Navitas Semiconductor Q326 Earnings - NAVITAS SEMICONDUCTOR CORP

Financial Performance:

• Q3 2026 total revenue reached $14.0 million, hitting the high end of guidance and representing 33% sequential and 38% year-over-year growth; this marks the first YoY growth since 2024.
• High-power markets drove growth, increasing >50% sequentially and growing at triple-digit rates YoY, effectively replacing declining mobile/low-end consumer segments which are now insignificant.
• Non-GAAP gross margin expanded 50 bps sequentially to 40.0%, up 130 bps YoY, driven by favorable revenue mix and scale; this was the third consecutive quarter of margin expansion.
• Operating expenses increased to $16.5 million (up ~6% sequentially) to fund targeted R&D and acquisition integration; loss from operations narrowed to $10.9 million from $11.4 million in Q2.
• Net loss per share was $0.03, improving from $0.04 in Q2 and $0.05 YoY.
• Cash and cash equivalents stood at $531.5 million (pre-CLAROS payment); pro forma cash post-$126M CLAROS payment is approximately $405 million. The company carries zero debt.
• Inventory increased to $32.4 million (including $11.2M long-term) to build TSMC wafer buffers and support anticipated AI infrastructure demand.
• Full-year 2026 revenue is expected to grow approximately 7% YoY, with high-power growth significantly offsetting the exit of the mobile business.

Guidance and Future Outlook:

• Q4 2026 revenue guidance is $16.0 million (midpoint), representing 14% sequential growth and 119% YoY growth, marking the fourth consecutive quarter of double-digit sequential growth.
• Q4 non-GAAP gross margin is expected to expand to 40.5% (±100 bps), extending the margin expansion trend.
• Q4 non-GAAP operating expenses are projected at $25.0 million (±$1M), reflecting the inclusion of CLAROS overhead and incremental R&D investments.
• AI infrastructure revenue is projected to exceed 40% of total revenue in Q4 and is expected to triple in calendar 2027.
• Management anticipates double-digit sequential growth to accelerate in 2027–2029 as multiple architectural inflection points (800V, native 800V, grid modernization) converge.
• Break-even analysis suggests a potential path to profitability with ~50% blended gross margin (including IVR impact) and operating leverage, targeting a low-60% revenue break-even point over the medium term.

Business Segments and Product Lines:

• High-Power Pivot: The strategic shift from mobile/low-end consumer to high-power markets is effectively complete by year-end, with high-power segments representing nearly 100% of revenue.
• AI Infrastructure: Identified as the "North Star," encompassing AI data centers and grid infrastructure; growth is driven by power efficiency constraints where every saved watt enables more compute.
• Claros Acquisition: The October acquisition of Claros provides a complete Vertical Power Delivery (VPD) and Embedded Power Delivery (EPD) IVR stack, enabling power delivery directly to the XPU with modularity and no external capacitors/inductors.
• Technology Portfolio: Combines Ultra-High Voltage SiC (grid), High-Voltage SiC/GaN (AC/DC & DC/DC), and Claros IVR to create a full "grid-to-XPU" platform.
• Product Milestones:
  • Collaboration with NVIDIA on 800V to 12V power delivery boards.
  • Licensing of Genesys technology to MagnetShift for SiC adoption in Asia.
  • Partnership with Microchip for 800V reference designs.
  • Selected by U.S. government for next-gen 10kV SiPAR semiconductors.
  • First shipments of Gen 5 GANFAST technology with GlobalFoundries; strategic "Integrated Fabless Manufacturing" (IFM) model established.

Market and Competitive Landscape:

• Market Trends: Power is now the primary constraint on AI compute; efficient power delivery is critical to scaling capacity within power envelopes. The industry is transitioning from 48V to 800V architectures.
• Architectural Inflection Points:
  1. Pre-800V: Silicon replacement in AC/DC >3kW (SiC traction).
  2. Sidecar Rack: Rollout of ±400V/800V DC-DC and AC/DC systems expected to ramp in 2027.
  3. Native 800V: Required for >500kW racks; expected in production before end of 2027.
  4. Grid Modernization: Long-term growth via solid-state transformers and ultra-high voltage SiC.
• Competitive Position: Navitas differentiates via speed of execution, modularity (Claros), and full-stack ownership (SiC, GaN, IVR). Management claims to be growing faster than the high-power market, targeting >40% share in specific PSU segments and sole-source positions in sidecar applications.
• SAM Expansion: The addition of IVR technology doubles the 2030 Serviceable Addressable Market (SAM) to over $8 billion.

Risks and Challenges:

• Supply Chain: Reliance on TSMC for wafers and potential capacity constraints in the foundry sector, necessitating significant inventory buffering.
• Technology Execution: Success depends on rapid adoption of complex 800V architectures and successful integration of IVR into various XPU designs (GPU, TPU, custom).
• Litigation: Ongoing patent disputes in the industry (referenced but deferred by management to previous statements).
• Customer Concentration: Heavy dependence on hyperscalers and major OEM/ODM partners for design wins and volume ramps.
• Execution Risk: Integrating Claros operations and managing the "Integrated Fabless Manufacturing" model without becoming a pure IDM or traditional fabless entity.

Management Commentary and Tone:

• Tone: Highly confident and optimistic; management describes the "Navitas 2.0" transformation as a reality rather than a plan, citing the successful pivot to high-power and the imminent "take-off" of AI infrastructure growth.
• Strategy Clarity: CEO Chris Alexander emphasized that "power is capacity," positioning the company as an essential enabler for the AI compute boom.
• Operational Discipline: CFO Tanya Stephens highlighted financial discipline, noting OPEX growth is being managed as a fraction of revenue growth to ensure a path to profitability.
• Strategic Shift: The company has moved from "consuming" capital for transformation to deploying capital for scale and capacity acquisition.

Analyst Questions and Answers:

• Growth Visibility: Q&A confirmed backlog extends into mid-2027 with a book-to-bill ratio of 2.0 in Q3; growth in 2027 is fueled by inflection points 1 and 2, with inflection 3 (native 800V) already showing significant volume orders (e.g., 1 million GaN units).
• Claros Differentiation: Claros was chosen for its "true IVR" stack (3.3V start, 3:1 conversion, modular, no external components) designed specifically for XPU density, unlike competitors adapting computing IVR for AI.
• Margin Structure: Margins are driven by complexity and solution value rather than specific technology (SiC vs. GaN); blended margins are expected to exceed 40-50% as IVR scales and replaces lower-margin silicon.
• IVR Design Cycle: IVR requires early involvement with processor vendors (hyperscalers) to design power delivery directly into substrates or PCBs; modularity allows adaptation to diverse XPU architectures.
• Break-even Model: A conservative model suggests a 50% gross margin and OPEX scaling could lead to break-even around $60M+ in annualized revenue, assuming continued scale and margin expansion.
• Market Share: Management avoids specific blended market share percentages but notes they are gaining share in PSU replacements (2-3 vendor competition) and securing sole-source positions in complex sidecar applications due to speed of execution.
• Claros Pipeline: Claros has a major anchor customer (vertically integrated) ramping in late 2028/2029; the broader pipeline is expanding rapidly across the industry as a "race for density" continues.

Other Key Points:

• IFM Model: The "Integrated Fabless Manufacturing" model involves owning critical IP and equipment while partnering with foundries for capacity, offering a capital-efficient middle ground between IDM and Fabless.
• Full Year 2026 Outlook: The company expects to exit 2026 as a pure high-power entity, with mobile revenue effectively zero.
• Strategic Timing: Management views the current period as the optimal intersection of AI compute growth and the re-architecture of power infrastructure from grid to XPU.
• Long-term Vision: The goal is to maximize the utility of every available megawatt for AI compute through end-to-end power efficiency.