AI data center revenue grew 50% sequentially, but nobody on the call ever said how many dollars that is.
Thesis: The $168M print contains last-time-buy shipments from a fab that no longer exists, and Q3 guides to $140-160M confirming sequential decline. Gross margin is negative 34% with $48M per quarter of underutilization drag and no numeric guidance on improvement. The AI data center story, which the entire bull case rests on, carries no absolute revenue figure. Second sourcing by customers post-Chapter 11 is a structural franchise loss that Toyota partnerships do not offset.
Verdict: SHORT — Conviction: MEDIUM
Catalyst: Long-range plan update in calendar 2026 with long-term financial targets. If it quantifies AI data center revenue and a credible path to positive gross margin, the thesis changes.
Key Risk: Fresh start accounting $70M raw material step-down flowing through future quarters could create a margin inflection illusion that gets misread as demand-driven recovery, triggering a short squeeze.
The Tell: When Rolland asked directly what AI data center revenue consists of today at 50% QoQ growth, Feurle gave a lengthy answer about rack power scaling from 100kW to 600kW, solid-state transformers, and UPS applications. He never provided a dollar figure, a percentage of total revenue, or even a range. The specific question went unanswered.
Friction Level: HIGH_FRICTION — Bull sees AI data center traction and clean restructuring execution as a trough-to-recovery setup. Bear sees declining revenue, negative 34% gross margin, and an unquantified AI narrative on a business burning $43M per quarter.
2026Q2 · Feb 4, 2026
3 of 5 readings agreed3 SHORT · 1 AVOID · 1 HOLDRead the full call
Guided Down 17%, Still Pitching AI With No Revenue Number
The $168M print contains last-time-buy shipments from a fab that no longer exists, and Q3 guides to $140-160M confirming sequential decline. Gross margin is negative 34% with $48M per quarter of underutilization drag and no numeric guidance on improvement. The AI data center story, which the entire bull case rests on, carries no absolute revenue figure. Second sourcing by customers post-Chapter 11 is a structural franchise loss that Toyota partnerships do not offset.
When Rolland asked directly what AI data center revenue consists of today at 50% QoQ growth, Feurle gave a lengthy answer about rack power scaling from 100kW to 600kW, solid-state transformers, and UPS applications. He never provided a dollar figure, a percentage of total revenue, or even a range. The specific question went unanswered.
CEO spent significant time on 800V architectures, solid-state transformers, 300mm wafers for AR/VR, and aerospace. Zero absolute revenue figures given for any of these. The AI data center claim of 50% QoQ growth has no dollar denominator against a $168M base.
Script leads with momentum: AI traction, restructuring complete, four verticals. Guide confirms contraction: $140-160M vs $168M, negative gross margin declined as unguideable, revenue drivers explicitly include pull-in reversal and second sourcing.
Fresh start accounting masks underlying economics. $39M one-time drag this quarter, but the offsetting $70M raw material step-down is deferred over 'several quarters.' CFO declined numeric GM guide citing accounting mechanics.
-30
CAUTIOUS"Feurle repeatedly used hedging language on EV demand: 'muted,' 'hard to predict,' 'dynamic environment.' CFO declined to give a numeric gross margin guide. When asked for AI revenue specifics, CEO pivoted to applications without numbers."
Financials are specific and detailed (negative 34% GM, $43M burn, $48M underutilization). AI data center revenue, the key bull driver, has zero absolute figure attached despite being the headline metric.
Long-range plan update in calendar 2026 with long-term financial targets. If it quantifies AI data center revenue and a credible path to positive gross margin, the thesis changes.
Fresh start accounting $70M raw material step-down flowing through future quarters could create a margin inflection illusion that gets misread as demand-driven recovery, triggering a short squeeze.
“We generated $168 million of total revenue, in line with the midpoint of the guidance range we provided last quarter.”
“During 2026, we expect revenues between $140 million and $160 million.”
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Descriptive only: none of this feeds the verdict.
Restructuring execution is genuinely clean. Durham 150mm shutdown landed a month early, removing $5M of margin drag. OpEx down $200M run-rate. CapEx collapsed from $400M to $31M. $1.3B cash against $600M net debt with interest down 60%. Balance sheet provides roughly 30 quarters of runway at current burn. AI data center revenue growing 50% QoQ with structural driver behind it: 400V to 800V transition and rack power scaling to 600kW by 2029. Physical constraints in legacy silicon cannot solve this. First Western company manufacturing exclusively on 200mm. $70M raw material step-down and $30M quarterly D&A reduction create a mechanical margin bridge nobody is modeling.
Revenue $168M includes last-time-buy pull-in from Durham closure that will never repeat. Q3 guides to $140-160M, a 5-17% sequential decline confirmed by management. Gross margin negative 34%. Adjusted EBITDA negative $82M. Operating cash flow negative $43M. Underutilization is $48M per quarter against a factory base built and idle. Customers are second-sourcing post-Chapter 11, a structural franchise loss. Materials segment facing 'tightening demand environment with increased competition.' No numeric gross margin guide. AI data center revenue has no dollar figure. $700M tax refunds are funding the burn, not growth. Interest rate step-up mid-2026 adds cost as revenue declines.
AI Label Without Substance
Say/Do Gap
Mix Shift Spin
Short medium size. Revenue declining with a mechanical floor at $140M, GM deeply negative with no guidance, and the growth story has no dollar figure attached. $1.3B cash means no bankruptcy risk, so keep the position small and manage for the long-range plan catalyst. Add if GM print in Q3 is worse than minus 25% ex-fresh-start items or if AI revenue gets quantified below $10M.