Joe
Full-Time Trader/Investor, Semiconductor & AI Infrastructure Inflection Points
Vicor's Google renewal could pay 3-4x the original license if discounted to encourage direct Gen2 VPD purchases. Without a discount, doubled NBM payments plus VPD at 4x the original NBM amount could total 6x, as 2027 TPU volumes may double. Roth says Vicor alleges Google received its tailored VPD design in 2018. Gen2 wins beyond Cerebras remain key; Google and AMD look likely to ramp from late 2027. $1.8B revenue in 2028 is a realistic bet. x.com x.comÂ
Cerebras's Oracle fast-inference cloud partnership is good for Vicor because Cerebras uses its Gen2 vertical power delivery. Per @boring_invest, Cerebras is partnering with Oracle; that eases fears Cerebras was failing to deliver after Nvidia's OpenAI deal. Cerebras's first Gen2 order was confirmed on last week's call. Vicor likely owns the point-of-load architecture everyone will use, though nothing is certain. x.com x.comÂ
Broadcom's integrated optics, electronics and packaging could lower custom-chip system costs because separate suppliers defend their margins. At similar valuations, it is preferred to Nvidia and has more room to beat expectations. Google TPU v8 shipments are estimated at 4.0M in 2026 and 9.1M in 2027, assuming Broadcom supplies the remainder after MediaTek's allocation. Anthropic financing adds value only if chip profit exceeds likely support payments and capacity costs. x.com x.comÂ
Google's growing TPU business needs HBM, adding demand for Micron, SK hynix and the memory basket. Google Cloud rose 103% to $30.7B, with backlog reportedly $561B; UBS models more than double TPU shipments in 2027. Micron should be up double digits on this, not down 5%. Samsung has LTAs with 10 major customers and is negotiating with 10 more, targeting 60-70% of capacity under contract. x.com x.comÂ
Vicor's accelerating business and operating leverage make it look like an incredible winner. Revenue and operating margins jumped in the latest quarter; CEO Patrizio Vinciarelli said two additional sites would enable nearly 5x Fab-1's capacity across three fabs, citing demand and industry trends. But QQQ could fall to 660 before a meaningful bounce, so Vicor is not a buy in this tape. x.com x.comÂ
Vicor's AI power-delivery royalties now have evidence of recurring usage, while funded fab expansion offers a path past capacity constraints. Q3 royalties reached $66.6M, including catch-ups; management set a conservative Q4 baseline of at least $43M without new deals, and early orders from the same AI OEM confirm ongoing payments. Backlog reached $406M, and planned Fab 2 and Fab 3 sites could bring capacity to nearly 5x Fab 1, with initial Fab 2 equipment targeted for Q2 2027. x.comÂ
Vicor's patent licensing now earns twice the profit of its factory, making the once-doubted royalty engine its larger profit contributor. CVResearch_ reports $66.6 million of royalties last quarter, versus $21.7 million a year earlier; per its report, the CFO expects at least $43 million next quarter without new licenses. Rough market conditions have battered the stock despite that fundamental improvement. x.com x.comÂ
Vicor stands to capture demand for high-density power delivery as power conversion moves closer to the AI processor, with royalties as early optional upside. On September 30 Vicor raised Q3 growth guidance from over 20% to over 30% because VPD royalties beat expectations, as Delta, Eaton, EPC and others showed 800VDC systems and converters with fewer conversion stages. Vicor expects to finish modernizing FAB1 by year-end, expanding capacity for high-density modules, with FAB2 to follow. x.comÂ
Vicor's vertical power delivery is converting from a technology story into real royalty revenue. Vicor raised Q3 sequential revenue growth guidance again, from 20%+ to 30%+, on higher VPD licensing royalties, its second raise in about nine days. Against Q2 revenue of $143.4M, that implies roughly $186M+ in Q3, and repeated raises suggest AI power architecture demand may be stronger than expected. Q3 earnings should show royalty scale, VPD customer expansion and 2027 growth visibility. x.comÂ
AI's physical buildout looks set to last to 2030 and beyond because financing, memory supply and hyperscaler spending still point to sustained demand. Nebius has $3.5B in financing, IREN $2.4B at 9% and SPCX $40B for GPU equipment. SK hynix expects shortages through at least 2030 and Samsung says HBM4 is sold out; memory companies are making money with no sign of slowing. Google expects 2027 capex to rise significantly. Pullbacks can happen; be cautious and buy quality. x.comÂ
Vicor has one of the cleaner potential re-rating mechanisms in AI hardware, combining product revenue, licensing and royalties with an increasingly strategic position. x.comÂ
Vicor is a way to play AI chips needing more power, through its high-density power delivery for AI GPUs and ASICs. Analysts expect revenue to grow from $453M in 2025 to $1.18B by 2028 and EPS from $2.61 to $8.04, the earnings acceleration a Stage 2 uptrend needs, after a big Stage 1 breakout. Vicor just raised Q3 revenue guidance to over 30% sequential growth. Space plus AI power delivery could make it an interesting long-term story. x.com x.com x.com x.com x.comÂ
