Industry Update March 26, 2026 Philip Shen Managing Director, Sr. Research Analyst pshen@roth.com (949) 720-7198 Matthew Ingraham, CFA Associate, Research mingraham@roth.com (949) 720-7149 Oscar Chim Associate, Research ochim2@roth.com (949) 720-5781 Sales (800) 933-6830, Trading (203) 861-9060 Click here to request a meeting with an analyst Sustainability 38th Annual ROTH Conference Takeaways – ITC Extension Perspective, Tax Equity, 232/FEOC Guide Later, Permitting Could Accelerate, AI/DC, & More We recently concluded our 38th Annual ROTH Conference where we hosted ~15 renewables and EPC companies, including ENPH, FSLR, FTCI, Luminous Robotics (private), MWH, NXT, Otovo ASA (AB: OTOVO, NC), PRIM, RUN, SEDG, SHLS, SPWR (NC), TE, TYGO, and ZEO (NC). We also hosted a number of industry experts/contacts including Nick Iacovella, Executive Vice President, Head of Public Affairs & Communications at Coalition for a Prosperous America (CPA, private), Neil Chatterjee, former FERC Chairman, Keegan Moyer, Principal at Energy Strategies (private), and Ryan Galeria, Chief Development Officer at Copia Power (private). Lastly, we hosted 14 fireside chats, five of which were recorded. See the replay links for our panel discussion and fireside chats that were recorded inside this note. See all ROTH conference fireside chat replays here. ITC Extension Perspective – Delaying Safe Harbor May Be Risky In our latest Snapshot, we published that the Democrats introduced an ITC extension bill. We wrote, "This past week, a group of Democrats introduced a clean electricity bill that would extend the ITC. The upcoming midterms could represent a significant positive upside catalyst for our coverage universe assuming we get a split Congress." While November could serve as a positive catalyst for stocks, we highlight that this would likely be in anticipation of what a potential split Congress might do with the ITC. It is important to note that an ITC extension bill likely would NOT move this year. One of our contacts shared, this Democrat Clean Electricity Bill is "simply a messaging bill and the Democrats setting up for the possibility that they retake a chamber in Congress and that they will attempt to negotiate with the Trump White House. If Trump does do a deal with them, it will be most likely on something small, and the Democrats will continue to introduce bills and pass them through the House to try and set up for 2028, a similar situation when they controlled everything, and they were able to pass the IRA." A potential ITC extension bill appears to already be "distorting behavior. In particular, some developers are delaying safe harbor decisions in anticipation of a potential extension." A DC contact indicated, "That is a mistake. Making safe harbor decisions based on an expectation that this legislation will pass in 2026 is flawed. There is no certainty around timing or outcome [of an ITC extension]. Delaying projects on that assumption is not a sound strategy —and any resulting decline in demand for domestic, compliant products is counterproductive to both near-term execution and longer-term supply chain objectives." Tax Equity Update A highly credible source from our conference affirmed that there may be four major banks that may have paused some 48E investments including 2025 safe harbored projects. This is in line with what we've written in the past. See here and here for some recent notes we've written on the topic. We remind you that the market is pivoting to Tier 2/regional banks and alternative sources of capital and tax credit buyers, which results in higher costs. Our high level view remains that the FEOC/PFE uncertainty may be worse for resi solar as it is a flow business that may already depend on the 48E with FEOC/PFE requirements, while utility scale solar has a longer lead time business that gives the segment more time/ways to navigate. For utility scale solar, the metric we're following is whether project timelines from LNTP to NTP, which are typically 5-8 months, are taking longer than expected. If these timelines start extending in the coming year due to the tax equity pause on 48E, then this tax equity situation could adversely impact 2027 and 2028 utility sale solar volumes. For now, we think it is fine for utility scale solar as developers pivot to more expensive money, but this will be important to track in the coming quarters. Important Disclosures & Regulation AC Certification(s) are located on page 7 of this report. Roth Capital Partners, LLC | 888 San Clemente Drive | Newport Beach CA 92660 | 949 720 5700 | Member FINRA/SIPC Sustainability - Philip Shen March 26, 2026 Section 232, FEOC, Permitting Updates The Section 232 on polysilicon may get pushed into May or June as the Administration focuses on replacing IEEPA Section 301 tariffs as well as updated Section 232s on steel and aluminum. Although initial internal targets pointed toward mid-February and then late March, the sequencing of higher-priority trade actions has pushed the timeline out, yet our contacts believe it remains unlikely to slip beyond early summer. We continue to believe that a critical poly 232 factor to follow is the implementation of an import licensing mechanism. We believe licenses could be preferentially granted based on non-FEOC status as well as to trusted manufacturers that already operate substantial US facilities (e.g. FSLR, Qcells). This structured approach deliberately avoids flooding the market with existing warehoused inventory and instead creates a managed environment that incentivizes continued onshoring. FEOC guidance is not expected to come out until the fall/Q4 as Treasury may be prioritizing long-term enforceability over short-term market certainty. Our work from the conference suggests Treasury may be taking its time in order to observe how companies adjust ownership structures, IP licensing arrangements, and tax strategies before finalizing guidance. Treasury may "spend the next 4-6 months watching companies and what they do." Rather than issuing prescriptive entity-based sanctions immediately, Treasury's strategy could be to monitor market behavior to surface potential loopholes, which would enable more comprehensive and robust coverage once the guidance/rules are issued. The approach reflects an understanding that entity-based restrictions have historically proven difficult to enforce, as demonstrated by experience with UFLPA. Our contacts also indicate that Treasury is having consistent conversations with major banks to ensure that systemically important institutions are not inadvertently disrupted. Utility-scale project permitting could accelerate over the coming months, driven by converging pressures that include rising energy prices, hyperscaler demands for immediate grid interconnection, and advocacy from Republican governors. In line with what we have written in the recent past, the administration is expected to favor projects directly supporting AI/datacenter, those located in politically aligned districts, and those demonstrating verifiable domestic content. While public messaging likely continues to emphasize conventional dispatchable resources, the practical outcome is expected to be a quiet but measurable acceleration in approvals for solar and storage facilities. These technologies represent the fastest route to incremental generation capacity at a moment when natural gas and nuclear additions remain years from commercial operation. Recent signals suggest this shift is already underway, with the pace expected to quicken in response to both economic necessities and mounting political incentives to address headline inflation risks tied to energy costs. AI/Datacenter and Power Demand High-voltage transformers, switchgear, and large-frame combustion turbines face multi-year lead times, with domestic supply chains remaining tight and manufacturers reluctant to expand capacity aggressively amid boom-bust concerns. Transmission network upgrades represent another bottleneck, as building or reinforcing 345kV and 500kV infrastructure requires years of planning, permitting, and construction while costs escalate rapidly. Finally, fuel supply and pipeline capacity for natural gas have seen little expansion over the past two decades, raising risks of volatile or higher delivered fuel prices that could undermine the economics of thermal generation. These constraints suggest that headline GW-scale datacenter announcements may outpace actual near-term deliverability and limits how quickly and at what scale new hyperscaler demand can be served. Hyperscalers could accelerate access to power if they commit to bringing, building, or funding new generation resources alongside their loads. In negotiations with utilities, large loads that offer to deliver surplus capacity—often through solar + storage, batteries, or thermal plants—may unlock faster interconnection and higher served volumes than they ultimately consume. This “growth pays for growth” dynamic is gaining traction, with hyperscalers increasingly pledging to front-load costs for transmission upgrades, system reinforcements, and capacity additions. In practice, developers and hyperscalers could structure deals in which they inject GWs of new resources at a shared substation, allowing the utility to serve the datacenter load while maintaining overall grid balance. Such arrangements could serve as a pragmatic bridge, particularly in markets like Arizona or the broader Southwest, where colocated or co-developed generation helps utilities manage reliability without shifting incremental costs onto existing ratepayers. Utilities could increasingly offer "97% uptime" interruptible or partial-reliability service provided the hyperscaler supplies on-site capacity resources to cover the remaining delta during peak or contingency events. Solar paired with storage, four- hour batteries, demand response, and grid-enhancing technologies could serve as the fastest and most economical bridge to reliable supply, delivering incremental megawatts far more quickly than new combined-cycle gas plants or nuclear facilities. In this model, batteries not only provide capacity accreditation but also enable arbitrage and firming of variable output, while interruptible training loads offer additional flexibility. Over time, as utilities upgrade their systems, these bridge solutions could transition toward full "five- lines" reliability delivered directly by the utility. This approach may prove especially valuable in stressed markets, allowing hyperscalers to begin operations years earlier while market forces and economics continue to favor solar-plus-storage as the cheapest marginal electron for baseload and peaking needs alike. Page 2 of 7 Sustainability - Philip Shen March 26, 2026 A substantial share of announced datacenter load could prove non-viable or significantly delayed, echoing the ~70-80% attrition historically observed in generation interconnection queues. Hyperscalers frequently file interconnection requests in multiple locations to test viability, knowing that most are likely to be withdrawn once power realities, lead times, and costs become clear. Our contacts at the conference estimate that only ~25% of queued positions may translate into operational load within the next five years, with many headline multi-GW hubs remaining conceptual or early-stage. Pure-play or less sophisticated developers may still operate under outdated assumptions that power will simply be available, while even more advanced players face hard constraints around equipment, transmission, and fuel. Greater transparency around contracted power, financing milestones, and actual construction progress could help separate firm demand from placeholder filings and allow utilities and grid operators to allocate resources more effectively. Outlook for Potential FERC Changes and Managing Load Demand We hosted Neil Chatterjee, former FERC Chairman, on a fireside chat. See below for our takeaways. The recent ratepayer protection pledge signed by major hyperscaler CEOs at the White House could carry more practical weight than its non-binding label implies (see here). Although the commitment does not surrender any legal rights, it provides FERC with significant optical leverage. Under Section 206 of the Federal Power Act, the commission could initiate proceedings against PJM or individual utilities to embed concrete safeguards that hold large loads financially accountable for incremental costs they impose on the system. Parties that publicly endorsed the pledge might find it politically difficult to challenge such measures in court, even if they retain the technical ability to litigate. This dynamic could translate into enforceable obligations that protect ordinary ratepayers while still enabling the rapid deployment of generation needed to support AI/datacenter. In October 2025, the Department of Energy (DOE) issued an advanced notice of proposed rulemaking that outlines a framework for addressing surging datacenter demand (see here). Under this approach, hyperscalers seeking expedited interconnection would receive faster access to the grid in exchange for agreeing to curtail load during peak periods and covering the costs of necessary system upgrades. The proposal appears to be an "elegant," technically driven solution capable of balancing speed- to-market with grid reliability, but FERC may choose not to convert the notice into a nationwide final rule. Instead, the commission could focus its efforts narrowly on the PJM region, where interconnection pressures and resource-adequacy risks are most acute, while allowing other independent system operators such as MISO, SPP, ERCOT, and CAISO to develop their own tailored processes. 2026 could be the year "antiquated" partisan framing around energy sources gives way to a more pragmatic recognition that reliability and affordability imperatives transcend traditional left/right divides. Policymakers on both sides of the aisle may increasingly stand aside and allow market forces, rather than prescriptive mandates, to determine the optimal mix of resources. This evolution could prove especially consequential for solar + storage, which benefits from improved aggregation rules, virtual power plants, and demand-response capabilities. The year ahead may mark the point at which DERs and grid-enhancing technologies finally receive the policy space they need to scale, driven less by ideology and more by the urgent necessity of keeping lights on and costs manageable for all consumers. Coverage Universe Takeaways FSLR: Section 232 on polysilicon is expected to be released by mid-2026. We believe FSLR's bookings could be light in Q1 and potentially in Q2 as the company waits for the 232 tariff outcome before committing to higher volumes. Management noted that its SEA plants are running at ~20% utilization. The company's strategy for its SEA capacity focuses on multi-year agreements with customers claiming PTCs instead of ITCs due to the lack of need for domestic content. For Series 6 international volume, FSLR is holding firm on a target ASP of ~30c/W as the company is prioritizing long-term agreements vs. chasing low-margin, short-term deals. On its partnership with Oxford PV (private), FSLR shared that it has a development line that can produce 200–300 units of small-form- factor modules per day, and a full-size Series 6 perovskite pilot line is scheduled for early 2027. NXT: We think bookings can remain strong and healthy despite the tax equity pause for now. We believe software continues to represent ~2% of total revenue with attractive software margins. NXT sees its software attach rate growing over time along with other non-tracker products. By integrating eBOS, steel frames, proprietary software, and other products, management sees its bundling offerings could be ~15% cheaper vs. "à la carte" procurement. PRIM: We expect bookings to remain robust, with strong verbal awards in Q1 converting to new contracts through Q2-Q4. The company expects renewables revenue in 2026 to be flat to slightly down vs. a strong 2025, primarily due to ~$500mn of pull- forward from accelerated concurrent project builds in 2025. The typical LNTP to NTP cycle is about ~8 months. While we see risk for this cycle to extend slightly due to the ongoing tax equity related slowdown, thus far, projects have not pushed out and utility scale solar's long lead times give the industry the ability to deal with tax equity pause to find alternative sources of capital. Management reiterated that no project suspensions or cancellations are anticipated. PRIM developed its EBOS solution, Premier PV, three years Page 3 of 7 Sustainability - Philip Shen March 26, 2026 ago, which they deploy on their projects as well as sell to third-party EPCs. We expect this product to gain traction as PRIM opens a $30mn manufacturing facility in Q4'26, which is expected to expand capacity from 1.5GW to 4.5GW. See our fireside chat replay here. MWH: According to the company, MWH-built power plants consistently outperform third-party constructed assets in its O&M portfolio. Management attributes this performance edge to a unique continuous engineering feedback loop, whereby real world O&M data directly informs future project design, SCADA architecture, plant controls, sensor density, and optimization strategies. While 2026 GMs were inline with consensus expectations, upside drivers to margin for the year include scale efficiencies from higher EPC volumes and unforecasted corrective maintenance work such as hail or tornado repairs. Strong project execution could allow the release of 1-3% of margin due to project contingencies. SHLS: The 2026 revenue guide, in our view, feels very conservative. SHLS is down ~30% vs. TAN ~4% since reporting Q4'25 earnings due to the company lowering its GM target from 35-40% to 30-35%. Current levels represent an attractive entry point. However, we believe the strategy behind the lower margins is sound as the company is growing revenue/gross margin dollars and capturing incremental wallet share. This strategy has resulted in a mix shift by increasing sales of long-tail BLA, combiner boxes, and home run solutions vs. the historical mix of 80% high margin BLA solution. Essentially, SHLS can make extra revenue by selling an additional ~500ft of lower margin feeder cable (long-tail BLA) in addition to the higher margin ~130ft BLA trunk. We also see upside to SHLS's $580mn 2026 revenue guide given the FFQ BLAO of $603mn, and the company typically sees ~$50-70mn in book-and-turn business that is not factored into the guide. We believe the difference between the guide and the FFQ BLAO is conservatism around the conversion of the $67mn BESS backlog as the company is new to this market. On the datacenter opportunity, a 1GW datacenter could represent ~$60mn in revenue to SHLS assuming a 1:1 datacenter power to BESS power ratio. FTCI: Management sees a meaningful shift by customers toward tracker supply diversification. The company has been added to 8 of the top 10 EPC Approved Vendor Lists (AVLs) and secured bookings with two of these EPCs in 2025. Management anticipates new bookings from these AVLs throughout 2026. Management highlighted a key differentiator for FTCI’s trackers is delivering ~40% labor savings with an installation rate of ~0.053 man-hours per module, outperforming a top-tier tracker, which had ~0.09 man-hours. Furthermore, the company reaffirmed its target for adjusted EBITDA breakeven at a quarterly revenue run rate of $50–$60mn. On 3/23/26, FTCI successfully resolved its Q4'25 "technical breach" by securing a lender waiver in exchange for a $10mn scheduled debt repayment and the implementation of new covenants. See details on page 55 of FTCI's 10K here. See our fireside chat replay here. TE: There remains a funding gap of ~$375mn for Phase 1 of the G2 cell facility, and the EU portfolio could represent source of cash in the medium term. TE reiterated that the capex for G2 (Phase 1) is ~$425mn. With ~$50mn already secured via customer deposits, the funding gap for Phase 1 stands at ~$375mn. The company is working through all of its options, and we'll learn more on TE's upcoming Q4'25 earnings call, which is scheduled for March 31st. Beyond this core funding, management is positioning its EU portfolio—specifically the Giga Arctic facility in Norway and potentially a permitted land site in Finland—as high-value divestiture targets for datacenter hyperscalers. The Giga facility has already restored ~50MW of power, with a trajectory to reach ~396MW capacity within 1-2 years (see here). ENPH: ENPH noted that adoption of its new PPL&L products via Propel is scaling with >150 installers across four US states. Management emphasized that the IQ9 (GaN) microinverter offers superior performance at a ~10% lower cost compared to the IQ8 and is sold at the same price, allowing the product to be immediately margin accretive. Furthermore, the 5th generation resi battery delivers a ~40% cost reduction and ~50% higher energy density over the 4th generation. In the Netherlands, the sunsetting of net metering and the introduction of export penalties create a ~$2-3bn retrofit TAM. Based on ENPH’s existing ~500k Netherlands installed base, we estimate 2026 retrofit revenue could reach ~$125-250mn (assuming 5-10% conversion at an ASP of ~$5,000 per system). We believe ENPH has the necessary technology to compete in the 800V DC datacenter vertical, leveraging its IQ9/GaN architecture and bi-directional EV charging expertise. With NVDA’s (NC) Rubin Ultra expected to drive major architectural shifts by end of 2027, we believe ENPH may unveil datacenter-specific products within the next 12 months. SEDG: We believe that higher power offerings, efficiencies from the single SKU strategy, US C&I growth, EU recovery, and cost-advantaged US exports represent healthy tailwinds for SEDG. The EU market has started to show positive demand signals for resi solar due to the recent war in Iran with changing energy dynamics leading to higher power prices. EU channel inventory levels have normalized to healthy levels, and the rollout of new products like the Nexis platform are helping SEDG regain momentum and market share. The company's single SKU strategy may be a meaningful driver of margin upside as 45X inverter/optimizer benefits of 11c/W may be recognized on SEDG's inverter nameplate capacity. The latest single SKU inverter has a nameplate of 14.5kW and the new Nexis residential inverter is expected to be 20kW. We note that the average US resi solar system size is 8-10kW. Additionally, in the C&I segment, the inverter is expected to be significantly higher-powered. Management sees the US C&I market expanding with strong underlying demand and rising storage attach rates. RUN: Trading at 0.4x 30-year NAV, we remain bullish on RUN. The company's assets are being marked at 1x or better through its quarterly sales to its large infrastructure partner. Since reporting Q4'25 earnings, RUN stock is down ~36% vs. TAN ~1%, with investors disliking the 2026 cash gen guide miss and the ~40% cut in affiliate volumes. We see three main drivers for the lower cash gen Page 4 of 7 Sustainability - Philip Shen March 26, 2026 outlook: (1) weakness in the tax equity market, where clearing prices have declined potentially into the mid-to-high 80c range vs. prior of low-to-mid 90c; (2) rising equipment costs, particularly for domestic content and FEOC-compliant modules, which have pressured unit economics; and (3) a modest ~5% y/y decline in blended origination volume. We believe the underlying fundamentals for RUN remain strong despite these headwinds and expect RUN to outperform vs. TPO peers as volumes shift to its higher-quality direct business. Otovo ASA (AB: OTOVO, NC): Otovo serves the US/EU resi and C&I markets. Otovo generates revenue through a tiered membership program called OtovoCare, with monthly fees of €9, €29, or €49 corresponding to response times of 48, 24, or 12 hours. Membership includes system monitoring and priority service. Additional streams include discounted repairs (10% off for members), battery and equipment upgrade sales, and in deregulated markets such as Texas, retail power sales and virtual power plant aggregation. Unit economics assume ~$1,400 average annual revenue per residential customer and roughly double that for commercial. The company expects to reach >30,000 customers by early April following the Energy Aid acquisition, a European JV, and a major undisclosed OEM service contract. The target is 275,000 customers by the end of 2028 with 45% GM, 25% net margin, and happy customers (unhappy customers reduce retention and upsell potential). Page 5 of 7 Sustainability - Philip Shen March 26, 2026 Covered companies mentioned in this report: Company Name Ticker Rating Price Enphase Energy, Inc. ENPH BUY $42.53 First Solar, Inc. FSLR BUY $193.51 FTC Solar FTCI BUY $4.50 SOLV Energy, Inc MWH BUY $29.48 Nextpower Inc. NXT BUY $130.42 Primoris Services Corporation PRIM BUY $149.40 Sunrun Inc. RUN BUY $13.04 SolarEdge Technologies, Inc. SEDG NEUTRAL $51.28 Shoals Technologies Group, Inc. SHLS BUY $6.91 T1 Energy Inc. TE BUY $6.68 Tigo Energy, Inc. TYGO BUY $4.34 Page 6 of 7 Sustainability - Philip Shen March 26, 2026 Regulation Analyst Certification ("Reg AC"): The research analyst primarily responsible for the content of this report certifies the following under Reg AC: I hereby certify that all views expressed in this report accurately reflect my personal views about the subject company or companies and its or their securities. I also certify that no part of my compensation was, is or will be, directly or indirectly, related to the specific recommendations or views expressed in this report. For important disclosure information regarding the companies in this summary report, please contact the Director of Research at (800) 678-9147 or write to: ROTH Capital Partners, LLC, Attention: Director of Research, 888 San Clemente Drive, Newport Beach, CA 92660 Disclosures: ROTH makes a market in shares of Enphase Energy, Inc., First Solar, Inc., FTC Solar, Nextpower Inc., Sunrun Inc., SolarEdge Technologies, Inc., Shoals Technologies Group, Inc. and Tigo Energy, Inc. and as such, buys and sells from customers on a principal basis. Within the last twelve months, ROTH Capital Partners, or an affiliate to ROTH Capital Partners, has received compensation for investment banking services from SOLV Energy, Inc and T1 Energy Inc.. Within the last twelve months, ROTH Capital Partners, or an affiliate to ROTH Capital Partners, has managed or co-managed a public offering for SOLV Energy, Inc. Shares of Tigo Energy, Inc. may be subject to the Securities and Exchange Commission's Penny Stock Rules, which may set forth sales practice requirements for certain low-priced securities. Distribution of IB Services Firmwide IB Serv./Past 12 Mos. as of March 24, 2026 Rating Count Percent Count Percent Buy [B] 396 77.65 114 28.79 Neutral [N] 84 16.47 6 7.14 Sell [S] 3 0.59 1 33.33 Under Review [UR] 27 5.29 6 22.22 Our rating system attempts to incorporate industry, company and/or overall market risk and volatility. Consequently, at any given point in time, our investment rating on a stock and its implied price movement may not correspond to the stated 12-month price target. Ratings System Definitions - ROTH Capital employs a rating system based on the following: Buy: A rating, which at the time it is instituted and or reiterated, that indicates an expectation of a total return of at least 10% over the next 12 months. Neutral: A rating, which at the time it is instituted and or reiterated, that indicates an expectation of a total return between negative 10% and 10% over the next 12 months. Sell: A rating, which at the time it is instituted and or reiterated, that indicates an expectation that the price will depreciate by more than 10% over the next 12 months. Under Review [UR]: A rating, which at the time it is instituted and or reiterated, indicates the temporary removal of the prior rating, price target and estimates for the security. Prior rating, price target and estimates should no longer be relied upon for UR-rated securities. Not Covered [NC]: ROTH Capital does not publish research or have an opinion about this security. ROTH Capital Partners, LLC and its affiliates expects to receive or intends to seek compensation for investment banking or other business relationships with the covered companies mentioned in this report in the next three months. The material, information and facts discussed in this report other than the information regarding ROTH Capital Partners, LLC and its affiliates, are from sources believed to be reliable, but are in no way guaranteed to be complete or accurate. This report should not be used as a complete analysis of the company, industry or security discussed in the report. Additional information is available upon request. This is not, however, an offer or solicitation of the securities discussed. Any opinions or estimates in this report are subject to change without notice. An investment in the stock may involve risks and uncertainties that could cause actual results to differ materially from the forward-looking statements. Additionally, an investment in the stock may involve a high degree of risk and may not be suitable for all investors. No part of this report may be reproduced in any form without the express written permission of ROTH. Copyright 2026. Member: FINRA/SIPC. Page 7 of 7