A CFO who joined six weeks ago volunteered that profitability may be accelerating, while the company burns $16M a quarter on $8.6M of revenue.
Thesis: Navitas is a pre-revenue AI story where the content math is structurally real (2x power = 5x SiC content in 800V PSUs) but every AI dollar claim is forecast, not booked. The gap between $8.6M quarterly revenue and $38M+ breakeven is 4.4x, with $64M annualized burn and a fully replaced executive team. The single event that matters is whether Phase 1 800V design wins convert to backlog by end of 2026. Until then, this is unsizeable.
Verdict: AVOID — Conviction: MEDIUM
Catalyst: Phase 1 800V HVDC PSU ramp starting end of 2026 into early 2027. Management stated final production samples shipped and board-level testing underway, with proof points expected 'very quickly' in backlog. Phase 2 in-tray GaN DC/DC proof points targeted Q1-Q2 2027.
Key Risk: Timeline slippage on the 800V HVDC design wins past 2H 2027 while cash burn continues at $64M annually. Breakeven requires 4.4x current revenue with no near-term bridge. Complete executive team replacement during the ramp adds execution risk on an already tight timeline.
The Tell: New CFO Tonya Stevens, who joined in late March, volunteered unprompted: 'Nothing's changed in our thoughts around profitability. In fact, potentially accelerated a bit.' A CFO with six weeks of tenure claiming profitability acceleration either received a strong internal briefing or is being positioned to deliver a specific message. Combined with Chris stating 'we are a bit ahead of where I think we're gonna be' on the same call.
Friction Level: HIGH_FRICTION — Bull sees beat-and-raise with a structural content step-up (2x power = 5x SiC content) about to hit the P&L. Bear sees pre-revenue AI labeling on a $34M annualized top line with 4.4x revenue gap to breakeven. Fundamental disagreement on whether the 800V HVDC design wins convert on timeline.
2026Q1 · May 5, 2026
3 of 5 readings agreed3 AVOID · 2 HOLDRead the full call
Revenue Down 39% YoY, Called a Return to Growth
Navitas is a pre-revenue AI story where the content math is structurally real (2x power = 5x SiC content in 800V PSUs) but every AI dollar claim is forecast, not booked. The gap between $8.6M quarterly revenue and $38M+ breakeven is 4.4x, with $64M annualized burn and a fully replaced executive team. The single event that matters is whether Phase 1 800V design wins convert to backlog by end of 2026. Until then, this is unsizeable.
New CFO Tonya Stevens, who joined in late March, volunteered unprompted: 'Nothing's changed in our thoughts around profitability. In fact, potentially accelerated a bit.' A CFO with six weeks of tenure claiming profitability acceleration either received a strong internal briefing or is being positioned to deliver a specific message. Combined with Chris stating 'we are a bit ahead of where I think we're gonna be' on the same call.
Total revenue is $8.6M quarterly, down 39% YoY from $14.0M. AI infrastructure content claims ($10-15K/MW GaN, 2.5x content acceleration) are all pre-production forecasts. Design wins are in sampling and qualification stages with no disclosed backlog dollar figures.
Management frames 'return to growth' and 'momentum' while year-over-year revenue declined 39%. Cash fell from $237M to $221M in one quarter ($16M burn). Breakeven requires revenue in the 'high 30s' per quarter versus current $8.6M run rate.
Inventory rose 12% sequentially to $14.9M against a revenue base 39% below prior year, approximately 156 days of sales. Management calls it 'measured investment' but growth in inventory is running ahead of demonstrated demand.
Revenue exceeded high end of guidance at $8.6M, gross margin expanded 30bps to 39.0%, and Q2 guide of $10.0M midpoint is another 16% sequential gain with margins ticking up 25bps. Management commits to sequential growth every quarter through 2026.
Content per megawatt is architecture-locked: GaN at $10-15K/MW in 800V DC/DC, SiC at $5-8K/MW in AC/DC PSUs. Management stated 2x power leads to 5x SiC content. This is pricing power embedded in the architecture, not a price list negotiation.
Chris walked through specific design qualification steps for Phase 1 (final samples shipped, board testing underway, system reliability validation next) and Phase 2 (in-tray GaN proof Q1-Q2 2027) with concrete timeline markers when Quinn pressed repeatedly.
Entire leadership team replaced including CFO who joined mid-quarter in late March. A 4.4x revenue ramp to breakeven must be executed by a completely new bench while simultaneously transitioning foundry from TSMC to GlobalFoundries in 2027.
+60
CONFIDENT"Chris volunteered the 50% AI infra QoQ growth number unprompted, then volunteered Q2 tracking faster, then volunteered profitability may be pulled forward. New CFO stated 'nothing's changed in our thoughts around profitability, in fact potentially accelerated a bit.' Paired with 'far too early to declare victory' and declining to guide profitability dates."
Specific content-per-megawatt figures ($10-15K GaN, $5-8K SiC), quantified sequential growth rates, clear Phase 1/Phase 2 design win timeline, and direct answers from CEO. Offset by zero backlog disclosure, no dollar figure on design wins, and AI infra growth is a combined segment metric management chose to aggregate.
Phase 1 800V HVDC PSU ramp starting end of 2026 into early 2027. Management stated final production samples shipped and board-level testing underway, with proof points expected 'very quickly' in backlog. Phase 2 in-tray GaN DC/DC proof points targeted Q1-Q2 2027.
Timeline slippage on the 800V HVDC design wins past 2H 2027 while cash burn continues at $64M annually. Breakeven requires 4.4x current revenue with no near-term bridge. Complete executive team replacement during the ramp adds execution risk on an already tight timeline.
“We expect revenue to increase sequentially to between $8 million and $8.5 million.”
“Revenue in the first quarter of 2026 exceeded the high end of guidance, increasing 18% sequentially to $8.6 million on a GAAP basis.”
Companies tend to guide low: the median company we cover reports +1.7% against its guidance midpoint, so a small beat is the norm. A transcript can drop part of a guided range, so read both quotes.
Changed more than 71% of calls
Wording compared with the Feb 25, 2026 call, ranked among 97 calls reporting on Q1 2026. A measure of how much the story moved, not a signal about the stock.
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Language change is computed from word frequencies, with no model involved. Descriptive only: none of this feeds the verdict.
Beat and raise with structural mix shift driving margins. Revenue $8.6M beat the high end of guidance, gross margin expanded 30bps to 39.0%, and Q2 guides to $10.0M with another 25bps of margin expansion. AI infrastructure (data center + grid) grew 50% QoQ while the total company grew 18%, meaning the non-AI business is flat or down and the piece that matters is compounding. Management states Q2 AI infra growth accelerates further before the content step-up even hits. The content math is asymmetric: 2x PSU power = 5x SiC content, and moving DC/DC inside the rack is a GaN-only locked-in architecture. $221M cash, zero debt, sequential growth commitment through 2026, and a new CFO volunteering profitability acceleration. The street is modeling linear growth on an exponential content curve.
Revenue is $8.6M, down 39% YoY from $14.0M. The 'return to growth' is sequential against a trough. Every AI content claim ($10-15K/MW GaN, 2.5x content acceleration, Kyber ramp) is pre-production with Phase 1 ramping end of 2026 and Phase 2 proof points not until Q1-Q2 2027. Cash burned $16M in one quarter. Breakeven needs 'high 30s' revenue versus $8.6M current, a 4.4x gap with no bridge. Inventory up 12% against a revenue base 39% below prior year. Fabless with zero control over TSMC capacity, wafer cost, or SiC yields. Complete executive team replacement during the critical ramp. No backlog dollar figure disclosed. This is a pre-revenue AI story on a $34M annualized top line.
AI Label Without Substance
AI infrastructure content claims are entirely forecast. Phase 1 800V ramp starts end of 2026, Phase 2 proof not until Q1-Q2 2027. Today's revenue comes from legacy grid and early industrial. No backlog dollar figures disclosed despite CEO saying proof will appear in backlog.
Say/Do Gap
Management says 'return to growth' while revenue declined 39% YoY. Cash fell $16M in one quarter. Breakeven requires 4.4x current revenue. CEO states profitability is 'maybe a little ahead' while declining to give specific dates.
Inventory Divergence
Inventory rose 12% to $14.9M against revenue 39% below prior year, ~156 days of sales. Management attributes to 'measured investment' but inventory is building ahead of demonstrated purchase orders, not just design wins.
Generation-Skipping Execution Risk
Complete leadership team replacement including CFO who joined mid-quarter. A 4.4x revenue ramp to breakeven plus a foundry transition from TSMC to GlobalFoundries in 2027 must be executed by a fully new bench simultaneously.
AVOID. The content math is real but pre-production. You cannot size a position where breakeven requires 4.4x revenue growth, burn runs $64M annually, and the catalyst is a design win conversion timeline that management themselves place at end of 2026 to Q2 2027. The asymmetry exists on both sides: if 800V converts on time, the stock re-rates hard; if it slips two quarters, the cash burn narrative dominates. Binary risk, unsizeable. Wait for the first backlog print that includes Phase 1 800V dollars.