Navitas’ beat is not the story; the investable debate is whether a smaller mobile company can survive long enough to become a high-power supplier

The thesis from this print is that Navitas is no longer a near-term revenue acceleration story, but it is also not a liquidity-stress story after the November financing. What was priced in was weak Q4 demand and a loss-making transition: the Street expected $6.9 million of revenue and EPS of -$0.05, and the company delivered $7.3 million and -$0.05, which means the only clean beat was revenue at +5.1% while EPS surprise was 0.0%. What actually surprised was not the absolute quarter, because revenue still fell -27.8% QoQ and -59.4% YoY, but the composition and commitment embedded in the call: management said high-power represented the majority of revenue for the first time, mobile will “become insignificant by the end of ’26,” and cash ended at $237 million after approximately $96 million of net proceeds. That combination matters because it separates the equity question from the income-statement question. The stock should not get paid for Q4 demand, which is still deteriorating, but it also should not trade as if the company lacks the capital to execute the pivot through the next few quarters.

That distinction is important because the historical trajectory makes the headline beat look smaller than the transition underneath it. Revenue peaked at $26.1 million in Q4 FY2023, then declined to $18.0 million in Q4 FY2024 and $7.3 million in Q4 FY2025; the latest quarter’s -59.4% YoY decline is the largest YoY contraction in the history table, and it came after Q3 FY2025 had already declined -53.4% YoY. The market likely had weakness priced in given the $6.9 million estimate, but a $7.3 million result does not by itself change the slope of the business, especially with Q1 FY2026 in the history table at $8.6 million, still down -38.7% YoY. The variant view is that investors should stop debating whether the legacy mobile reset is worse than feared and start measuring whether the remaining revenue base can support high-power gross margin while management keeps opex near $15 million. If Q4 was merely a mobile air pocket, the revenue line would need to rebound faster. If Q4 was the beginning of a cleaner high-power mix, then gross margin and design traction become the leading indicators.

The margin evidence supports the pivot more than the revenue evidence does, although it is not yet enough to prove operating leverage. Q4 FY2025 gross margin was 38.1%, recovering from -8.9% in Q3 FY2025 and 16.1% in Q2 FY2025, despite revenue falling from $14.5 million in Q2 FY2025 to $10.1 million in Q3 FY2025 and $7.3 million in Q4 FY2025. That recovery is the central bull data point in the quarter: it says the company can still produce a high-30s margin on a shrunken base after distribution and product clean-up. The conflict is that the quarterly history shows Q1 FY2026 gross margin at -17.4%, while the call key point says Todd Glickman guided first-quarter gross margin to 38.7%, plus or minus 25 basis points. Those are different bases in the data pack, and they cannot be blended. For the investment debate, the company’s own guide is the number to watch because it is management’s forward commitment, while the history-table gross margin reminds us that reported margins have been volatile enough to punish any assumption of a smooth transition. Per CFO Todd Glickman: “Gross margin for the first quarter is expected to be 38.7%, plus or minus 25 basis points.” The wording earns attention because it is a precise margin band around the same high-30s profile as Q4, not a broad qualitative assurance.

The opex and restructuring numbers explain why margin stability alone is not enough. Q4 operating expenses fell from $15.4 million to $14.9 million after a 19% workforce reduction, with SG&A at $6.8 million and R&D at $8.1 million, but loss from operations still widened to $12.1 million from $11.5 million because revenue fell faster than costs. The restructuring was not cosmetic: GAAP results included a $16.6 million restructuring and impairment charge, made up of approximately $10 million of distribution contract terminations, $4 million of fixed asset impairments, and $2 million of workforce reduction expenses. That is the income-statement cost of exiting the old channel and product emphasis. The upside case is that management has already taken the severance, impairment, and distribution-contract pain while keeping R&D above SG&A, which leaves the organization pointed at high-power designs rather than simply shrinking. The downside case is equally numeric: with Q1 operating expenses expected to remain approximately $15 million and Q1 revenue guided to between $8 million and $8.5 million, the company is still structurally loss-making unless revenue scales materially beyond the current range.

The balance sheet changes the risk-reward because it gives the 2027 argument time, but it also raises the bar for evidence. Chris Allexandre said the company completed a private placement of common stock in November with net proceeds of approximately $96 million, contributing to quarterly-end cash of $237 million. That cash position is the reason the equity can be valued on design conversion rather than near-term solvency, even after full-year revenue fell to $45.9 million from $83.3 million and full-year loss from operations was $46 million versus $49.7 million last year. Working capital also moved in the right direction: accounts receivable fell to $3.6 million from $9.8 million in Q3, reducing DSOs to 45 days, and inventory decreased to $13.3 million from $14.7 million. Those numbers matter because a revenue reset accompanied by rising receivables or inventory would imply channel stuffing or stranded stock; instead, the company is shrinking working capital while cutting distribution exposure. The cost is dilution from the common-stock placement, but the benefit is that the next two to four quarters can be evaluated on high-power design wins and margin quality rather than financing risk.

The customer signal is narrow but meaningful, because management put concrete wattage and project counts around computing without naming customers. Allexandre said Navitas has “more than 15 projects in production and approximately twice that number in designing across 170-watt, 200-watt, 250-watt, 240-watt and up to 360 watts with leading global computing companies.” The phrase matters less than the numbers: more than 15 production projects are already outside a pure sampling story, while approximately twice that number in design-in creates the bridge to the 2027 revenue comment. In AI infrastructure, management framed the serviceable addressable market at $3.5 billion by 2030, split roughly 50-50 between GaN and high-voltage SiC, with a combined CAGR of more than 60%. It also gave a content framework of $25,000 to $35,000 per megawatt of total content for Navitas, with about $10,000 to $12,000 outside of data center. The investable question is not whether those market sizes are attractive; it is whether a company doing $7.3 million in Q4 revenue can convert production and design projects into revenue before operating expenses around $15 million consume the evidence.

That same customer signal has second-order implications, but the data pack limits the conclusion because Navitas disclosed no named customers and no named suppliers in the supply-chain section. The absence itself matters: there is no data-supported read-through to a specific customer’s unit build, purchase commitment, or supplier allocation, so the only defensible read is at the competitor level. If high-power was the majority of Navitas revenue at $7.3 million, then the dollar base is still tiny relative to power-discrete peers such as VSH at $839.2 million revenue and DIOD at $405.5 million revenue in their latest reported quarters. However, Navitas’ Q4 gross margin of 38.1% sits above VSH at 21.0% and DIOD at 31.8%, while its revenue YoY of -59.4% is far worse than VSH at +17.3% and DIOD at +22.1%. The comparative implication is precise: Navitas is not taking visible share at the revenue-line level yet, but if its high-power mix can sustain high-30s gross margin, the threat to incumbents is not near-term volume displacement; it is a future content mix where GaN and SiC attach at higher margin once projects move from design to production.

The call delivery reinforced the same split between prepared conviction and Q&A caution. The tone history shows Q1 FY2026 sentiment rose to 0.34 from 0.26 in Q4 FY2025, guidance_tone jumped to 0.66 from 0.27, and prepared_sentiment increased to 0.62 from 0.42. Yet qa_sentiment was unchanged at 0.11, tone_confidence was only 0.38, and qa_evasiveness moved to -10.2 from -40.1, a +29.9 change. That combination says management’s scripted message became much more assertive while answers did not improve in sentiment, which is exactly what one would expect in a transition where the TAM and architecture story is clearer than the revenue timing. The AI framing also became more positive, with ai_optimism moving to 0.36 from 0.25, while uncertainty was roughly stable at 31.0 versus 30.7. In other words, the call sounded more confident on guidance and strategy, but not because Q&A removed the timing risk.

The most important wording on timing came when Allexandre said, “But I think you’ll start to see a significant revenue growth starting ’27.” That is not a Q1 or Q2 revenue promise, and PMs should not trade it as one. It is a calendar marker that pushes the payoff beyond the next reported quarter, which makes interim evidence more important, not less. The company’s own Q1 guidance of revenue between $8 million and $8.5 million implies sequential improvement from the $7.3 million company-reported Q4 call basis, but it does not change the fact that Q4 FY2025 revenue was down -59.4% YoY and Q1 FY2026 in the history table was down -38.7% YoY. The charitable read is that mobile is falling out faster than high-power can enter, while gross margin and cash provide the bridge. The skeptical read is that high-power design activity is still too early to absorb an opex base that remains approximately $15 million per quarter.

Mobile is the fulcrum of that timing risk because management is deliberately letting it fade. Allexandre said, “We expect Mobile to continue going down as a percentage of quarterly revenue and become insignificant by the end of ’26.” That statement is more useful than a generic diversification claim because it gives a deadline and confirms that revenue mix degradation is intentional, not just cyclical weakness. It also means investors should penalize any future quarter where mobile still props up revenue without high-power proof, because the strategic plan depends on exiting mobile relevance by the end of ’26. The high-power product activity is concrete, with accelerated sampling of 100-volt GaN and 650-volt GaN targeted at AI data center, 800-volt HVDC, and 48-volt IBC HV buck architecture. Those are not revenue numbers, but they are specific enough to track against production-project growth and margin behavior. If the company keeps talking architecture without moving the revenue base above the current $8 million to $8.5 million guide, the thesis weakens.

The right way to own or avoid the stock after this print is therefore to focus on proof points rather than the beat. Confirmation next quarter would be Q1 revenue at or above the high end of the guided $8 million to $8.5 million range, gross margin within the 38.7%, plus or minus 25 basis points guide, operating expenses still approximately $15 million, and continued working-capital discipline after accounts receivable reached $3.6 million, DSOs reached 45 days, and inventory reached $13.3 million. The thesis would break if revenue slips below $8 million, if gross margin fails to hold the guided 38.7%, plus or minus 25 basis points range, or if management’s 2027 “significant revenue growth” marker is not supported by an increase from more than 15 production projects and approximately twice that number in design. By the next earnings date, the market should not reward another small revenue beat against a low Street number unless it comes with high-power evidence; the actionable variant perception is that Navitas has bought enough time with $237 million of cash, but only gross margin stability and named or numerically expanded high-power traction can turn that time into equity value.