FSLR - NASDAQ

Good afternoon, everyone, and welcome to First Solar's Q4 and Full Year 2024 Earnings and 2025 Guidance Call. This call is being webcast live on the Investors section of First Solar's website at investor.firstsolar.com. All participants are in a listen-only mode and please note that today's call is being recorded. I would now like to turn the conference over to your host, Byron Jeffers, Head of Investor Relations. Please go ahead, sir.

Good afternoon, and thank you for joining us on today's earnings call. Joining me today are our Chief Executive Officer, Mark Widmar, and our Chief Financial Officer, Alex Bradley. During this call, we will review our financial performance for 2024, discuss our future business outlook for 2025. Following our remarks, we will then open the call for questions. Before we begin, please note that some statements made today are forward-looking and involve the risks and uncertainties that could cause actual results to differ materially from management's current expectations. We undertake no obligation to update these statements due to new information or future events. For a discussion of factors that could cause these results to differ materially, please refer to today's earnings press release and our Form 10-K filing with the SEC. You can find these documents on our website at investor.firstsolar.com. With that, I'm pleased to turn the call over to our CEO, Mark Widmar. Mark?

Good afternoon, and thank you for joining us today. Beginning on Slide 3, I will share some key highlights from 2024. From a commercial perspective, 2024 saw us sustain a highly selective approach to contracting, which we foreshadowed at the start of the year, securing full year net bookings of 4.4 gigawatts at a base ASP of $0.305 per watt, excluding adjusters, India domestic sales and terminations. This led to a year-end contracted backlog of 68.5 gigawatts. We are pleased to have sold a record 14.1 gigawatts of modules in 2024. Our record net sales of $4.2 billion, in line with our prior earnings call guidance, represented a 27% increase year-on-year. Our full year diluted EPS, which included an after-tax impact of approximately $0.42 per share from the December sale of 2024 Section 45X tax credits, which was not included in our October guidance, came in below the low end of our guidance range at $12.02 per share. Alex will provide more details regarding our 2024 financial results later in the call. From a manufacturing perspective, we produced 15.5 gigawatts in 2024, including 9.6 gigawatts of Series 6 modules and 5.9 gigawatts of Series 7 modules. We began producing and selling our first CuRe modules from our lead line in Ohio in Q4. And we progressed our technology roadmap in 2024, commissioning a new dedicated R&D innovation center in Ohio, featuring a high-volume manufacturing scale production pilot line and began wrapping a new perovskite development line capable of producing small form factor modules at our Perrysburg campus. Our growth continued during 2024 as we exited the year with approximately 21 gigawatts of global nameplate manufacturing capacity, an increase of over 4 gigawatts over 2023, driven by the addition of our new Alabama facility and throughput optimization in Ohio. Additionally, we continue to construct our $1.1 billion Louisiana manufacturing facility over the course of 2024, which remains on track to begin commercial operations in the second half of this year. Once ramped, it is expected to increase our global nameplate manufacturing capacity to over 25 gigawatts by 2026. Turning to Slide 4, I will discuss our most recent shipments and bookings. At the end of 2023, our contracted backlog reached 78.3 gigawatts with an aggregate value of $23.3 billion or approximately $0.298 per watt. In 2024, we recognized sales of 14.1 gigawatts and contracted an additional 4.4 gigawatts of net bookings, resulting in a year-end contracted backlog of 68.5 gigawatts with an aggregate value of $20.5 billion or approximately $0.299 per watt. Since our previous earnings call, we have contracted a net 0.5 -- 0.4 gigawatts of new volume. This includes 0.3 gigawatts in India, approximately 40 megawatts of inventory below our current contracted backlog minimum win requirement. The sale was through a non-traditional revenue-sharing contracting arrangement with a module distributor. The residual net bookings include 0.6 gigawatts of sales to our traditional US utility-scale customer base, an ASP of $0.305 per watt, excluding adjusters, were up to $0.314 per watt assuming the realization of adjusters were applicable, partially offset by 0.5 gigawatts of terminations. For the full year 2024, including our 4.4 gigawatts of net bookings were approximately 5.1 gigawatts of gross bookings to our traditional US utility-scale customer base at an ASP of $0.309 per watt, excluding adjusters or up to $0.328 per watt, assuming the realization of adjusters were applicable, a 0.1 gigawatt of lower bin module sales through the aforementioned distributor, 0.6 gigawatts of domestic India volume, and 1.4 gigawatts of module agreement terminations. In addition, we saw 1 gigawatt of contract terminations in India, which were included in our contract subject to conditions present in number, but not in our bookings backlog for total 2024 module contract terminations of 2.4 gigawatts. A substantial portion of our backlog includes opportunities to increase the base ASP through the application of adjusters if we realize achievements within our current technology roadmap as of the expected timing of delivery of the product. By the end of Q4, we had approximately 37.1 gigawatts of contracted volume with these adjusters, which we estimate could generate up to an additional $0.7 billion or approximately $0.02 per watt, a majority of which would be recognized between 2026 and 2028. This amount does not include potential adjustments, which are generally applicable to the total contracted backlog for both the ultimate bin delivered to the customer, which may adjust the ASP under the sales contract upwards or downwards, and for increases in sales rate or applicable aluminum and steel commodity price changes. As reflected in Slide 5, our total pipeline of potential bookings remained strong with bookings opportunities totaling 80.3 gigawatts and a decrease of approximately 1.1 gigawatts from the previous quarter. Our mid to late-stage bookings opportunity decreased by approximately 2.5 gigawatts to 21 gigawatts and now includes 18.5 gigawatts in North America and 2.3 gigawatts in India. Within our mid to late-stage pipeline, our approximately 3.9 gigawatts of opportunities that are contract subject to conditions precedent. As a reminder, signed contracts in India will not be recognized as a booking until we have received full security against the offtake. As stated on previous earnings call, given the long-dated timeframe into which we are now selling, we need to align customer project visibility with our balanced approach to ASPs, payment security, and other key contractual terms. And the uncertainty related to the policy environment from the recent US elections, we will continue to leverage our position of strength in our contracted backlog and be highly selective in our approach to new bookings this year. We intend to continue forward contracting with customers who prioritize long-term relationships and appropriately value our points of differentiation. I'll now turn the call over to Alex who will discuss our Q4 and full year 2024 results.

Thanks, Mark. Beginning at Slide 6, in Q4, net sales were $1.5 billion, a $0.6 billion increase from the prior quarter. For the full year, net sales were $4.2 billion, an increase of $0.9 billion compared to the previous year, driven by higher volumes sold. Our Q4 net sales were benefited by $20 million from customer contract terminations and our full year net sales results were benefited by $115 million from customer contract terminations and reduced by warranty charges of $56 million due to the Series 7 manufacturing issues discussed on our prior earnings call. Gross margin was 37% in the fourth quarter compared to 50% in the prior quarter. Full year gross margin increased 5 percentage points from 2023, leading to a full year 2024 gross margin of 44%. Our Q4 and full year gross margin was lower than forecast due to several factors. Firstly, in December of 2024, we entered into a transaction with Visa, resulting in the sale of $857 million of Section 45X tax credits. In Q4, we discounted the carrying value of these credits by $39 million, which impacted Q4 gross margin by 3 percentage points and full year gross margin by approximately 1 percentage point to reflect the cumulative discount from the transaction. We received $616 million in cash proceeds in Q4 with the remaining $202 million expected to be received in Q1 of 2025. Secondly, warranty charges relating to manufacturing issues affecting the initial production of Series 7 modules as discussed on our prior earnings call, are estimated to result in total charges ranging from $56 million to $100 million, $50 million of which was recognized in the third-quarter and $6 million of which was recognized in Q4, primarily driven by additional impact in modules sold during the fourth quarter. Separate but related, we have taken corrective actions following the identification of the manufacturing issues in our initial Series 7 production that led to this warranty charge, which we believe has remediated these issues. While we continue to collaborate with customers to advance resolution, there is the potential that formal disputes arise. In addition, certain shipments to customers of product identified as impacted prior to shipment have been delayed as part of this remediation process. This resulted in the deferral of the sale of approximately 250 megawatts of modules from Q4 of 2024 into 2025. In aggregate, the $6 million warranty charge, the 250 megawatt module sale delay, the aforementioned sale of approximately 40 megawatts of [low in] (ph) inventory through a module distributor, and the aforementioned contract termination payments in aggregate negatively impacted the fourth quarter gross margin by approximately $16 million relative to our forecast. In addition, shipment delays both within the quarter and from Q4 2024 into 2025, in large part due to the aforementioned manufacturing issues resulted in approximately $36 million of incremental demurrage, detention, grounding, warehousing, and other logistics costs in Q4. Note, as of year-end, we held approximately 0.7 gigawatts of potentially impacted Series 7 modules in inventory. Thirdly, our newest operating facility in Alabama experienced higher than anticipated ramp related charges including material usage and spare parts consumption amounting to approximately $4 million. Despite these unforeseen costs, the plant continues to exceed the ramp-up pace of all previous Series 7 facilities in achieving its full production capability, underscoring our commitment to supporting the rapid growth trajectory of US manufacturing. In Ohio, we incurred approximately $18 million in unforecasted underutilization charges, yield losses, and other ramp-related expenses associated with the conversion to our CuRe technology on a high-volume manufacturing line. And lastly, in India, we incurred an unforeseen cost of $5 million due to a recent clarification of import regulations that imposed a specific duty on cover glass entering India, which applied to our 2024 cover glass imports. SG&A, R&D, and production startup expense totaled $11 million in Q4, a decrease of approximately $12 million relative to the prior quarter. Note, Q4 production startup was $9 million above the midpoint of our guidance range, largely related to our Alabama and Louisiana expansions. For full year 2024, operating expenses totaled $465 million, an increase of $14 million compared to the prior quarter -- prior year. This includes a $39 million increase in R&D expense, primarily driven by higher depreciation and maintenance costs, employee compensation due to headcount growth, and increases in material and module testing costs, in each case related to our significant investments in our R&D facilities and equipment. Additionally, start-up expense increased by $20 million, mainly due to the ramp-up of our Alabama and Louisiana facilities and the implementation of the lead line for our CuRe program. These increases were partially offset by a $9 million reduction in SG&A expense attributable to lower performance-based compensation expense as well as the non-recurrent $36 million litigation loss from the prior year. Our fourth-quarter operating income was $457 million, which included depreciation, amortization, and accretion of $124 million, ramp-in underutilization costs of $39 million, production start-up expense of $15 million, and share-based compensation expense of $6 million. For the full year 2024, our operating income was $1.4 billion, which included depreciation, amortization, and accretion of $423 million, rampant underutilization costs of $82 million, production start-up expense of $84 million, and share-based compensation expense of $28 million. Interest income, interest expense, other income, and foreign currency losses totaled $10 million in expense in Q4 and $12 million of income for the full year. We recorded income tax expense of $53 million in Q4 and $114 million for the full year. Our tax expense in Q4 and full year 2024 included a reserve of approximately $6 million for state taxes in jurisdictions that do not follow the federal tax provisions of the IRA for tax exemption of Section 45X credit sales. Residual increase in our tax expense was mainly due to a shift in the jurisdictional mix of our earnings towards higher tax jurisdictions and adjustments to our tax provision based on differences between estimated and actual tax liabilities. Q4's earnings per diluted share were $3.65 compared to $2.91 in the previous quarter. For the full year 2024, earnings per diluted share were $12.02 compared to $7.74 in 2023. Turning to Slide 7, I'll review select balance sheet items and summary cash flow information. Aggregate balance of our cash, cash equivalents, restricted cash, restricted cash equivalents, and marketable securities was $1.8 billion at the end of the year, an increase of $0.5 billion from the prior quarter and a decrease of $0.3 billion from the prior year. Our year-end net cash position, which includes the aforementioned balance left debt was $1.2 billion, an increase of $0.5 billion from the prior quarter, a decrease of $0.4 billion from the prior year. The increase in our net cash balance in the fourth quarter was primarily driven by the collection of $0.6 billion in proceeds from the sale of Section 45X tax credits and positive module segment operating cash flows, partially offset by capital expenditures associated with our Alabama and Louisiana facilities. The decrease in our net cash balance for the full year 2024 was primarily due to capital expenditures, partially offset by module segment operating cash flows. Cash flows from operations were $1.2 billion in 2024 compared to $0.6 billion in 2023. This increase was primarily driven by $1.3 billion in proceeds from the sale of Section 45X tax credits, partially offset by an increase in payments made to suppliers compared to the prior year and lower cash receipts from module sales in the current year. Capital expenditures were $314 million in the fourth quarter compared to $434 million in the third quarter. And capital expenditures were $1.5 billion in 2024 compared to $1.4 billion in 2023. I'll now turn the call back to Mark who will provide an update on technology and policies.

Thanks, Alex. Turning to Slide 8, I will now provide a brief overview of our technology strategy. As we'll further discuss in a moment, we believe that the age of electrification is upon us, where electricity is a lifeblood of the modern economy and the way of life. While meeting this unprecedented demand for electricity is going to require diverse sources of energy generation, we believe that solar will be a key part of the solution mix, providing the opportunity to develop and commercialize the next generation of solar technologies. Optimizing across efficiency, energy, and cost, we believe the future of solar has a dependency on thin-film technologies. As a result, we have embarked on a focused technology strategy concentrated on three core pillars. The first pillar centers around improvements to our core single-junction CadTel semiconductor technology. As previously noted, in Q4 2024, we commenced a limited commercial production run of modules employing our CuRe technology, which we expect will be completed in Q1 of 2025, and have deployed the first of these modules in the field. Upon successful field performance validation to confirm the results of our accelerated life testing, we intend to permanently convert the Ohio lead line to CuRe. Previously forecasted in Q4 of 2025, we now expect to convert the Ohio lead line back to CuRe in Q1 of 2026 to address the key manufacturing learnings from the initial production run and to allow further time for field validation and intend to begin a phase replication of the technology across our fleet in early 2026. The second pillar relates to developing the next generation of thin-film semiconductors able to be deployed at commercial scale. This research is focused on perovskite technology and performed at our California Technology Center, also supported by the associates brought over through the Evolar acquisition. We expect the effort to benefit from our new dedicated perovskite development line in Ohio, which is expected to be fully operational by Q2 of this year. To date, we have been able to achieve reliability results that we believe are comparable with best-in-class R&D efforts, and we continue to advance our work on improving efficiency and stability in the ways to develop a viable and commercially scalable perovskite product. Our third pillar centers around the next-generation tandem device combining two semiconductors, which optimize to a different range of solar spectrum to create a very high-efficiency module. While tandems can utilize a range of available PV semiconductors, we believe at least one needs to be thin film and that the optimal solution will require that both semiconductors are thin films. In other words, our view is there is no tandem without thin film. While we previously explored the possibility of a crystalline silicon CadTel tandem product, we believe the energy inefficiency benefits of a fully thin film approach provides a more likely path to a future transformative device and are therefore prioritizing our research into this structure. We believe this three-pillar framework enables us to compete in the near term with the best-in-class crystal and silicon technology through advancing our CuRe technology platform. And in the long-term, we believe our thin-film technology and manufacturing leadership and expertise positions us as well as -- it positions us well in the race to commercialize and scale a perovskite-based thin-film semiconductor. I would like to take a moment to address the intellectual property issues that continue to challenge the crystal and silicon industry. Since our last earnings call, during which we shared a sustainable overview of the IP landscape, we continue to see leading manufacturers assert patent-related claims against one another. Most recently, Trina Solar sued Canadian Solar's China subsidiary, CSI Solar for patent infringement in China. JinkoSolar sued LONGi in Australia, China, and Japan and Visan in a California court and Longi sued JinkoSolar in the US. As we disclosed in 2024, First Solar also possesses a TOPCon patent portfolio through our acquisition of TetraSun in 2013. Our TOPCon patent portfolio includes issued patents in the United States, Australia, Canada, China, the European Union, Hong Kong, Japan, Mexico, Malaysia, Singapore, South Korea, United Arab Emirates, and Vietnam, with validities extending to 2030 and beyond. The portfolio also includes patent-pending applications in the European Union, Japan, Hong Kong, United Arab Emirates, and Vietnam. We continue to enforce our rights under this patent portfolio. Earlier today, following numerous commercial engagement efforts, we filed a complaint with the United States District Court for the District of Delaware against various JinkoSolar entities alleging infringement on one of our US TOPCon patents. This lawsuit is consistent with our previous statements that we intend to actively enforce our intellectual property rights against companies that exploit infringing TopCon technology. At the same time, we continue to advance our discussion with certain other companies that have expressed interest in resolving the issue commercially. For example, earlier today, we announced that we entered into our first agreement for the licensing of our US TOPCon patents with Talon PV, a US manufacturer of solar cells, which recently announced a 4 gigawatt TOPCon cell manufacturing facility scheduled to commence operations in the first quarter of 2026. The Jinko lawsuit, combined with our commercial discussions with companies that have expressed interest in securing a license to First Solar's TOPCon patent portfolio as well as the decision by some manufacturers to pivot to the Perc node from TOPCon give their own and their customers potential exposure to lawsuit risk reinforces our belief that developers, project owners, and PPA offtakers as well as debt and tax equity financing parties must evaluate the risk of procuring TOPCon solar products that could be subject to IP-related legal challenges. Once again, this environment underscores the benefits of First Solar's unique, highly-differentiated CadTel semiconductor technology over highly commoditized crystalline silicon panels. Turning to Slide 9. A word about the overall market conditions and our policy environment. Beginning with the macro-environment, President Trump has defined an expansive economic mandate that could reshape the US economy over the next four years, particularly in terms of electricity production and consumption. His administration's goals of accelerating economic growth, reducing inflation, establishing global energy dominance, bringing American manufacturing jobs back and championing innovation, including artificial intelligence require abundant stable power generation. Forecasts show that the US will need 128 gigawatts of new capacity by 2029 to meet high summer peak demand. And while President Trump is expected to preside over the first meaningful growth in electricity demand this century, it will not be without challenges, the most significant of which is the time it takes to expand power generation capacity and grid infrastructure. Considering the new natural gas capacity could take half a decade to come online and cost twice as much as it did five years ago, thanks to supply chain constraints, including the shortage of turbines. Large-scale nuclear power plants take more than a decade to permit, construct, and commission. While these decommissioned nuclear plants are an option, reportedly, only three assets can be economically recommissioned by 2028. And while hyped and anticipated to be quicker to deploy, small modular reactors are not expected to operate commercially at gigawatt scale before 2035. Quite simply, in order to avoid potential energy price related inflation, maintain its economic and innovation competitiveness, and secure its energy independence, the country cannot wait that long. Given its attributes of low cost and speed to deployment relative to other sources of energy generation, solar should clearly be a significant part of the near-term solution mix. The administration and Congress must ensure that this unprecedented growth in power generation capacity is not deep in the country's dependence on China and that American manufacturers have access to a level playing field. The threat from China is existential and cannot be addressed through market factors, economics, and innovation alone. Given this backdrop, we continue to advocate for decisive actions to address China's dominance of the global solar supply chains and weaponization of subsidy fuel overcapacity to undermine American manufacturing, energy security, and enduring middle-class jobs. This includes establishing Foreign Entities of Concern or FIAC laws that exclude companies tied to the Chinese Communist Party from assessing US taxpayer-funded incentives. Considering the large number of Chinese manufacturers that have set up low-value-added US assembly shops importing high-value overseas components to secure billings and incentives, FIAC legislation not only prevents China from unfairly accessing US taxpayer-funded incentives but it's also impactfully reduces the cost of programs like 45X advanced manufacturing tax credit, while ensuring that value created by domestic manufacturing is retained in the US and not remitted to China. Moving to trade, the Southeast Asia AD CVD case filed in April of 2024 has made significant progress since our last earnings call. In November, the Department of Commerce released its preliminary anti-dumping determination, establishing higher-than-anticipated cash deposit rates for Cambodia, Malaysia, Thailand, and Vietnam, while issuing affirmative critical circumstances findings for Thailand and Vietnam. In addition, this past month, Commerce revised certain preliminary CVD rates upon examining the cross-border subsidies tied to Chinese wafer, glass, and solar space. In Cambodia, as a result of this examination, four solar manufacturers' imports now have post-preliminary CVD rates of 729% and a total AD plus CVD preliminary rate of nearly 850%, up from a rate of around 150% prior to this examination. We expect to see determinations from commerce in respect of such Chinese cross-border subsidizations concerning Malaysia and Vietnam in the near term, which may increase the CVD rates applicable to subject silicon manufacturers in those countries. Overall, the Trade Committee is pleased with the results related to Cambodia, Malaysia, Vietnam, and Thailand and continues to monitor import data and trade practices relative to all countries, including Laos and Indonesia, among other countries, and all trade remedy options remain on the table. I will now turn the call back to Alex, who will discuss our 2025 outlook and guidance.

Thanks, Mark. Before discussing our financial guidance, I'd like to reiterate our growth and investment thesis and our approach to backlog and bookings. Continue to focus on differentiation and are guided by an approach that balances growth, profitability, and liquidity. This framework informs our long-term strategic decision-making. It guided our strategy to exit the systems business at the end of the last decade and to significantly expand our domestic US R&D and manufacturing base. It also drives our approach to forward contracting volume, where we built a significant contracted backlog totaling 68.5 gigawatts of volume at year-end 2024 at an ASP of nearly $0.30 per watt. Our reported backlog, which includes US and rest-of-world bookings with our typical contractual security provisions, but excludes contracts signed in India unless backed by 100% security is made up of two types of contracts, those related to a specific asset or project and frameworks, which are typically larger multi-year and therefore often have less certainty of a specific delivery timing with customers having more flexibility to shift volume within and across delivery years. Common across these structures is a fixed-price structure, which may include adjusters for technology improvements and which typically include adjusters for bin-class, freight risk, and commodity risk. As of December 31, 2024, over 90% of our backlog had some form of steel and/or aluminum commodity cost protection and substantially all of our backlog had some form of freight protection. This long-term approach to our business model and customer contracting is especially important during periods of macro and industry uncertainty, such as we face today. As Mark discussed, we believe there are a number of drivers of sustained long-term growth in the demand for the energy generation, and with its relatively low-cost profile and speed to power, solar is well-positioned to be a fixture of the energy mix in advanced economies as we progress through the next decade. However, at present, continuing a trend that increased throughout 2024, there remains significant near-term uncertainty, largely driven by the still unresolved policy environment following the US elections in November. This uncertainty is also driving a pause in at least some domestic manufacturing expansions. Just a one example, earlier this month, Indian solar manufacturer Premier Energies announced it is pausing plans to build a cell plant in the US citing policy uncertainty. Given the multi-year lead-time required to build and commission a cell manufacturing facility, the uncertain policy environment also has a potential delay and increased presence of domestic high-value manufacturing competition. This uncertainty, however, is also driving customer caution, particularly as it relates to new procurements and a lack of clarity as to project timelines in 2025. This uncertainty is further reflected in our 2025 allocation position. Excluding India, we remain cumulatively oversold through 2026. As previously discussed, this oversold position is deliberate and in addition to providing us with revenue visibility in an industry that has historically experienced volatile pricing conditions provides us resilience to the uncertain timing of delivery inherent in some of our larger framework contracts, the natural tendency for delay in the project development process as well as the potential for incremental supply as we start-up and ramp new factories. To further out the delivery timeframe, the more comfortable we are with overallocation. The closer we get to delivery dates and as we enter any given year and undertake our annual planning process, the more we look to ensure demand can be met with available supply. Therefore, whilst beneficial in the longer term providing us with flexibility, the trade-off is that this over-allocation must be resolved in the near-term delivery window. As we progress through 2024 and as discussed on our previous earnings call, we saw increasing requests from customers to push out delivery schedules as a function of project development delays. We also saw 0.6 gigawatts of terminations for convenience and 1.8 gigawatts of termination for default, including 1 gigawatt of terminations for default to India domestic contracts, which were signed but not included in our bookings backlog, the majority of which incurred after the initial supply and demand allocation balancing for the year was completed. As mentioned in our third quarter earnings call, for terminations, we have not received the termination payment entitlement, we will litigate or arbitrate seeking to enforce our full termination payment rights under the respective contracts. While we enter 2025 in a strong position with respect to our US production, we are in an under-allocation position for our Series 6 Malaysia and Vietnam production. This is driven in large part by two factors, one, customers employing module delivery shift rights, and the other, the previously mentioned contract terminations that occurred in 2024. As it relates to module delivery shift rights, as previously mentioned, multiyear framework contracts typically have less certainty over specific delivery timing. As we enter the year, approximately 1 gigawatt of contracts with shift rights allowing product to move out of 2025 have had this option exercised. In addition, customers with intra-year flexibility are in many cases requesting deliveries in the second half of the year, which we expect will drive a back-end weighting of our full-year revenue and shipment results. Additionally, due to the aforementioned contract terminations in 2024, we are not delivering international product that we intend to deliver in 2025 under those terminated contracts. While typically, we would seek to mitigate the impact of these factors by reallocating shifted or terminated deliveries to other customers, our ability to do so in the near term, particularly as it relates to our Southeast Asia produced product is constrained by the policy environment in Europe, India and the US. In Europe, the combination of China's strategy of product dumping and oversaturation with the aim of capturing the European market, combined with a lack of EU block political action to employ responsive trade remedies to level the playing field has resulted in the EU market becoming a less attractive destination for our international production. As a consequence of this environment, in Q4, we made the decision to shut down our EU-based sales operation, which we occurred approximately $3 million in severance charges in Q4. In India, while we applaud the efforts of the Indian government to address China's efforts to dominate its domestic market and are encouraged by its tariff and non-tariff measures, including the expected 2026 expansion of the existing approved list of models and manufacturers to cover cells as well as modules, these measures have the effect of largely eliminating the India market as a destination for our Malaysia and Vietnam product. Finally, in the US, there is uncertainty relating to the post-election policy environment, including the increased prospect of tariffs and potential revisions to the IRA, currently subject to a reconciliation process, which according to public reporting is not expected to be resolved until the second half of 2025. In this environment, our customers generally view our domestic modules as advantage relative to our international products. While taken together, these items create near-term headwinds for our international production, assuming current policy remains unchanged, we continue to see long-term opportunities to place international products in the US and optimize our allocation position. The IRA domestic content bonus provisions create significant economic value for our customers. This is enabled by way of the more recently issued points-based domestic content bonus guidance. We believe the points-based guidance provides our customers with the clearest route to enable and finance their domestic content bonus qualification efforts. For First Solar, we expect the points-based guidance will provide us with greater supply-chain flexibility in the US and greater opportunity to optimize allocation across our entire fleet while maintaining the ASP and the original module sale agreement. In summary, while we're encouraged by the long-term opportunities to optimize the entirety of our global production fleet, the near-term combination of increased project delays and uncertainty experienced by our customers, promoting their utilization of module delivery shift rights both inter and the termination in 2024 of contracts that included modules expected for delivery in 2025, together with the imbalanced demand for domestic versus international product given the current policy environment in key markets leads to a challenging full-year and quarterly supply-demand allocation and balancing position as we enter 2025. So, with this context in mind, I'll next discuss the assumptions included in our 2025 financial guidance. Please turn to Slide 10. As it relates to growth and production, our factory expansions and upgrades remain on schedule to increase our expected global nameplate capacity to 25 gigawatts by 2026. In Ohio, we completed our footprint expansion in Q1 of 2024. In Alabama, our factory is expected to exit the ramp phase at the end of Q1 2025. Our newest facility in Louisiana is forecast to be in startup into Q3 of 2025 and to begin ramping production in the second half of this year. Combined, this leads to a forecast domestic production of 9.2 gigawatts to 9.7 gigawatts. Internationally, given the supply-demand imbalance that Southeast Asian product just detailed and the impact of the current Europe, India, and US policy environment, we've made the decision to reduce output of Series 6 international products from our Malaysia and Vietnam factories by a combined total of 1 gigawatt this year for a total forecast production of 5.8 gigawatts to 6.1 gigawatts. In India, we're forecasting 3 gigawatts to 3.2 gigawatts of production, approximately 70% destined for the US market and the remainder for domestic sales. This leads to a combined full-year 2025 production forecast of 18 gigawatts to 19 gigawatts. Growth-related costs are expected to impact operating income by approximately $110 million to $130 million. This comprises startup expense of $60 million to $70 million since that's substantially all in connection with our new factory in Louisiana, an estimated ramp in underutilization costs of $50 million to $60 million across our factories in Alabama, Louisiana, Malaysia, and Vietnam. Note, in connection with the aforementioned 1-gigawatt reduction in Southeast Asian production, we are temporarily deploying a portion of our experienced engineering and manufacturing associates to our new US facilities. While contributing to growth-related costs, this is expected to support the start-up and ramp of these factories. Combined with inventory drawdown, including India volume shipped to the US in late 2024, we forecast module sales of 18 to 20 gigawatts, of which 9.5 gigawatts to 9.8 gigawatts is produced in the US and approximately 1 gigawatt is assumed to be domestic sales in India. For the full year, we expect to recognize an ASP of approximately $0.29 per watt, including domestic India sales and the value of certain technology, commodity, and freight adders. Included in our ASP assumption and within the guidance range is approximately 1.4 gigawatts of international product, approximately evenly split between Series 6 International and Series 7 India domestic product, which is forecast to book and bill within the year at an ASP below the current backlog average. From a cost perspective, full year 2025 cost per watt produced is forecast to be approximately $0.20 per watt, an approximate $0.01 per watt increase over 2024. It's driven by six key factors. Firstly, in Ohio, following on from the fourth quarter of 2024, we anticipate ongoing underutilization and yield loss impacts from our CuRe technology run on a high-volume manufacturing line, which is expected to conclude by the end of the first quarter. Secondly, in Alabama, as the factory exits the ramp phase forecast to occur in Q1, but continues to produce at less than normal expected production capacity as is forecast for the remainder of 2025, we expect to see an increase in cost produced capitalized in inventory until such normal production capacity is reached. Thirdly, as it relates to our India factory, product destined for the US market is made with a tracker mounting structure and a higher production cost point relative to the fixed tilt structure used for India domestic sales. In 2025, approximately two-thirds of our India production is destined for export versus approximately half in 2024, leading to higher expected blended production cost at this factory. Fourthly, the scheduled reduction of 1 gigawatt in total production at our Vietnam and Malaysia factories is expected to increase fixed cost per watt at these sites as a function of lower throughput. Fifthly, on a combined basis, we anticipate a negative impact to fleet cost per watt from the relative mix of higher cost to US versus lower cost international production, which increases both as a function of the 1 gigawatt aggregate Malaysia and Vietnam production decrease as well as the increased US capacity coming online. And lastly, we expect to see a cost headwind from the recently announced Section 232 tariffs imposed on aluminum imports into the US at a rate of 25%, given the aluminum frame rails for our US Series 6 product are imported from a Malaysian supplier, after which they undergo a manufacturing process in our domestic machinery to produce the ultimate frame. The administration also recently announced it is assessing the implementation of reciprocal tariffs to items from countries currently applying import duties to American products. Note, our module sale contracts for international product deliveries typically have some form of tariff protection if new tariffs are imposed on the importation of modules into the US, whether in the form of First Solar termination right or tariff absorption that is either shared with the customer or exclusively borne by the customer customers. Related to the topic of tariffs, a word on the export controls concerning five critical minerals, including products containing tellurium announced by China's Ministry of Commerce earlier this month. Although tellurium and products containing tellurium, including among them cadmium telluride, our key raw materials used in our module production process. We have over the past decade employed a strategic sourcing strategy to diversify our tellurium supply chain to mitigate a sole sourcing position in China and are undertaking additional measures to mitigate dependencies on China for certain products containing tellurium. While we continue to evaluate whether there will be any operational impacts from China's decision, this latest development emphasizes the urgent need for the United States to accelerate the strategic development of copper mining and processing of its byproduct materials, including tellurium. Moving to cost watt sold, we're forecasting fleet average sales rates, warehousing, ramp, underutilization, and other period costs of approximately $0.04 per watt, a reduction from 2024, but above our previous long-term cost assumptions. As it relates to sales rates, although we've seen some increase in freight rates, we are largely contractually protected from these impacts. However, we do expect some incremental freight charges as we increase our volumes sold from India to the US. We have seen an increase in our warehousing and storage needs driven by our increase in production capacity as well as the allocation balancing challenges referenced earlier in the call, which are expected to result in a back-ended shipment and sales profile in 2025. At the same time, warehousing rates have increased given capacity constraints, driven by a combination of increases in domestic manufacturing, combined with the surge of imports as manufacturers seek to mitigate the expected tariff risk following the November election. As it relates to the aforementioned Series 7 manufacturing issue, we've estimated warranty losses of $56 million to $100 million. As noted, as of year-end, we held approximately 0.7 gigawatts of potentially impacted Series 7 modules in inventory. The combination of production and period costs results in a forecasted full year 2025 cost per watt sold of approximately $0.24. From a capital structure perspective, our strong balance sheet has been and remains a strategic differentiator, enabling us to both weather periods of volatility as well as providing flexibility to pursue growth opportunities, including funding our Series 6 and Series 7 growth. We ended 2024 in a strong liquidity position and with forecasted operating cash flows from module sales, coupled with residual operating cash flow from the sale of 2024 Section 45X tax credits in December of 2024 and advanced payments from module orders, we expect to be able to finance our currently announced capital programs without requiring external financing. As it relates to our 2025 Section 45X credits, we are not forecasting the sale of these credits in 2025 and therefore assuming no discount to the value of these credits for a sale to a third party. But as in previous years, we will continue to evaluate options and valuations for earlier monetization. I'll now cover the full year 2025 guidance ranges on Slide 11. Net sales guidance is between $5.3 billion and $5.8 billion. Gross margin is expected to be between $2.45 billion and $2.75 billion or approximately 47%, which includes $1.65 billion to $1.7 billion of Section 45X tax credits and $50 million to $60 million of ramp-on underutilization costs. SG&A expense is expected to total $180 million to $190 million versus $188 million in 2024, demonstrating our ability to leverage our largely fixed operating cost structure while expanding production. R&D expense is expected to total $230 million to $250 million versus $191 million in 2024. R&D expenses increasing primarily due to commencing operations at our R&D innovation center and perovskite development line and the expectation of adding headcount to our R&D team to further invest in advanced research initiatives. SG&A and R&D expense combined is expected to total $410 million to $440 million and total operating expenses, which includes $60 million to $70 million of production startup expense are expected to be between $470 million and $510 million. Operating income is expected to be between $1.95 billion and $2.3 billion, applying an operating margin of approximately 38% as inclusive of $110 million to $130 million of combined ramp costs and plant start-up expense and $1.65 billion to $1.7 billion of Section 45X credits. Turning to non-operating items, we expect interest income, interest expense, and other income to net to zero. Full year tax expense is forecast to be $100 million to $120 million. This results in full year 2025 earnings per diluted share guidance range of $17 to $20. Note from an earnings cadence perspective, we expect between 2.7 gigawatts and 3 gigawatts of module sales in the first quarter at a gross margin similar to the full year average, resulting in first-quarter earnings per diluted share of between $2.20 and $2.70. Capital expenditures in 2025 are expected to range from $1.3 billion to $1.5 billion. Approximately half of our CapEx is associated with capacity expansion, majority of which relates to our Louisiana plant with the remainder driven by R&D programs, technology replication, and maintenance. Our year-end 2025 net cash balance is anticipated to be between $0.7 billion and $1.2 billion. Turning to Slide 12, I'll now summarize the key messages from today's call. With respect to 2024, while our full-year diluted EPS came in below our expectations, this result was largely attributable to the incurrence of discrete costs in the pursuit of achieving the growth and liquidity principles of our strategic decision-making framework. Growth in the case of ramp costs associated with our Alabama facility as we expand US manufacturing as well as lower throughput and higher yield losses connect to our initial conversion to high-volume CuRe manufacturing and liquidity in connection with the sale of $857 million of Section 45X tax credits as we significantly strengthened our industry-leading balance sheet. The other cost driver impacting 2024 margin was similarly discrete, resulting from increased logistics costs associated with delayed shipments, largely in connection with the manufacturing issues affecting the initial production of Series 7 modules. That said, from a revenue perspective, full-year 2024 net sales came within our guidance range for the year and we exited the year maintaining a significant contracted backlog totaling 68.5 gigawatts with an average ASP of nearly $0.30 per watt. In addition, our full year 2024 diluted EPS result represents a 55% increase over the prior full year results. As we enter 2025, we are in a position of strength with respect to our US production, although together with the rest of the industry, we are confronted with uncertainty in the current policy environment in key markets, leading us to an under-allocated position with respect to our international product and increased costs associated with reduction in our Southeast Asian production. That said, assuming the current construct of domestic content provisions remains unchanged, we are encouraged by the long-term opportunities to optimize the entirety of our global production fleet. For the full year 2025, we're forecasting an earnings per diluted share guidance range of $17 to $20, the midpoint of which would represent an approximately 50% increase over 2024. With that, we conclude our prepared remarks and open the call for questions. Operator?

[Operator Instructions] Our first question today is Brian Lee, Goldman Sachs.

Hey guys, good afternoon. Thanks for all the information and color. I wanted to start off first question just on the guidance range. It seems a little wider than usual. I know Mark, Alex, you talked through some of the uncertainties and moving pieces. But how much is really tied? And I'm focused kind of on the shipment and revenue guidance, how much is tied to selling the 1 gigawatt of India volume left over from last year? Maybe walk through how that process is playing out, what ASPs you're expecting to realize? And then second question would be around just safe-harbor potential. Are you seeing any of that? What are your customers doing in this uncertain environment where is there even potential as you move through the year that you kind of turn output back up from Southeast Asia to allow customers to take advantage of Safe-Harbor maybe blend? Just curious on your thoughts around if any of that is happening and what you would see as a potential impact if it did later through the year? Thanks, guys.

Hi, Brian. I'll take the Safe-Harbor question. I'll let Alex talk about the guide on the shipment and the revenue. The -- look, I think right now, Brian, I’ve been meeting a lot of customers over the last several weeks in particular. And everybody's -- look, a lot of people have already safe harbor. They safe harbor as of the end of last year, some of our customers have used modules, but most of them have used transformers and the like high voltage equipment to do their safe harboring. And so most of safe harbor [indiscernible] guidance and they've got a multiyear kind of portfolio that they would execute against that. But everybody is kind of in a position right now if there's so much uncertainty, let's not change anything at this point in time. So the projects, I think are going to continue to move forward. Clearly, they're running into some obstacles or potential headwinds as every day there seems to be something new coming out from the administration that could have some implication or a reason that you have to step back and reassess and reevaluate. As it comes to, if we can get better clarity as an example, and the sooner we get certainty, the better, the biggest thing this industry needs and you can pretty much talk to anyone in this industry is just the certainty of which how to move forward. And we don't have that clarity right now. But once we have that certainty, then we can start thinking through optionality and how do we optimize our customers will look at it from that lens, we'll look at from that lens. But if you get into a conversation from achieving a point requirement for domestic content, which would then allow for latitude of both the domestic product and potentially international product to achieve the points required at the project level. Then you get into conversations of, well, I don't want to change anything right now because I don't know exactly what could evolve differently as we move forward. And then the other is, okay, if we have that conversation, what happens if tariffs are imposed? Today, I've got largely a domestic contract and there's still uncertainty around tariffs as Alex mentioned in his comments and what are the implications around that? And do I want to take on that additional risk right now? And neither counterparty wants to take 100% of that risk. I mean the way our construct is done right now with our module sale agreement, we generally share in that risk. But knowing that you could be going into an environment that has greater certainty of tariffs, then it's because more of a conversation with the customer, well, who's also going to take that risk? Today, they got a commitment for a domestic module. If we're going to talk to optimization, which could include international modules, that's a risk allocation conversation to have with a customer and neither one of us really want to step into that with that amount of uncertainty. So I would say, yes, there's been a lot of activity for safe harbor. I think that bodes very well for projects to continue to be built out this year. But I think the amount of uncertainty we have right now, people are somewhat saying, hey, let's just sort of stay where we are, continue to assess once we have that amount of certainty, we can determine how to move forward from there.

Hey, Brian, on the shipment piece, so we guided 18 gigawatts to 20 gigawatt range. We said there's about 1.4 gigawatts of unsold dependency for the year. And then we said it's about half, half split between domestic India and international Series 6. So on top of that, you've got some uncertainty around US. I think with the guide, we had 9.5% to 9.8% on US sold. I would say in general, the other risk we run is that the year is very back end and we haven't given a split across the full year. We gave a Q1 number. But even from the Q1, you can see it's not a run rate if you were to divide the full year by four. And as you back-end and have more risk at the back-end of the year, there's always risk of things slipping out. We also said that there's negotiations ongoing with certain customers related to follow-on questions from the warranty issue we disclosed. We believe those will get resolved and that has also resulted in some delays in shipments. So generally, I'd say there's 1.4 of international book-and-bill dependency for the year. There's a little bit of US and then there's just the overall profile of the year, which adds some risk.

Yeah. And I would just say -- just add to that, Brain, I don't think there's -- and we have the assumption in our plan for the India volume at a current market price, Randy. I don't know if there's much risk of the sell-through on India, just the mechanics of generally how the India market works. And as we indicated, we won't recognize the booking until we have 100% payment security. So I think there's really not much risk there. And we are assuming that there will be better clarity as we exit the second-half of this year or as we enter into the second-half of this year that will allow some of this optimization and mix of international into with domestic product. So that's how we anticipate for that book and bill to clear and it's not a lot of volume.

The next question comes from Philip Shen, ROTH Capital Partners.

Yes, thanks for taking my questions. A few topics here. First one on warranty. There was a lot of speculation on the Street heading into the earnings about your warranty expenses and how they could impact your results. Can you confirm that you guys have indeed solved the production issues? In other words, are the Series 7 modules coming off the line today expected to perform as expected? Your 10-K suggests the warranty expense risk may be capped at $100 million. Is that indeed the case? And what is the risk that the overall warranty expense could be meaningfully higher than $100 million? And if so, higher, can you quantify how much higher? So that's the first category. Second topic here is about the guide just very simply, can you share to what degree tariffs are included in the guide? And then finally, back to warranty for a little bit here. Some of the customers that have bought the modules that are impacted by a lower production. They're sharing with us some of the production might be like 5% lower for some of these systems and just 7% lower production could bankrupt the projects. So while you may be buying down the warranty, these customers may be left with underperforming projects and systems and maybe reluctant to buy First Solar modules again. So what are your thoughts on these challenging situation that some of these customers, including Tier-1 IPPs and regulated utilities may be left with underperforming projects? I know that's a lot. Really appreciate it. Thank you.

Yeah. All right. So starting off on the warranty and I think even in our last call, Phil, that some of the impact, some of the changes that we identified were already corrected, right? So we identified in November that there was a couple of items. One is the engineering performance margin and calculation around that. That one was actually fixed and implemented before the even last earnings call. The other items that are impact -- that were impacting the Series 7 underperformance, for example, the dual time on the washer. And again, it didn't impact all of our factories, in particular, our Alabama factory, which was just ramping up, didn't have the same setup and the layout. So we didn't have the dual impact on product coming out of Alabama. But across the fleet, those items have been corrected in the products that are being produced, the Series 7 product is being shipped right now. It will -- based off of our accelerated testing and everything else we did, we'll perform in line with the expectations of the specification of which we've highlighted for that product. We've also used a third party to come in and do an independent assessment of the root cost corrective action and implementation. That study is still ongoing right now because some customers would like to have a third-party validation. We'd expect to have that here in the near term to help provide additional comfort to our customers and independent engineers and obviously financing parties and the like. So there's a little bit of work still being done on that. The range is the best range that we have right now. Based off of the information we have, the understanding of the underperformance, the testing that we've done in our laboratory measurements, it's still the best indicator of the range. And again, we booked to our low end of the range because we don't have a better estimate of -- within the range of -- that would be more likely than not than what we currently have estimated. So that's the basis for the reserve. It's $56 million. There's a range up to $100 million. And again, that's the best information we have at this point in time. As it relates to tariffs, we're not assuming any tariffs on modules coming into the US for First Solar's product coming from either Malaysia, Vietnam, or India. We are assuming other adverse impacts of tariffs, for example, on aluminum, as Alex mentioned, we are assuming that our aluminum components that we do have imported into the US will be subject to the tariff rate, which I believe is 25%. There are some other noise that came across like even recently this week about some additional charges that will be associated with Chinese-owned or manufactured freight carriers. And so we are assuming that adverse impact. So it's a moving target right now, Phil, in terms of the total impact that we could see and we'll have to see how it continues to play out from that standpoint. But to the extent we know something right now, we have tried to accommodate for that. Look, as it relates to the warranty and how we see this, look, our relations with our customers as we generally have -- we've stated this before and most of our customers, I think would see it the same way that this is a true partnership. And there are -- there will be times where there's going to be stress on both sides of the house. In 2022, from our standpoint, we had substantial stress on our side of the house, where we had to deal with abrupt changes in an ocean freight market, carrier cost for containers going up 5, 6, 7 times. We had explosion of commodity cost increases, aluminum in particular. And we had some pretty substantial adverse impacts on our financial results, which was about $500 million, I think is what we disclosed for 2022, headwinds that we took. And look, we absorbed all that. That's the obligation. We have a contract and there's frameworks associated with that contract and that's the measurement of which we engage with our customers. Same thing this year. I'm going to take some changes and some hits on commodity cost increases and tariffs and other things that are going to come through on my input costs, which I will absorb and still deliver and fulfill 100% of my obligations. The contract that we have with our customers around our warranty is that there's an obligation at which we'll stand 100% behind the technology within the parameters of that contract. If you try to expand beyond that and try to get into a field of consequential damages using that terminology, it's just not within the parameters of what we agreed to between the two counterparties. We'll do everything we can to assess any shortfall that we have on the technology and the product. We'll stand behind the product and we'll stand behind it in accordance with the framework of the warranty that we have. I mean, that's really the four corners of what we have to operate and manage by. And again, our customers see this as a long-term journey. And obviously, First Solar has been a very credible counterparty and partner and I think they still see the value of a long-term relationship. I don't see it having a substantial adverse impact on our relationships and long-term commitments that we have to our customers or our customers to us.

We'll take the next question from Mark Strouse, JPMorgan.

Yes, good evening. Thanks very much for taking our questions. I want to go back to the difference between cost per watt produced and sold. The cost per watt produced in the 2025 guide of $0.20, it seems like you're tracking ahead of your Analyst Day expectations from about a year-and-a-half ago or so. I'm just trying to think about -- I think at the Analyst Day, you laid out 2026 targets as well. Just kind of how to think about that following this year, if you think you're still on track to outperform that number? And then the cost per watt sold, the extra $0.04, a lot of that, in my opinion, seems to be fairly transitory. Looking out to 2026, not looking for formal guidance, but do you think 2026 cost per watt produced and cost per watt sold look close? Thank you.

Yeah. So if you go back down a produced basis, I think we were somewhere in the mid-19, high-19. So it's potentially slightly over this year, but it's pretty close to where we expected to be back at the time the Analyst Day. And we do have some challenges on the produce side looking in this year, there certainly weren't -- we wouldn't have been aware of the time of the Analyst Day. So you go back to making more products from India coming into the US. If you look at some of the tariff impacts we're seeing, there's a significant piece in the cost of watt produced related to curtailing about a gigawatt of production in our Southeast Asia factories. So it's not enough of a curtailment that comes into a period cost of underutilization, but it does mean that our cost of watt produced is higher. So you see lower absorption of the fixed costs going into there as you've got lower throughput. So on the produced side, we're relatively close to where we expect it to be. What you're seeing is a significant expansion on the period costs relative to where we thought we would be. If you look at that, roughly $0.04 at period costs across 19 gigawatts, that's $750 million or so. The sales rate historically for us has been around $0.02 a watt, so somewhere in the range of $400 million roughly of that is sales rate. That's about where the historical numbers would have been. If you look at, we've got about $55 million of ramp costs that we talked about specifically that's included in that number. So a little bit of warranty. It's pretty small. And then you've got other period expenses that happens. So we have some overhead that sits above the line versus an OpEx, some scraps and LCM, some ongoing hedging variance and cycle count, all that noise that generally sits in that bucket. But the big piece that has changed significantly is around warehousing. So if you go back to the Analyst Day, we didn't have a significant assumption of warehousing. If you look at this year, that warehouse number is going to be close to $250 million. And I think a significant portion of that, as you said, maybe transitory. I'd say maybe because we do have a strategy of bringing product into the US through distribution centers. We generally find that's the right approach in terms of optimizing across the fleet. But we are finding ourselves right now storing significant volume of products in inventory because of the back-end loading that we saw in 2024. Some of the shipment delays that we saw at the end of '24 as a function of customers looking to move things around and associated with the warranty issue that we talked about. And then as you get into 2025, we still have some of that lagging issue from the warranty happening. We have the fact that there was undersold volume from 2024 coming into 2025. And then we talked about in 2025, the shift rights that customers can deploy even within the year to try and push product back, which means that as we produce across the year on a relatively consistent rate at our factories, if we're not shipping at that same consistent rate with storing inventory, then that's creating a lagging effect. So the sales freight piece is pretty standard. The residual piece within that cost what period cost is not unusual. The big change here is around warehousing.

Yeah, the only I'll just add just a little bit just in terms of the core CPW because as we already highlighted, I mean the tariffs now in terms of things that have changed a little bit, right? So now I've got tariffs on my aluminum that wasn't anticipated before. So that is close to a $30 million impact on our core costs. The other one just to highlight is that our glass cost has continued to be challenging, right, as we continue to scale and bring facilities out of, basically, they were decommissioned back into production. There's been incremental costs associated with that we're scaling up our US operations has become a little bit of a headwind. And just to put it in perspective, just so people understand, I mean, we're pushing close to almost 20% of the market for glass that's produced in the US. So we're a large consumer, but we're also taking assets that more brownfield than greenfield in some cases are driving a little bit higher costs on glass than we had anticipated. And then there's also a glass commodity driver that we're starting to see higher natural gas prices. I still think they're within the range of what we had anticipated in higher framework agreements with our glass suppliers, but to the extent that we see commodity cost pressure on natural gas, we could see a little bit of pressure on our glass costs as well. So another component to take in consideration.