plant as well. Then lastly, we've also structured a couple of deals recently where we have pretty back-end loaded cash flow profiles.
If you look at the cost of carry from a project perspective relative to the opportunity cost of the cash on the balance sheet at the moment, we've been using some of that cash, optimizing some of the project returns and timing of cash flow receipts from projects. You're seeing a lot on the project side that's causing noise for the year. Then later in the year, we'll give you an update on 2019.
Our next question will come from Ben Kallo with Baird.
Hi, good afternoon. Just on Series 6, I'm clear, because I heard you say, Mark, that you pulled forward or Vietnam's starting up faster, you have some extra costs there. Then you have the framing costs. I just want to make sure that we leave this call understanding where you are in the technology level. At the same time, the cost level, we can model that out. Are you there with the technology?
Yes.
This is a one-time thing, then the cost is where you expect it to be, then we pulled forward some Vietnam because you guys did better than you expected.
Yeah. Let's talk about that. Part of the question was around the ramp delta from what we had last quarter to this quarter. The vast majority of the ramp delta, $40 million, $30 million of that was just the geography move from start-up into ramp because we were successful at pulling forward Vietnam into production faster. We're actually going to pull forward Vietnam 2 sooner and faster as well. We highlighted to that as well. That's just a geography shift in terms of where the cost goes. $10 million was incremental ramp cost for the reasons that I said, mainly associated with Malaysia and Ohio, mainly on the back end for the framing cell and some of the manual processes that we had to put in place there. That is all being normalized away.
As I indicated as well, in Perrysburg, we have one accumulator left to do, which will help sort of buffer the back end a little bit more, which will help drive higher throughput. More importantly, we're 70% through on the tooling upgrade, tool set upgrades that we need to do. We have identified and prioritized a number of upgrades that we need to do to the tool set. We're 70% through on that. I'm happy with where we are there. As we indicated, Perrysburg, even with not having been fully buffered, but clearly the benefit of having buffered a significant portion of the line, we went from kind of a 60% capacity number to around a 90% capacity number. I'm really happy with that. Malaysia made similar progress, as we progress through the balance of the year, we'll have accumulators installed at both factories.
As we indicated, we should be exiting the year where we need to be on both those factories. Feel good from that standpoint. As I said before, the tool set, when we looked at it, we communicated in the last call, two most critical things to me were really cycle time and performance of the tool set, and everything is still performing extremely well from that standpoint. Very confident that once we finish the buffering, get the availability where it needs to be, both those factories will be running well. The other thing that we indicated was when we started up Vietnam, it already has an enhanced. It has a different framing cell, and as a result of that, it has much more resilient capability. Therefore, we're seeing better performance in Vietnam.
Every new factory that we have going forward will have that revised framing cell, plus other modifications that will be made. That should help us ramp and get to full entitlement of our second Vietnam factory, then obviously Perrysburg after that. We're starting to reach 430-watt modules. I think when you look at it from that standpoint, we feel good. On the technology side, the one thing that we need to tighten up the distribution. We're seeing the top bin starting to approach 430, but we're still seeing tails on the bottom end that are not where we want to be. Part of that, as I alluded to in my prepared remarks, is that we will start, normally we'll start and we'll ramp through with our non-ARC product.
As a result of that, it's going to be a lower efficiency product for sure, because ARC is going to give you, if you got a 400-watt module or so, you're going to get 12 watts from the benefit of ARC. That delta will drive to a lower bin. It's about two bins lower for a non-ARC product. It also drives a higher cost. The more non-ARC that I make, it does add about $0.01 to my cost profile. When we get everything up and running fully on ARC, and we've got a new ARC application and product that we're using that will effectively will get to almost 100% of our production will be on ARC, that $0.01 goes away. Near term, I got a $0.01, mix issue between ARC and non-ARC.
As we indicated in the last call, we are dealing with issues around tariffs that are impacting the cost of steel and the cost of the frame. The frame's $0.01 or north of that of impact that we're dealing with on the module cost. That's a headwind for us. We're dealing with a little bit of headwind there, and I'm dealing with a little bit of headwind just because my throughput's not where it needs to be. As we drive the throughput up, we go from a non-ARC to an ARC product, and then we're still working through some solutions to try to get the cost out on the frame, get out from underneath the tariff, source differently, other options that may happen that can drive that cost down, then we'll be able to get to where we need to be on that standpoint.
I don't want you to leave the call with a view that we still have some cost challenges. They're well understood in what we need to do. We're going to have to deal with, and the one that's probably structurally going to be the bigger challenge would be the frame.
Next, we'll hear from Julien Dumoulin-Smith with Bank of America Merrill Lynch.
Hey, good afternoon. Thanks for the question. First question here, just going back to the systems side of the business. Obviously, you've had continued success. Can I ask you, in terms of margins, obviously, we've seen pricing come down on the module side. How are you thinking about the margin profile for systems as you're continuing to scale this up and you see a little bit of a different geography? Is it still kind of in the same ballpark that you all have seen historically, part one? Maybe the second question I'll throw it out now. On the actual Series 6 panel, you kind of hit at it a little bit a second ago, but you kind of alluded to on the call the competitiveness of the panels themselves actually improving a little bit.
Any sense of the magnitude overall versus what you've been contemplating perhaps earlier this year, just as you think about it?
I'll let Alex take the system one. As it relates to the competitiveness of Series 6, we're very happy with how it's positioned in the market and the value that it's capturing. As I said, just on our bookings this last quarter, Series 6 versus Series 4, we've got a 10% premium of our Series 6 product over our Series 4 product, which when you couple a 10% premium with obviously a much lower cost profile entitlement then that's a very attractive product. The indications that we're getting from our customers, I was at recently at SPI, and one of our structure providers came over to me and was just very excited about Series 6 and the implications it has on helping them drive cost out as well on their structure, as well as the installation velocity and what we're hearing from our EPC partners.
They're very happy with the ease of the installation, the ease of the wiring and the connections associated with it. We're seeing the same thing in our own projects that we're self-developing and executing on right now. What I'm seeing on Series 6 right now has been very much in line, maybe slightly more favorable than what we had anticipated, but the product's been very well received in the marketplace.
Yeah, Julien, on the gross margin side, I'll refer you back again to the Analyst Day last year. I think at that point, we talked about our contracted pipeline having greater than 50% gross margins on the pure development piece. Now, the way we did that, make sure it's clear, is we broke out what we thought the entitlement for the module was, assuming a third-party module sale, the total EPC, and then the residual development piece is a very small number. But on that development piece, we were seeing contracted margins greater than 50% on the existing pipeline and new contracted development assets. On the development piece, we were seeing over 20% gross margins. Clearly, it's come down over time, but we're still seeing, I'd say, healthy gross margins on the development piece there.
The other thing to say is that it depends a little bit who the buyer is and what the structure of the deal is. If it's a contested auction in California for a busbar long-term PPA, where there's a huge amount of competition. We may no longer be the best buy because we're not going to sacrifice an acceptable gross margin profile for volume. What we're seeing is there are a lot of opportunities out there for us to deploy the skills that we have and the team and the opportunity set that we've built up over time in both sites and human capital, through more complex PPAs, for instance, with C&I buyers who have a different profile rather than contracting a significant amount of product and assuming that maybe some PPAs will fall away and just recontracting them.
Renewable energy buyers, corporate buyers, have a much more reputationally focused procurement process where they're looking to make sure that everything they procure will actually go through. There's a stronger counterparty risk piece that they place, and that positions us very well for that. Then finally, also on the UOG side, where again, we can bring talent to bear and skills to bear that we have as a company that aren't necessarily the same across the board, and we can see acceptable margins there. I think you got to look through where the procurement's happening. In general, I'd say that especially with the backlog we have today of systems procured, we're not going to go out and chase margins down for the sake of volume. We're seeing enough opportunity at acceptable margins today to maintain that gig on a year that we've talked about.
Yeah, the only thing I'll add is that, look, I think there's still quite a bit of demand for high-quality assets. You look at the capital stack and across it, whether it's the tax equity side, I think we thought that maybe it would be less competitive given some of the tax reform that happened last year. We're not seeing that at all. We're seeing even more getting involved in the tax equity side here in the U.S. The cash equity, there's a lot of money being raised around the world that people are looking to deploy. These are great quality assets from that standpoint. We're even seeing on the debt side aggressive pricing from that standpoint as well. All that creates higher value to the projects and the value that we can capture.
We've got a very strong pipeline right now, not only here in the U.S., but Japan assets and Australia. Very happy with what we have, and we're going to continue to build upon that. I think the safe harboring opportunity that's in front of us, and again, an opportunity to deploy our balance sheet and to carry multiple gigawatts of opportunities out into 2023 is going to be a very good position for us to be in.
Our final question will come from Michael Weinstein with Credit Suisse.
Hi, guys. Thanks for the question. With capacity expansion and CapEx already locked in and project development pipeline now pretty much above the 1 gigawatt per year target, this is a use of cash question. Can we expect M&A or maybe capital return next year? Regarding M&A specifically, do you see any opportunities for project pipelines? Are there any technologies or new verticals of interest out there that you might be looking at?
We continue to see most things that are happening in the space. We're always very happy to look at development portfolios. I think Mark mentioned earlier, we acquired a small development portfolio on the Southeast this year. We're always happy to look at both early-stage and contracted assets. We'll continue to do that. In terms of technology, given that we have a unique and different technology relative to most players in the market, it's hard to see what could be out there that would be beneficial to us. There are things we will look at. For instance, we did make an acquisition last year that's helped develop our antireflective coating technology, and it's been advantageous there. We'll continue to look at things around the core technology. The other thing that comes up often is storage.
As of today, we haven't seen anything that we think is differentiated enough to make it worth our while investing in the storage technology. We clearly will make sure we invest in understanding how to integrate storage and how to contract storage in a plant. Today, I don't necessarily see us investing in the core storage technology. On the second piece of that around capital return, we'll continue to look through the future opportunities and uses of cash. We talked before about having a waterfall and how we look at that, funding our core operations, the capacity expansions, any M&A of the development business, and then having a reserve, given the cyclicality of the industry that we're in. We'll look a bit more through that.
The other piece, I'd say, is towards the end of this year, we're going to be talking a little bit more about the opportunity of safe harbor. Because we look out to the end of 2019, there could be considerable opportunity for us to safe harbor, either through our module business or through using some of the balance sheet. That could give us a strategic advantage not available to other players without the financial resources we have. We want to always make sure that we're looking to use that cash as advantageously as possible. If we go through that full waterfall, we can't see an opportunity to expand the business. We have laid out a CapEx profile against a current capacity profile.
If the future of solar is as we believe, there's no reason that that is a stopping point for us from a capacity perspective. That's something we're going to continue to monitor and look at when would be the right time to consider additional capacity, which obviously we'd prefer to do versus returning capital. If we get to a point where we can't feel we can use that cash creatively at acceptable returns through that profile, we'll at that point look at options to return capital.
Yeah. The only thing I'll add to it too, in terms of that kind of use and priority is that one of the things now that as we move forward with Series 6 and where we are, and obviously there's a number of programs still in our efficiency roadmap, we're going to obviously look at our California team, which is our advanced research team, just to continue to think about the next evolution. Where's the next evolution with our current technology? Where can we take its fullest potential to go above and beyond? As we do that, there may be opportunities or opportunities may come up that some form of an acquisition or capability that we don't have today.
You got to remember, one of the things that helped us with our core technology to where we got it today was the acquisition of the IP from GE a number of years ago. Alex referenced the acquisition of a new antireflective coating that is obviously beneficial to our product, not only from the standpoint of giving a better efficiency benefit to the product, but also enabling us to capture almost 100% of the market with that product. I would imagine as we think forward, especially with the creativity and capability of our advanced research team in California, there will probably be some other opportunities as we think about, again, how do we evolve this technology to its fullest capability.
That does conclude our question and answer session for today and today's conference. We thank you for your participation.