Industry

OpenAI unveils Jalapeño chip with Broadcom to reduce Nvidia reliance

OpenAI announced plans for Jalapeño, a custom inference chip developed with Broadcom, as part of a wider industry trend toward bespoke silicon to lower dependence on Nvidia.

OpenAI unveils Jalapeño chip with Broadcom to reduce Nvidia reliance

AI Newsroom · 26 June 2026 · 2 min read

OpenAI has revealed plans for a custom inference chip called Jalapeño, developed in collaboration with Broadcom. This move is part of a broader pattern in which major tech companies — including Google, Apple and SpaceX — are building proprietary silicon to reduce reliance on a single supplier.

The aim is less about a clean break from existing vendors and more about hedging risk: custom silicon offers greater control, hardware that can be tuned to specific workloads, and potential performance gains similar to those Apple achieved when it moved away from Intel.

The topic was discussed on an episode of TechCrunch’s Equity podcast, hosted by Kirsten Korosec, Anthony Ha and Sean O’Kane. The episode covered several other industry developments:

  • Groq’s $650 million fundraising round was highlighted; the raise was framed as a possible comeback after Nvidia recruited away the company’s top talent.
  • The conversation touched on AI agents and looping behaviors: Boris Cherny, creator of Claude Code, argued that such loops are “just as important and as big a step” as the transition from source code to agents.
  • The episode also asked whether public markets are warming to humanoid robots, noting that Agility Robotics intends to go public via a SPAC.
  • Additionally, it was reported that A24 has received investment from Google DeepMind to develop a new AI toolkit for filmmakers.

Why this matters: building custom chips can lower strategic vulnerability to a single supplier, improve performance and cost-effectiveness for certain workloads, and enable bespoke features. At the same time, this trend does not imply an immediate end to Nvidia’s market dominance; for many companies, in-house silicon is a hedge rather than a full exit from existing ecosystems.

The TechCrunch Equity episode provides further detail and analysis on these developments and highlights several deals and debates that could shape the future of AI hardware.

Key players mentioned in the episode

  • OpenAI
  • Broadcom
  • Nvidia
  • Google
  • Apple
  • SpaceX
  • Groq (noted $650M round)
  • Boris Cherny (Claude Code)
  • Agility Robotics (SPAC plan)
  • A24 and Google DeepMind
  • TechCrunch Equity hosts: Kirsten Korosec, Anthony Ha, Sean O’Kane

Related stories