Navitas Semiconductor Corp: Tracking, but proofpoints still pending

Core Conclusion

Navitas is making incremental progress shifting its revenue mix toward high-power data center and grid infrastructure, but the stock’s >100% YTD rally has priced in a favorable 2027 AI power outcome without formal design-win evidence, clear 800V GaN timing, or resolved foundry transition risks. We remain Underweight with a $13.70 price target (22% downside from $17.55).

What the Market is Overlooking

The market discounts revenue growth far beyond what near-term proofpoints justify. At ~50x EV/CY27 sales (vs. SMID-cap GaN peers at lower multiples), the valuation implies near-certainty that Navitas will capture significant share in NVIDIA's 800V racks and other high-power opportunities by 2027. However, company commentary indicates phase one is SiC-led, GaN step-up depends on later 400V/800V architectures, customer qualification, and production allocation. No design wins have been announced. The dual/triple-sourcing norm in power semiconductors makes implied market share highly uncertain.

Evidence Chain

High-power revenue growth is rapid but from a trivial base; material AI DC revenue remains a 2027+ story. Data center + grid infrastructure grew ~50% q/q from Q4 2025 to Q1 2026, yet total Q1 2026 revenue was only $8.6mn, down 38.7% y/y. Q2 2026 guidance of $10mn midpoint modestly beat consensus of $9mn. Company commentary clearly states material data center revenue is not expected until 2027 at the earliest. The growth trajectory is positive, but the absolute revenue level and absence of publicized design wins mean the current valuation implies a step-change that has not yet been demonstrated.

800V GaN adoption timeline is more nuanced and longer than the stock prices. Management framed the near-term opportunity as phase one (50V busbar, top-of-rack) that is more SiC-led. The larger GaN step-up is tied to higher-density in-rack DC/DC converters requiring 800V-class architectures, which depend on customer qualification and production allocation. 400V architectures may move faster, but even those do not guarantee near-term GaN revenue acceleration in AI racks. The market appears to assume a rapid GaN ramp that company guidance does not support.

Intense competition and foundry transition create execution risk the valuation ignores. Power customers typically dual- or triple-source, meaning early GaN wins may not be locked in. On the supply side, the foundry path adds risk: 100V products are currently Powerchip-based; 650V GaN is moving to GlobalFoundries first, then 100V/LV later. Each step requires customer re-validation and carries ramp timing risk. Combined with established competitors (STMicro, Infineon, Renesas) making inroads via acquisitions, Navitas’s competitive position is far from assured.

Valuation is extreme even after a target multiple reduction. The stock trades at ~50x EV/CY27 revenue, down from 60x but still a significant premium to peers. The price target of $13.70 is derived by applying 50x to our revised CY27 revenue estimate of $65.5mn (up from $47.7mn). At the current price of $17.55, downside of ~22% is implied. Yet consensus rating distribution shows 67% Equal-weight and only 11% Underweight, indicating an optimistic view that lacks fundamental backing.

Key Risks to the Underweight Rating

A smartphone market recovery or faster GaN adoption in consumer could pull revenue forward. A surprise design win at a major AI customer (e.g., NVIDIA’s 800V racks) earlier than modeled would change the revenue trajectory. Higher-margin GaN mix could improve gross margins faster if 400V TPU racks ramp sooner. Further multiple expansion driven by AI hype remains possible despite the lack of proofpoints.

Valuation / Trade Implication

We lower our target price to $13.70 by applying a 50x multiple (down from 60x) to our revised CY27 revenue estimate of $65.5mn. The multiple reduction better aligns with peers while retaining a premium for the potential GaN opportunity. Current share price of $17.55 implies ~22% downside. Maintain Underweight. The risk/reward is unfavorable until credible design wins and revenue backlog conversion materialize.

Appendix: Peer Valuation Comparison (EV/CY27 Revenue Multiples)