Jed Dorsheimer

Jed Dorsheimer asked how the Pentagon's $29.2 billion spending appropriation, including a new submarine, compares to BWXT's expectations and if there were any surprises in the budget allocation. He also questioned BWXT's capital allocation strategy on the commercial side, considering its involvement in CANDU, AP1000, and various SMRs, asking about growth pace, regional support, and resource allocation.

Answer

CEO Rex Geveden stated that the Pentagon's appropriation does not influence BWXT's business, as their programs are funded through different lines, and the shipbuilding schedule remains unchanged. Regarding commercial capital allocation, he noted potential capacity constraints at the Cambridge facility within a couple of years, leading to a high emphasis on acquiring or building new U.S. capacity (e.g., at Mount Vernon). He also sees opportunities in Europe for SMRs, with investment depending on localization demands. CFO Mike Fitzgerald added that investments also target technologies to improve factory throughput through operational excellence, not just expanding physical footprint.

Jed Dorsheimer inquired about potential supply constraints in the radiopharma business for the Lutetium-177 precursor (Ytterbium-176) during the quarter, and expectations for radiopharma growth.

Answer

Rex Geveden, President and CEO, stated they were not supply-constrained for Ytterbium-176, being far downstream in the process. He expects Lutetium growth to continue to accelerate, leading to higher future demand. Jed Dorsheimer also asked if an RFP win for a Rolls-Royce SMR or AP1000 would primarily drive backlog without significantly contributing to 2026 growth, implying that most reactor-side growth in 2026 would come from Bruce and OPG in Canada. Rex Geveden and Mike Fitzgerald confirmed this, noting such scope is not in the 2026 forecast and would be additive. Rex emphasized a low-risk revenue outlook for 2026, with a focus on driving margins.

Jed Dorsheimer inquired about the memory market, specifically the new high current system order from a North American manufacturer, and the anticipated capacity expansion ramp-up, as well as the percentage of the silicon carbide installed base that has converted from 150mm to 200mm wafers and the remaining upgrade potential in the CS&I business.

Answer

Russell Low, President and CEO, explained that DRAM demand is driven by AI, currently limited by clean room space, with significant capacity increases expected from new fabs in 2027. He noted that only a small portion of the silicon carbide installed base has transitioned to 200mm, indicating ample opportunity for high-value upgrades, including system enhancements like the Purion Power Series Plus within the same footprint.

Jed Dorsheimer asked about the high current system order from a U.S. memory manufacturer, inquiring if it was a second-sourcing opportunity and how to interpret the Q4 to Q1 transition given anticipated stronger memory performance. He also questioned the percentage of the silicon carbide installed base that has converted from 150mm to 200mm wafers and the remaining market potential for such upgrades in the CS&I business.

Answer

Russell Low (President and CEO, Axcelis Technologies) explained that memory demand, particularly DRAM, is driven by AI and currently limited by cleanroom space, anticipating significant DRAM growth in 2026 and momentum into 2027 as new fabs open. Regarding silicon carbide, he stated that only a small portion of the installed base has converted to 200mm, primarily leaders outside China, indicating ample opportunity for high-value field upgrades, especially with the Purion Power Series+ platform. Jamie Coogan (EVP and CFO, Axcelis Technologies) added that the Q4 upgrade included an upgrade on top of an upgrade, reducing cost of ownership within the same footprint.

Jed Dorsheimer of William Blair sought clarification on the implant technology mix (high-energy vs. high-current) for silicon carbide in China versus the rest of the world, and asked about other applications driving high-energy demand.

Answer

President & CEO Russell Low detailed that planar SiC primarily uses high and medium current tools, whereas advanced trench and superjunction devices require high-energy implanters. He added that while high-energy is used in image sensors and memory, the implant density in silicon carbide is a significant current driver for the technology.

Jed Dorsheimer of William Blair questioned the geographic mix of the silicon carbide business, particularly China's exposure, the portion of backlog secured by deposits, and the specific technology differentiators in medium/high current tools for memory and logic applications.

Answer

President and CEO Russell Low stated that Axcelis has global exposure in silicon carbide and does not believe the business will be impacted by recent export rules, which target advanced logic and memory. EVP and CFO James Coogan explained that customer deposits are typically required for new entrants, representing a smaller portion of the total backlog. Low added that for advanced logic, Axcelis is targeting new applications, and for memory, its tools are 'battle-hardened' and optimized for high-volume applications.

Jed Dorsheimer inquired about Centrus Energy's ability to achieve its End-of-a-Kind cost target with less than the previously assumed 3.5 million SWU capacity, seeking more framework on the relative capacity required.

Answer

Amir Vexler, President and CEO, confirmed that the End-of-a-Kind cost will be achieved well before the 3 million SWUs, emphasizing this as a remarkably important announcement and a significant achievement for the company. He refrained from disclosing full cost details.

Jed Dorsheimer inquired about achieving nth-of-a-kind costs, specifically if it can be reached with less than the previously assumed 3.5 million SWU capacity, and requested more framework on the relative capacity needed for this target.

Answer

Amir Vexler, President and CEO, confirmed that nth-of-a-kind costs would be achieved well before 3 million SWUs, highlighting this as a remarkably important announcement and a significant achievement for a new market entrant. He refrained from disclosing full cost details.

Jed Dorsheimer of William Blair followed up on the L1 debt refinancing, asking for the potential scope of annual interest savings Wolfspeed anticipates from addressing this debt. He then inquired about the Siler City facility and the 300mm virtual lens opportunities, seeking a framework for the timing of utilization should the AR/VR market ramp up.

Answer

Gregor van Issum, CFO of Wolfspeed, stated that the exact amount of interest reduction from L1 refinancing would depend on the instruments used and the size of the initial steps, indicating it would be a material but not a one-go transaction. Robert Feurle, CEO of Wolfspeed, explained that Siler City's production would scale with demand, which is dependent on customer adoption and architectural choices in new applications like AR/VR. He highlighted that the factories are built, and the focus is on driving customer engagement and diversification to leverage the existing CapEx and facilities.

Jed Dorsheimer followed up on the potential interest savings from addressing the L1 debt, asking for a scope of annual savings (e.g., $50M-$150M). He then questioned the timing of Siler City fab utilization, particularly how potential AR/VR opportunities leveraging 300mm wafers might accelerate its ramp.

Answer

CFO Gregor van Issum stated that the exact amount of interest reduction from L1 refinancing would depend on the instrument and size of the first step, indicating it would be a material but not a one-go transaction. CEO Robert Feurle explained that Siler City's production would scale with demand and customer adoption of the technology, emphasizing that the facilities and CapEx are already in place. He noted that customer architectural choices and qualifications are key drivers, highlighting the importance of diversifying the customer base and understanding end applications.

Jed Dorsheimer sought to confirm his back-of-the-envelope calculation that implied Mohawk Valley is operating at approximately a 20% gross margin, based on the provided financial details.

Answer

CFO Neill Reynolds declined to confirm the specific margin figure but affirmed the underlying premise, stating, "the profitability coming out of Mohawk Valley is significantly better than what we're seeing at 150-millimeter in Durham." He attributed this to Mohawk Valley's status as a highly automated, state-of-the-art facility.

Jed Dorsheimer with William Blair asked about Oklo's 14-gigawatt backlog, inquiring if it has changed, the breakdown between utility and hyperscaler customers, and the nature of ongoing discussions. He also questioned if the Atomic Alchemy subsidiary and its fuel recycling plans have influenced supply chain risk discussions with customers. Finally, he asked if isotope production from EBR fuel could occur before the Viper reactor is operational, comparing it to Candu reactor capabilities.

Answer

CFO Craig Bealmear confirmed the 14 GW backlog, primarily hyperscaler and data center customers, with ongoing term sheet exchanges and discussions for prepayments. He noted Atomic Alchemy's progress on its pilot program and lab-scale facility, expecting single-million-dollar revenue from the latter in H1 next year, though these discussions involve different counterparties. CEO Jacob DeWitte explained that near-term isotope recovery from EBR fuel is possible without a reactor, but the Viper reactor (mid-2026) will unlock significant production, especially for thermal neutron isotopes, while fast reactors (Aurora, Pluto) will enable unique fast neutron isotope production.

Jed Dorsheimer from William Blair asked for an update on the 14 GW backlog, seeking color on the breakdown of customer discussions between utilities and hyperscalers. He also questioned whether the Atomic Alchemy initiative and fuel recycling capabilities have accelerated supply chain risk discussions with customers. Finally, he asked if actinide tailoring for isotope production could be done with EBR or if it requires the Viper reactor, comparing it to Candu's capabilities.

Answer

Craig Bealmear, CFO of Oklo, confirmed the 14 GW backlog is still predominantly data center and hyperscaler customers, with potential for growth. He noted active term sheet exchanges and progress on prepayments for power or fuel. Jacob DeWitte, CEO of Oklo, clarified that Atomic Alchemy discussions are with different counterparties, and while lab-scale facilities can generate early revenue, reactors like Viper unlock significant performance differentials for isotope production. He emphasized Viper's versatility for thermal neutron isotopes and the unique capabilities of fast reactors for certain isotopes.

Jed Dorsheimer of William Blair inquired about the pipeline's mix between behind-the-meter and front-of-the-meter opportunities and asked which specific radioisotopes Oklo has a competitive advantage in producing for the radiopharma market.

Answer

Co-Founder, CEO & Director Jacob Dewitte stated that while behind-the-meter applications are compelling, near-term deployments may favor grid-tied or hybrid solutions. For radioisotopes, he explained that the strategy includes extracting isotopes like strontium-90 from recycled fuel and using Atomic Alchemy's cost-effective reactor design to produce known isotopes more cheaply, aiming to create an ecosystem of isotopic abundance.

Jed Dorsheimer of William Blair & Company, L.L.C. inquired about the drivers behind the quarterly gross margin decline, the source of management's increased optimism regarding the AI processor market, and the nature of discussions with semiconductor foundries.

Answer

President & CEO Gayn Erickson and CFO Chris Siu attributed the gross margin pressure to an unfavorable product mix, the temporary cost burden of the InCal facility prior to its consolidation, and lower manufacturing utilization. Erickson explained the heightened AI optimism stems from successfully proving their wafer-level burn-in technology with a lead customer, which has generated significant inbound interest from other major players who now see a clear, cost-effective path to improving yield on expensive, complex chiplet-based processors.

Jed Dorsheimer of William Blair & Company requested details on the backlog composition, the expected revenue balance between systems and WaferPaks, the attach rate for automated aligners, the mix of 150mm versus 200mm silicon carbide wafers, and the reasons new customers are now opting for FOX-XP systems directly, bypassing the typical FOX-NP evaluation phase.

Answer

Executive Gayn Erickson declined to provide a detailed backlog breakdown for competitive reasons but confirmed a significant portion is from the new Incal acquisition. He expects a more balanced revenue mix between systems and WaferPaks going forward, with growth in non-SiC applications. Erickson estimated over half of future systems will include automated aligners and noted that while the FOX platform is wafer-size agnostic, they are testing an increasing number of 200mm wafers. He attributed customers moving directly to the FOX-XP to their increased comfort with the technology, which has been bolstered by Aehr's extensive in-house wafer testing services.

Jed Dorsheimer inquired about the potential for further cost reductions in the core business, regional demand trends for various wafer sizes, and the market impact of the GT Advanced Technologies bankruptcy.

Answer

President and CEO Bill Weissman responded that another 15% or more in cost reductions over the next 6-12 months is not unreasonable. He noted that 2-inch demand is almost exclusively from China, while 4-inch is primarily Taiwan and Korea. Regarding the GT bankruptcy, Weissman stated they are not seeing direct market implications but noted continued interest from smartphone manufacturers in sapphire faceplates.