Vicor Q4'25 vs Q1'26 — Delta analysis (for critical review)

Purpose: identify what weakened, went silent, or decelerated between Q4 2025 (released 2026.02.19) and Q1 2026 (released 2026.04.21). Not looking for explicit negatives — earnings releases don't contain those. Looking for:
- signals that dropped from the messaging
- metrics that decelerated despite broader acceleration
- framing choices that hide rather than reveal

1. Headline YoY / QoQ comparison

*Derived from $98.0M vs $83.2M Q1'25 product revenue.

The story:

• Backlog +70% QoQ is the single most bullish structural signal — bookings running dramatically ahead of revenue.
• Gross margin compressed 20bp sequentially (55.4% → 55.2%). Small, but the direction is contrary to the "operating leverage" narrative. The YoY +800bp comparison is flattering because Q1 2025 was a depressed base.
• Operating cash flow turned negative due to a $28.6M litigation payment. Ex-payment, OCF would be ~+$24.7M — still lower than Q4's $15.7M? Actually it would be higher, so genuine underlying cash generation is strong. But the headline OCF number doesn't show that.
• CapEx more than doubled QoQ ($5.5M → $12.4M). This is consistent with the second-fab narrative accelerating.
• Q4's $45M one-time patent settlement creates an optical "net income decline" QoQ ($46.5M → $20.7M). On an underlying operating basis, Q1's $16.9M op income is actually healthy vs Q1'25's near-zero.

2. Messaging items Q4'25 highlighted that Q1'26 went silent on

• "Alternate sourcing partnerships for high-current-density 2nd Gen VPD modules" — Q4 featured this explicitly as a strategy to give licensed OEMs and hyperscalers redundant supply access. Q1 silent. Either in progress as planned or stuck — the press release doesn't differentiate.
• BCM6135 Electronics Excellence Award — Q4 talking point. Q1 has no new product-recognition highlight.
• "Gen 5 VPD progressing to Q1 2026 production launch for lead customer" — Q4 said this would happen in Q1. Q1 2026 press release does not explicitly confirm that Gen 5 VPD shipped or achieved first production volume for the lead customer. This is a silent "did it happen?" item that investors will probe on the call.
• "Record IP licensing revenues in 2026" — Q4 framing. Q1 royalty of $15M is consistent but the 2026 projection is not restated.
• "Second ITC investigation targeting NBCs" — Q4 described the second investigation explicitly. Q1 mentions IP enforcement is "having an effect" but doesn't restate the investigation's procedural status (Final Initial Determination? Final Determination pending?).

3. Messaging items that appeared fresh in Q1'26

• Backlog jump to $301M and its attribution across HPC / ATE / IAD.
• CapEx surge associated with second-fab prep.
• CEO insider sale of 40,000 shares (~$9.66M) executed 2026.04.18, three trading days before the earnings release (Defense World, 2026.04.19). Whether this is pre-planned 10b5-1 or discretionary is not specified in the press release and would need to be confirmed via Form 4. Either way, it's a data point that sits uncomfortably next to a "stage set for revenue growth" narrative.

4. The 800V HVDC question — an externality Vicor's press release doesn't address

NVIDIA publicly launched the 800V HVDC ecosystem for AI factories with Texas Instruments, Infineon, and STMicroelectronics as named partners (NVIDIA developer blog, 2025.05; Texas Instruments press, 2026.03). Vicor is not publicly named as a partner in the 800V HVDC ecosystem. Production deployment of 800V HVDC systems is targeted for Kyber rack-scale 2027.

Vicor's Factorized Power Architecture and VPD / GCM products are optimized for 48V-to-load direct conversion, which is excellent for the current 48V-intermediate-bus generation. The 800V HVDC transition reshuffles the competitive deck and the last-inch (at-load) power conversion role in a pure 800V rack architecture is still being defined.

Q1 2026 press release does not address this transition. This is a long-tail strategic consideration — not a Q1 issue — but worth reader awareness alongside the upbeat near-term numbers.

5. Genuinely strong items (so critique doesn't tip into opposite bias)

• $301M backlog represents ~2.7× book-to-bill (orders/revenue). Rare at this scale.
• Royalty revenue reversed trajectory from −7.8% YoY in Q4 to +38.9% YoY in Q1 — the ITC effect is no longer just a one-time settlement, it is producing recurring royalty flow.
• Operating leverage: cost of product +2% on +18% product revenue. Underutilized fab economics are starting to flip positive.
• Operating income swing: Q1 2025 near-zero loss → Q1 2026 $16.9M. Clear underlying profitability improvement independent of patent windfalls.
• Capacity signal: $12.4M CapEx + second-fab planning says the supply side is finally catching up, which is what a backlog-led narrative needs.

6. The two findings that most matter

1. Backlog +70% QoQ (genuinely bullish, structural) vs gross margin −20bp QoQ (minor but notable deceleration). The acceleration is in bookings, not yet in margin expansion. The next two quarters will reveal whether margin catches up.
2. Guidance language has gone quiet between Q4 and Q1. Q4 promised "record bookings, revenues, profitability" in 2026 (Q4 2025 release). Q1 says the backlog is "setting the stage for revenue growth." The "record" framing was not restated. This could just be natural quarterly cadence, or it could be a very subtle lowering of the flag.