Broadcom Inc.

🇺🇸NASDAQ Global Select
Back to all articles
Broadcom Inc. stock analysis AVGO overview
Bullish +60

OpenAI Is Building Its Own AI Chip With Broadcom. Should Nvidia Investors Be Worried?

🚀 OpenAI and Broadcom launched Jalapeño, a custom AI inference chip, as part of a plan to deploy 10 gigawatts of accelerators between late 2026 and 2029.

💰 Broadcom's AI semiconductor revenue surged 143% year-over-year to $10.8 billion in fiscal Q2 2026, with over $30 billion in orders booked that quarter.

📈 Broadcom reaffirmed expectations for AI chip revenue to exceed $100 billion in fiscal 2027, representing a doubling of current annual projections.

🤖 Nvidia's data center revenue reached $75.2 billion in fiscal Q1 2027, demonstrating that it still sells more hardware annually than Broadcom expects from its custom chip business over a full year.

⚙️ Jalapeño is an application-specific inference chip lacking CUDA support, making it less flexible than Nvidia's general-purpose processors for training and broad developer use.

📉 Custom silicon adoption by deep-pocketed buyers like Google and Meta could gradually pressure Nvidia's pricing power and its high gross margins of approximately 75%.

📊 Nvidia trades at about 22 times forward earnings, suggesting the market has already priced in some erosion of its dominance compared to Broadcom's low-30s P/E ratio.

🛡️ Analysts conclude that Nvidia's lead in performance and software scale remains enormous, meaning customer-designed chips will chip away at the edges rather than close the moat.

Bullish Signals
  • Broadcom's AI semiconductor revenue jumped 143% year-over-year to $10.8 billion in fiscal Q2 2026, driven by insatiable demand for custom accelerators and networking.
  • The company booked more than $30 billion of AI orders during the fiscal second quarter of 2026 alone, indicating a robust pipeline for future growth.
  • Broadcom projects its AI chip revenue will exceed $100 billion in fiscal 2027, signaling sustained high-growth momentum in its custom silicon segment.
  • Nvidia's massive scale allows it to sell more AI hardware in a single quarter ($75.2 billion) than Broadcom expects to generate from its entire AI chip business for the year.
Risk Factors
  • Custom inference chips are less flexible than Nvidia's general-purpose processors and do not support CUDA, potentially limiting their ability to compete for training workloads or developer loyalty.
Full Analysis
OpenAI and Broadcom unveiled Jalapeño, OpenAI's first custom AI chip, designed for inference workloads. This marks the beginning of a broader plan to deploy 10 gigawatts of OpenAI-designed accelerators between late 2026 and 2029, signaling a strategic shift where major tech buyers are developing their own silicon to reduce reliance on Nvidia's GPUs. Broadcom's custom-silicon business is experiencing explosive growth, with AI semiconductor revenue jumping 143% year-over-year to $10.8 billion in the fiscal second quarter of 2026. The company booked over $30 billion in AI orders during that quarter alone and projects its AI chip revenue will exceed $100 billion in fiscal 2027, roughly double the current annual expectation. Despite this momentum, Nvidia remains dominant with data center revenue of $75.2 billion in a single quarter (fiscal Q1 2027), selling more hardware annually than Broadcom expects to generate from its custom chip business over an entire year. While custom chips may erode Nvidia's pricing power and market share at the edges, the company retains a massive lead in performance, software ecosystem (CUDA), and scale. The article concludes that while the AI build-out wave is growing at a breakneck pace, Nvidia's moat is narrowing only at the edges rather than closing. The primary near-term risk to Nvidia stock is identified as a potential shift in overall AI sentiment rather than direct competition from customer-designed chips like Jalapeño.