Navitas (NVTS) Q3 2026 Earnings Call Transcript

Navitas reported Q3 2026 revenue of $14.0 million, up 38% year over year and 33% sequentially, with non-GAAP gross margin of 40.0%, 130 basis points above Q3 2025. It guided to Q4 revenue of $15.5 million–$16.5 million, or 119% year-over-year growth at the midpoint, and expects AI infrastructure to exceed 40% of sales by year-end; management said that business is poised to triple in 2027. The company paid approximately $126 million for Claros after quarter-end, leaving pro forma cash of about $405 million; Q3 non-GAAP net loss was $6.9 million.

Analysis

The key underwriting question is conversion, not market size: NVTS must turn design activity and backlog into repeatable shipments while growing far faster than its cost base. At Q4 guidance midpoint, a 40.5% gross margin implies roughly $6.5M of quarterly gross profit against $25M of non-GAAP OpEx—so the near-term growth narrative does not yet equal operating leverage. Management’s illustrative break-even framework requires a substantial revenue ramp and margin gains; verify actual quarterly progress rather than extrapolating the stated AI tripling, whose baseline and customer-level economics were not disclosed.

The strongest second-order risk is that IVR acquisitions across the industry validate the architecture but also intensify competition. Modular designs may help NVTS win multiple processor programs, yet hyperscalers and vertically integrated processor vendors gain leverage to qualify alternatives and dual-source. The 1M-unit order and mid-2027 backlog are encouraging, not equivalent to recognized revenue or durable share. NVDA collaboration is not, by itself, proof of a production award or material NVIDIA revenue. GFS could benefit if NVTS’s foundry model scales, but the call provides no economics to quantify that spillover; MCHP’s SiC license could broaden adoption while also widening the supplier ecosystem.

Near term, expectations may outrun the small revenue base. Over 1–3 months, test Q4 delivery, margin, and Claros-related OpEx. Over 6–18 months, the thesis depends on program conversion, supply execution, and AI power architecture adoption. The $405M pro forma cash balance reduces immediate financing risk, but continued losses and inventory commitments still make execution and working-capital discipline material.

Key Decisions for Investors

• Treat NVTS as a high-beta execution watch, not a clean AI infrastructure proxy. Avoid chasing an earnings-driven gap without valuation and price context; consider a staged position only if subsequent results confirm shipment conversion and margins.
• Track three falsifiers each quarter: revenue versus guidance, non-GAAP gross margin progression, and OpEx growth relative to revenue. Reassess if backlog/book-to-bill weakens, high-power growth fails to translate into sales, or Claros costs rise without customer ramps.
• For the next 1–3 months, verify the composition and cancellation terms of backlog, the timing and revenue recognition of the 1M-unit order, and Claros’s anchor-customer milestones. These are missing inputs—not yet grounds for a separate trade.
• Do not infer a material read-through to NVDA, GFS, or MCHP from the announcements alone. Monitor disclosed production awards, foundry volumes, and licensing economics before trading those names on this call.