+31 20 705 1010
EnglishNederlandsFrançaisDeutschEspañol
Models

Openai and broadcom develop jalapeño chip to reduce nvidia dependency

OpenAI is collaborating with Broadcom to design its first in-house artificial intelligence chip, codenamed Jalapeño. This custom silicon aims to optimise inference performance while diversifying the company's hardware supply chain beyond its current reliance on Nvidia GPUs.

Why it matters

This shift indicates a move towards vertically integrated stacks where custom silicon can lower inference costs and improve latency for production models. It suggests that enterprise teams may soon need to manage deployments across increasingly heterogeneous hardware architectures.

Key points

  • The Jalapeño chip is designed specifically for high-efficiency AI inference tasks
  • Broadcom provides the underlying silicon expertise and design integration for the project
  • OpenAI aims to secure dedicated manufacturing capacity through TSMC for production by 2026

Related solutions

Sovereign AI Generative AI Machine Learning
All AI news

More AI news

Models

Moonshot AI introduces Kimi K3.1 Pro and high-speed variant for commercial evaluation

Moonshot AI has officially launched its Kimi K3.1 Pro model alongside a dedicated high-speed version designed for faster processing. Organisations assessing the release are scrutinising operational costs and vendor data handling policies before committing to production deployment.

Models

Kimi K3.1 Pro secures fifth place on the SevenLab AI leaderboard

Kimi K3.1 Pro by developer Kimi has officially entered the top ten on the SevenLab AI leaderboard, securing fifth position overall. The benchmark tracks leading foundation models using a specialised value-adjusted methodology derived from Artificial Analysis evaluation data.

Models

Selecting enterprise hardware for local LLMs prioritises memory capacity and bandwidth

Enterprise teams evaluating hardware for local language model execution must focus on memory capacity and bandwidth over raw compute power. Unified memory architecture allows development workstations to load larger parameter models directly into memory. Planning leaders are assessing these specifications to ensure cost-effective on-device inference.

Ready to build something extraordinary?

15 minutes. No pitch deck. Just a conversation about what AI can do for your team.

Talk directly with our AI specialists

15 min, no strings
No sales pressure
Prototype in 7 days