Suji Desilva

Suji Desilva asked about the differences between hyperscalers Credo has penetrated at 10%+ levels versus those not yet at that scale. He also inquired about the manufacturing strategy for Active LED Cables (ALCs), specifically if it would be similar to AECs, involving modules versus cables, and any outsourcing.

Answer

President and CEO Bill Brennan explained that each hyperscaler is viewed as a distinct market with unique network architectures, and initial engagements typically lead to broader adoption. He highlighted the quarter's significance in showing a path to a more diversified business across customers, protocols (PCIe), and an expanded Total Addressable Market (TAM) with system-level solutions spanning various connection lengths. For ALCs, Bill Brennan confirmed a manufacturing strategy very similar to AECs, where Credo intends to own the entire stack and take full accountability for the system solution.

Suji Desilva asked for an update on the competitive landscape for AECs and inquired about the demand drivers and timing for Credo's longer-reach, 7-meter rack-to-rack solutions.

Answer

CEO Bill Brennan stated that the competitive environment has not changed meaningfully and that Credo's advantage lies in its system-level ownership, which enables faster innovation and time-to-market. He explained that the demand for longer rack-to-rack AECs is driven by next-generation data center designs using liquid cooling, which allows for higher compute density and creates a need to replace optical links to improve reliability. He cited one public customer, xAI, and a second customer ramping this year.

Suji Desilva inquired about the company's guidance for securing four design wins, specifically asking if these represent four separate OEMs and what factors contribute to the near-term visibility for these wins over the next three quarters. He also asked about the expected penetration of these programs into premium versus mainstream vehicle models and the factors influencing the wide Q4 2025 revenue guidance range.

Answer

Co-Founder and CEO Kobi Marenko confirmed that the four design wins are expected from four different OEMs, noting that decisions are anticipated from at least five OEMs in the next three quarters. He added that programs are expected to initially target premium cars, gradually expanding to non-premium models over time. CFO Karine Pinto-Flomenboim clarified that the Q4 revenue range is primarily influenced by the timing of NRE (Non-Recurring Engineering) program shifts, with earlier customer decisions potentially moving revenues into Q4 2025 rather than 2026.

Suji Desilva asked about Arbe Robotics' guidance for four design wins, specifically if these represent four separate OEMs and what factors contribute to the near-term visibility for the next three quarters. He also inquired about the expected penetration of these customer programs across premium, mainstream, or specific platforms, and the factors influencing the wide Q4 revenue range for the 2025 full-year guidance.

Answer

Co-Founder and CEO Kobi Marenko confirmed the four design wins are expected from four different OEMs, with decisions anticipated from at least five OEMs in the next three quarters. He clarified that programs would initially target premium cars, gradually expanding to non-premium models over time. CFO Karine Pinto-Flomenboim explained that the Q4 revenue variability depends on the timing of NRE (Non-Recurring Engineering) shifts, with earlier customer decisions potentially pushing revenues to the higher end of the guidance range.

Suji Desilva inquired about trends in Electron launch deals, specifically regarding the number of launches, contract length, and customer efforts to extend visibility into 2026. He also asked about the M&A environment, particularly if targets are showing increased willingness to consolidate with larger companies.

Answer

Peter Beck, Founder and CEO, Rocket Lab, observed that customers are generally booking multiple launches at once to secure capacity, and demand for Electron continues to grow. Adam Spice, CFO, Rocket Lab, added that larger bulk buys are more common on the commercial side, and the company aims to secure larger, long-term HASTE deals. Peter Beck noted an increasing trend of consolidation interest among smaller companies, with Rocket Lab becoming a preferred partner for scaling products.

Suji Desilva inquired about how Neutron-related costs will transition from R&D to COGS and impact gross margins, and asked about the strategy for building out the company's payload capabilities, specifically regarding organic versus inorganic growth.

Answer

CFO Adam Spice explained that modeling the cost transition will be challenging due to the shift from R&D for the first flight to COGS for subsequent vehicles, with reusability adding further volatility; ultimately, launch cadence will be the key driver of margins. CEO Sir Peter Beck stated that the payload strategy is a mix of both organic and inorganic efforts. He noted that acquiring best-in-class technology like GEOST is often more efficient for national security payloads that require extensive heritage.

Suji Desilva inquired about the drivers for gross margin improvement, the margin profile of the solar business, and whether the $55 million in new Electron bookings represented an acceleration in demand.

Answer

CFO Adam Spice explained that gross margins are currently similar across both Launch and Space Systems, and noted the solar business is steadily improving as it works through an older, low-margin contract. Both CEO Peter Beck and CFO Adam Spice characterized launch bookings as inherently 'lumpy' but confirmed it was a strong quarter for Electron, supported by a rising average selling price (ASP) that now stands at $8.2 million in the total backlog.

Suji Desilva with Roth Capital inquired about D-Wave's strategy regarding ancillary quantum opportunities such as communications, timing, memory, and security keys, asking if the company remains laser-focused on its core computing opportunity.

Answer

CEO Alan Baratz affirmed that D-Wave is laser-focused on quantum computing. He stated that ancillary areas like quantum networking, sensing, and quantum key distribution are not viewed as central or critical to D-Wave's systems, especially since their systems do not require quantum interconnects for the foreseeable future, unlike other modalities.

Suji Desilva of Roth Capital Partners, LLC asked about the next milestones or key activities investors should watch for to gauge D-Wave's progress in Quantum AI.

Answer

CEO Alan Baratz did not announce specific future milestones but elaborated on the company's strategic direction. He mentioned that D-Wave is actively working to apply its quantum systems to other modern AI approaches, such as transformer and diffusion models, with the goal of creating a comprehensive platform that supports these advanced methods.

Suji DeSilva inquired about the 10 POC customers for PCIe UALink for scale-up, asking if any are pursuing Ethernet-related solutions with Astera Labs, or if the roadmap is solely PCIe to UALink.

Answer

Sanjay Gajendra, President, Co-Founder, and COO, reiterated Astera Labs' focus on PCIe-like deployments and multi-generational engagements. He emphasized the company's belief in open standards and its readiness to support alternate implementations if customer requirements shift. Gajendra highlighted that critical data path and management functionalities, which remain constant regardless of the physical layer, are key to performance and deployment.

Suji DeSilva asked about the factors driving customers towards optical scale-up solutions, specifically comparing the pain points of bandwidth versus XPU density.

Answer

Jitendra Mohan, CEO and Co-founder, stated that customers prefer copper for as long as possible due to its reliability, lower power, and better TCO. He explained that optical deployments will become necessary in the 2028-2029 timeframe as scale-up networks evolve beyond single-rack limitations and copper's reach. Sanjay Gajendra, President, Co-Founder, and COO, addressed the 10 POC customers for PCIe UALink, stating that Astera Labs cannot comment on specific customer Ethernet plans but is heavily engaged in PCIe-like scale-up deployments. He emphasized the importance of data path and management functionality, which remains constant regardless of the physical layer, positioning Astera Labs to support alternate implementations if required.

Suji Desilva requested a framework for comparing retimer content in scale-out versus scale-up applications and asked if the 10 Scorpio X customer engagements imply multiple programs per customer.

Answer

CEO Jitendra Mohan explained that retimer usage depends on signal reach, not just architecture, with scale-up having more links but using retimers selectively on longer connections. President & COO Sanjay Gajendra confirmed the 10+ engagements are with unique customers, and within each customer, Astera is tracking multiple programs at various stages of development and qualification.

Suji Desilva asked CEO Kishore Seendripu about the adoption delays or push-outs for DOCSIS 4.0, seeking details on any technical issues that might be gating larger flagship adoption. He also asked CFO Steve Litchfield for an update on the ongoing arbitration process.

Answer

CEO Kishore Seendripu explained that the DOCSIS 4.0 delays are as expected, primarily due to the extensive network upgrades (nodes, amplifiers) required, leading to stability issues and a slow, incremental deployment process. He conjectured that 'ultra DOCSIS 3.0' (3.1) will see massive deployment, noting that 4.0 is largely a North American phenomenon. CFO Steve Litchfield provided an update on the arbitration, stating it is on track to commence this quarter and extend into next year, with a resolution hoped for in the first half of next year, and the company feels positive about it.

Suji Desilva requested an update on the arbitration process, noting that the time for it to commence was approaching.

Answer

CFO Steve Litchfield stated that arbitration is on track for this quarter but will extend into next year, with resolution hoped for in the first half of next year. He expressed positive feelings about the process.

Suji Desilva questioned MaxLinear's roadmap for Active Electrical Cables (AECs) and asked about the applicability of the Panther storage accelerator in AI infrastructure.

Answer

CEO Kishore Seendripu confirmed they are actively supporting customers with AEC designs, where MaxLinear's low-power technology is a key advantage. He strongly affirmed Panther's value in AI, stating it addresses critical power and latency bottlenecks and is becoming an AI-centric product.

Suji Desilva requested an update on the backlog mix between launch and spacecraft, specifically detailing the Alpha launch backlog split between national security responsive and other missions. He also asked for a roadmap of Blue Ghost's capacity to the Moon, including plans for larger cargo, infrastructure, and eventually manned missions, potentially paired with Elektra.

Answer

CFO Darren Ma stated that while the exact split isn't disclosed, Q2 revenue was majority spacecraft, and the backlog weighting is expected to shift towards spacecraft over time. CEO Jason Kim described a long-term lunar roadmap, emphasizing Firefly's ability to scale technology using carbon composites for larger systems, aiming for multiple annual missions by the end of the decade, including rovers, light terrain vehicles, and infrastructure.

Asked about the trend towards custom ASICs in the premium mobile market in China, the business model for these engagements, the timing and size of the one-time transcoding order, and the future structure of Pixelworks after the Shanghai subsidiary's strategic review.

Answer

Management explained that the custom ASIC trend is driven by the need for differentiation in a competitive market. The business model could be either production or IP/design services depending on the customer. The transcoding revenue is expected in Q4, but its magnitude was not disclosed. It's too early to comment on the company's structure post-strategic review.

Suji Desilva inquired about the trend of Chinese mobile customers preferring custom ASICs, the associated business model, the timing of a one-time transcoding order, and the potential future structure of Pixelworks post-Shanghai subsidiary transaction.

Answer

President and CEO Todd DeBonis explained that Chinese OEMs use custom solutions for differentiation in a competitive market. For ancillary chips, Pixelworks would manage production, while for integrated SOCs, it would be an IP and design services model. CFO Haley Aman confirmed the transcoding order revenue is expected in Q4. DeBonis stated it was too early to comment on the company's structure post-transaction.

Suji Desilva inquired about the revenue level required for the Pixelworks Shanghai subsidiary to achieve profitability, its associated operating expenses, the geographic focus of TrueCut device discussions, and the revenue model for new ASIC design services.

Answer

Executive Todd DeBonis explained that Pixelworks Shanghai's OpEx is expected to be around $7 million to $7.5 million per quarter in the second half of the year. Profitability depends on a mix of revenue from projectors, mobile, IP licensing, and design services. For TrueCut, he clarified the focus is on global premium devices for North American and European markets, and the recently certified brand is not Chinese. Regarding ASIC services, he described a flexible model from partial to full turnkey design, noting a full 12-nanometer SoC project could range from $10 million to $20 million in cost depending on the scope.

Suji Desilva from Roth Capital Partners, LLC asked for clarification on the equipment installation timeline for Q1 2026, the spending schedule for the $60 million in CapEx, details on the company's inorganic growth strategy, and its capacity to handle customized magnet requests for emerging high-volume applications like drones and data centers.

Answer

CEO Joshua Ballard explained that the core production equipment is being commissioned through year-end, with finishing equipment arriving in Q1 2026. CFO Rob Steele confirmed the bulk of the $60 million in CapEx for the first phase will be spent in calendar 2025. Regarding strategy, Mr. Ballard detailed plans to fill supply chain gaps through potential acquisitions, JVs, or investments to build scale and ensure supply. He also noted that the company will phase in capabilities to support diverse customer needs as it scales from 600 to 1,200 tons of capacity.

Suji Desilva from ROTH Capital Partners, LLC asked about the royalty stream from CEVA's flagship smartphone customer, questioning if its contribution would continue to grow into 2026. He also inquired about the key elements enabling CEVA to scale its Edge AI business and the potential opportunities in the data center market.

Answer

CEO Amir Panush declined to provide a 2026 forecast but expressed confidence in the continued success and ramp-up of their technology with the U.S. OEM customer. CFO Yaniv Arieli added that Q4 is historically the strongest quarter for smartphones, independent of this specific customer. Regarding AI, Panush highlighted the scalability of CEVA's hardware IP and its comprehensive software stack as key drivers. He explained that their low-power Edge NPU architecture is now finding applications in cloud inference, creating a new market opportunity.

Suji Desilva from Roth Capital Partners, LLC inquired about the eMotion3d acquisition, asking if it introduces a new software licensing revenue stream for Indy. He also asked about the customer profile for the new Photonics business unit in the quantum sector.

Answer

Co-Founder, CEO & Director Donald McClymont confirmed the eMotion3d acquisition represents Indy's first significant move into generating a meaningful software licensing revenue stream. He also noted that the Photonics unit is engaging with customers in both quantum communications and quantum compute, with some revenue already generated.

Suji Desilva of Roth Capital Partners sought clarification on the company's portfolio review process and inquired about the key competitive advantages of the Stalker UAS that contributed to its selection for the Army's LRR prototype program.

Answer

CFO Jonathan Baliff stated that program reviews are a regular, ongoing process and the significant EAC charge was identified during a standard review in July. CEO Peter Cannito highlighted the Stalker UAS's primary competitive advantages as its best-in-class flight duration and range, which are enabled by proprietary battery technology, as well as its extensive, combat-proven operational history with the U.S. Marine Corps and Special Forces.

Suji Desilva of Roth Capital Partners, LLC asked about Intuitive Machines' competitive advantages in the lunar satellite market and requested details on the highest probability pipeline opportunities and their potential magnitude over the next 6 to 12 months.

Answer

CEO Steve Altemus stated that the company's competitive edge stems from its sole-award Near Space Network (NSNS) contract, flight experience around the moon, and its existing global ground network. He identified the Lunar Terrain Vehicle (LTV) Services contract as the most transformative upcoming award, with a potential value approaching a significant portion of the program's multi-billion dollar ceiling.

Suji Desilva of Roth Capital Partners, LLC asked about the contribution of new products to revenue and their impact on gross margin expansion, and sought details on what enables SiTime to win higher dollar content in data center equipment.

Answer

CEO Rajesh Vashist explained that new products, such as the Elite and Cascade families, carry significantly higher ASPs and are driving growth, particularly in data centers, with 2026 expected to be a significant year for new product revenue. He attributed higher dollar content wins to SiTime's unique ability to provide a complete system solution of natively produced oscillators and clocks. This system-level approach solves complex architectural problems for customers, justifying the increased content value through customized and multiple clock chip additions.

Suji Desilva of Roth Capital Partners, LLC requested an update on the three previously announced automotive OEM design wins, asking about the next key milestones for investors to monitor.

Answer

CEO Gideon Ben-Zvi stated that the wins are with prestigious European OEMs, which is generating positive industry attention and helping to attract other potential customers. He emphasized that while the company is actively leveraging this success, further details on milestones or new engagements can only be disclosed once they become formal.

Suji Desilva from Roth Capital Partners, LLC inquired about customer traction and key milestones for ACM's business outside of mainland China.

Answer

CEO David Wang detailed ongoing evaluations with key customers in the U.S. and Korea for cleaning and plating tools, and noted strong interest from Taiwan for panel-level packaging. He highlighted the new Oregon facility as a key enabler for global expansion. CFO Mark McKechnie added that more tools are scheduled for delivery to the U.S. in Q3.

Suji Desilva of ROTH Capital Partners, LLC inquired about the lingering impact of customer pull-ins from the first half and whether the gross margin decline was primarily due to pricing pressure or factory utilization adjustments.

Answer

CEO Young-Joon Kim stated that the pull-ins, estimated at around $2 million, were largely accounted for in Q2, especially in the TV sector. CFO Shin Young Park confirmed that gross margin pressure stems from both severe pricing competition on older products in China and a consequently lower-than-forecasted factory utilization rate, which led to the softer outlook.

Suji Desilva from Roth Capital Partners inquired about the expected timeframe for a major customer like Daimler Truck to reach its full-scale annual revenue potential and whether another customer could ramp up concurrently.

Answer

CFO Saurabh Sinha stated that a customer typically reaches scale approximately two to three years after its Start of Production (SOP). CEO Soroush Salehian added that Daimler Truck's SOP is next year, with a ramp in 2027, and a major passenger car program is expected to follow about a year later, enabling concurrent scaling. Salehian also highlighted that industrial automation programs are progressing faster and are expected to ramp to production earlier than the automotive timelines.

Suji Desilva asked whether the gross margin could expand above the target range given the mix shift to automotive and enterprise, and also inquired about the R&D efforts for MonTitan firmware, questioning if it will remain a resource challenge.

Answer

President & CEO Wallace Kou stated that gross margin depends on product mix each quarter but expects to meet guidance and see better results in 2026. CFO Jason Tsai explained that as more MonTitan customer projects are completed, the company will build a wider range of firmware capabilities, though turnkey solutions will remain resource-intensive.

Suji Desilva from Roth Capital Partners, LLC asked if the favorable product mix from automotive and enterprise could push gross margins above the target range and questioned if managing MonTitan firmware development would be a persistent resource challenge.

Answer

President and CEO Wallace Kou responded that while some high-end products have better margins, the overall gross margin depends on the quarterly product mix, but he expressed confidence in meeting guidance and seeing better results in 2026. CFO Jason Tsai added that as more MonTitan customers are brought on, the company will build a wider range of firmware capabilities, creating leverage, especially as some customers will only require SDKs.

Suji Desilva from Roth Capital Partners, LLC asked for help in framing the size of the Edge AI server opportunity relative to end devices and inquired about how the competitive landscape in this new market differs from their traditional markets.

Answer

CEO Fermi Wang provided a framework, noting a large installed base of cameras could be upgraded via edge infrastructure boxes that integrate 16-32 cameras, with Ambarella's content per box being in the low three-digit dollar range. VP of Corporate Development Louis Gerhardy described the competitive landscape as new, stating Ambarella competes against general-purpose processors by offering a more efficient solution in terms of performance-per-watt.

Asked about the long-term segment contributors, the competitive advantage offered by the new single-crystal product, and requested an update on the company's ongoing litigation with Qorvo.

Answer

Executives see strong long-term performance from Wi-Fi (due to increased content in Wi-Fi 7) and defense (expected to triple over the next year), with mobile considered an upside to the current model. The single-crystal product provides performance advantages in power handling and harmonics and is a key building block for the advanced XP3F technology. Regarding litigation, Akoustis has filed for partial summary judgment against Qorvo in Delaware, and its own case against Qorvo in Texas is moving forward after a favorable court ruling.

Asked for details on the size of the WiFi 7 revenue opportunity compared to previous WiFi generations and inquired about the customer engagement and revenue ramp timing for the automotive market.

Answer

The WiFi 7 opportunity is expected to be larger than prior generations due to a 50-100% increase in filter content per system, driven by higher performance requirements for MLO. In the automotive market, the company is engaged with over five customers for CV2X and has components on a Tier-1 IC reference design for wBMS, but significant revenue from these new platforms is not expected for another couple of years.

Suji Desilva of ROTH MKM inquired about the 20 active Proof-of-Concept (POC) installations, asking for a breakdown of how many are near the validation or production stage. He also sought details on the production ramp, end market, and timing for the one customer in production, and questioned the visibility into the second-half revenue ramp needed to achieve the 20% annual growth guidance.

Answer

CEO Cesar Johnston clarified that of the 20 POCs, eight are in the more advanced proof-of-validation (POV) phase. He explained the one production customer is a Fortune 100 company in the logistics and warehousing sector, partnered with Thinaer, but could not disclose the name due to an NDA. Johnston confirmed the company's guidance for over 20% year-over-year revenue growth, weighted to the second half, and stated the expectation is for more than one production system to contribute to revenue in Q4, contingent on customer decisions.

Suji Desilva inquired about the 2023 revenue guidance ramp given the Q1 decline, the practical applications and network topology for the new 2-watt PowerBridge, the typical customer investment in a proof-of-concept (POC), and how the company counts its 14 POC installations across customers and sites.

Answer

CEO Cesar Johnston reiterated the company's 20% annual revenue growth guidance, stating that increasing POCs will convert to production revenue. He explained that the new 2-watt PowerBridge offers extended range and can be mixed with 1-watt units depending on application needs. Johnston noted that POCs vary significantly in scale, from single units to hundreds of transmitters, and clarified that the company counts deployed sites, with the growth from 2 to 14 POCs in recent quarters indicating strong strategic traction.