TNB Tech Minute: OpenAI And Broadcom Develop a Custom Chip for AI Inference

2026-06-24 · Source: WSJ Tech News Briefing · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Fundamental Awareness, quick

Summary

OpenAI and Broadcom are collaborating to develop Jalapeno, a custom AI chip designed for large language model inference. This initiative aims to enhance efficiency and reduce costs for serving AI systems at scale, forming part of OpenAI's strategy to lessen dependence on external hardware suppliers. Concurrently, Qualcomm has agreed to acquire AI software firm Modular in a \$3.9 billion stock deal. Modular's technology optimizes AI execution across diverse hardware architectures, which Qualcomm intends to use to make AI faster and more cost-effective for its clients, accelerating AI deployment from device to cloud. Separately, Amazon's Zoox unveiled a redesigned robotaxi, optimized for mass production, featuring a lighter interior and refined exterior. Zoox plans to commence large-scale manufacturing at its Hayward, California plant, targeting 100 vehicles weekly, with availability expected later this year, pending regulatory approval.

Key takeaway

For Directors of AI/ML evaluating infrastructure, these developments signal a critical shift towards vertical integration and custom solutions. You should assess your reliance on generic hardware and consider the long-term cost benefits of specialized AI chips and software optimization. This trend suggests that future competitive advantage in AI will increasingly depend on tailored infrastructure, impacting your build-versus-buy decisions for scaling AI services.

Key insights

Major tech firms are vertically integrating AI hardware and software to optimize efficiency and scale across inference and autonomous vehicle production.

Principles

• Vertical integration drives AI efficiency.
• Custom hardware reduces AI operational costs.
• Software optimization enhances AI deployment.

In practice

• Custom chips for LLM inference.
• Software for cross-architecture AI efficiency.
• Robotaxis optimized for mass production.

Topics

• AI Hardware
• Custom AI Chips
• Large Language Models
• AI Inference Optimization
• AI Software Acquisition
• Autonomous Vehicles
• Robotaxi Production

Best for: CTO, VP of Engineering/Data, AI Architect, Tech Journalist, Investor, Director of AI/ML

Related on AIssential

• OpenAI, Broadcom debut custom Jalapeño chip for AI inference — Custom AI inference chips, like OpenAI's Jalapeño, offer performance per watt gains and strategic market differentiation.
• TNB Tech Minute: Meta Unveils Four New Custom-Built Chips for Internal Use — Tech giants are developing custom hardware and navigating regulatory challenges while expanding AI and youth safety features.
• A PhD pivot, a stint in VC and a rumoured Anthropic deal: why Europe is watching Walter Goodwin — AI inference demand is growing exponentially, creating a strategic opportunity for specialized chip development to challenge incumbents.
• Deepseek is designing its own AI chip — Deepseek, like other AI leaders, is developing custom inference chips to reduce external hardware dependency and control costs.
• OpenAI, Broadcom Unveil Chip to Run Models Faster, Cheaper — OpenAI and Broadcom unveiled a chip to run AI models faster and cheaper.
• Meta’s Broadcom Liaison Enters Next AI Phase — Custom AI accelerators and specialized networking are critical for hyperscalers to achieve massive-scale AI compute.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by WSJ Tech News Briefing.