Vicor acquires New Hampshire sites for chip fab expansion

Vicor purchased a 334,000-square-foot building in Merrimack, New Hampshire, plus 54 acres in Hooksett, to build ChiP Fab-2 and Fab-3 with a combined footprint of nearly 1 million square feet. The expansion follows Fab-1 utilization approaching capacity; Fab-2 has a one-year lead time to initial deployment. The added U.S. manufacturing capacity is intended to support AI, hyperscaler and OEM demand while improving domestic sourcing and supply-chain resilience.

Analysis

The strategic value is not incremental square footage alone; it is whether VICR can convert proprietary power-delivery architecture into qualified design wins at hyperscalers and accelerator-platform OEMs before larger analog peers commoditize adjacent solutions. Domestic capacity can improve customer qualification and supply assurance, but it also raises fixed-cost absorption risk: a long construction and ramp cycle can depress gross margin and free cash flow well before revenue materializes. The key near-term read-through is therefore bookings and backlog quality, not management’s capacity narrative.

Over the next 1-3 months, the stock could rerate if AI-related orders demonstrate that demand is constrained by supply rather than limited by customer adoption. Conversely, a capacity announcement without a corresponding uplift in backlog, customer concentration disclosure, or forward revenue guidance is unlikely to support a durable multiple expansion. MPWR and Infineon (IFNNY) are better diversified power-semiconductor alternatives if AI power demand broadens, while VICR carries greater upside—and materially greater execution risk—from a single-technology adoption curve.

The contrarian issue is that vertical power delivery may be technically compelling yet remain tied to platform transition timing controlled by a small set of OEMs and chip ecosystem partners. Construction commitments made ahead of visible volume awards create a 6-18 month risk of underutilized facilities, elevated depreciation, and return-on-invested-capital dilution. Falsification of the bullish thesis would be two consecutive reporting periods without AI/HPC booking acceleration, a lower gross-margin outlook during the ramp, or capex materially exceeding internally funded cash generation.

Key Decisions for Investors

• Do not chase VICR solely on the facility announcement. Place on a 1-3 month catalyst watchlist; initiate only if the next earnings release shows AI/HPC backlog or revenue acceleration plus management ties capacity spending to funded customer programs.
• For a confirmed demand signal, build a long VICR position over 3-6 months with a 3-5% portfolio risk budget and a defined exit on a gross-margin guide-down or evidence that capex is outpacing operating cash flow. The upside case is multiple expansion from credible AI infrastructure exposure; the principal downside is fixed-cost deleveraging during a slow ramp.
• Use MPWR as the lower-beta alternative for broad power-content growth in AI servers and data-center infrastructure. A relative-value expression—long VICR / short MPWR—should only be considered after verified VICR design-win evidence, since MPWR’s broader end-market mix makes it a poor short against an unproven VICR capacity thesis.
• Monitor quarterly capex, depreciation guidance, utilization, backlog conversion, and top-customer concentration. If capacity ramps without corresponding order visibility, avoid VICR and reassess whether the expansion is a return-on-capital headwind rather than an AI revenue catalyst.