Broadcom's Jalapeño bet vs Nvidia's 14.4× valuation

The article frames Broadcom and Nvidia as two ways to invest in the same AI infrastructure build-out. Broadcom's business centers on helping large customers design custom chips for specific workloads, and that approach is showing results. In fiscal 2026's Q3 (ended Aug. 2), Broadcom's AI semiconductor revenue jumped 221% year over year to $16.7 billion, 56% of total revenue, and management now expects AI chip revenue of $115 billion in fiscal 2027 and $230 billion in fiscal 2028.

Nvidia, by contrast, sells a broader computing platform and is growing from a much larger base. Its revenue surged 106% year over year to $96.2 billion, while Data Center revenue rose 117% year over year to $89 billion in fiscal 2027's Q2 (ended July 26). The article notes that Nvidia's growth is broadening: revenue from AI cloud, industrial, and enterprise customers jumped 138% year over year, faster than the 102% growth in its hyperscale business.

The competitive tension runs through custom chips. Some large AI companies are developing their own accelerators to reduce computing costs and optimize inference workloads. OpenAI's Jalapeño chip, co-developed with Broadcom, is one example — built specifically to run large AI models more efficiently and planned for deployment across very large data centers, according to Reuters. The article says a shift toward custom accelerators could pressure Nvidia's share of inference computing while creating revenue opportunities for Broadcom. The threat appears most direct among hyperscalers, which have the scale and resources to develop custom chips, though Nvidia's broader customer mix and its lower forward valuation — roughly 14.4 times versus about 18.8 times for Broadcom as of Sept. 10 — are presented as counterweights.