Saturday, July 4, 2026
DarkSubscribe
AI Infrastructure · News & Analysis
HomeChips & HardwareReport
Chips & Hardware · Report

Broadcom powers OpenAI's Jalapeño custom AI accelerator chip, advancing proprietary silicon strategy

Extends custom silicon ecosystem beyond Nvidia, reduces inference cost/latency via optimized ASICs
Trade pressSlicast · June 30, 2026 · US · Source: Google News
importance 90

Broadcom kicked off June with its fiscal Q2 2026 results on June 3. Revenue hit $22.19 billion, up 48% year-over-year, but came in slightly below the $22.27 billion consensus estimate. Earnings per share of $2.44 beat the $2.40 consensus. AI semiconductor revenue surged 143% year-over-year to $10.8 billion, driven by increasing demand for custom AI accelerators and AI networking.

On June 24, Broadcom and OpenAI unveiled Jalapeño—OpenAI's first custom AI inference chip. Broadcom handled the silicon implementation, networking, and connectivity technologies, while OpenAI designed the architecture around its deep understanding of LLM inference needs. The chip went from initial design to manufacturing tape-out in just nine months, potentially the fastest ASIC development cycle ever achieved in high-performance semiconductors.

Early testing shows Jalapeño will deliver performance per watt substantially better than current state-of-the-art alternatives, with initial deployment targeted for the end of 2026 at gigawatt scale with Microsoft and other data center partners.

Despite the earnings per share beat, record AI revenue, and the Jalapeño reveal, Broadcom is down approximately 20% this month. The revenue miss, the strategic pivot away from integrated AI systems, and questions about the stock's premium valuation may all be contributing factors.

At the time of publication, Broadcom shares were trading 2.19% higher at $373.00.

Read the original
Broadcom powers OpenAI's Jalapeño custom AI… · Slicast