AI Processors Quarterly Update Report

Description

As of the end of the quarter, 135 suppliers—from industry giants to VC-funded start-ups—are chasing this $493 billion market. Presently, the market is entering a new phase marked by consolidation, strategic pivots, and differentiated bets on future compute paradigms.

The report contains the following content:

• A comprehensive list of the 135 AI processor suppliers
• Analysis of the shifting U.S.–China competitive landscape
• Profiles and interviews with notable start-ups and established players
• Trends in VC investment and market maturity

The quarterly AI Processors Quarterly Update report is a valuable resource for understanding the AI processor market, providing insight into the capabilities of AI processors in use today and coming in the near future.

Q1’26 AI Processors Quarterly Update Table of Contents

• Introduction 
  • AI processor start-ups: IPO faster than getting acquired
  • AI companies rake in the VC money
  • AI training vs. inferencing vs. memory bandwidth
  • AI processor architectures are evolving almost as fast as LLMs are
    • Would a photonic AI inference processor be better?
    • Would a neuromorphic AI inference processor be better for inference?
  • AI takes over cryptocurrency transaction monitoring
  • Can AI save science?
  • The RISC-V vector extensions for AI
  • The heart of an AI processor: Is it the NPU and GPU?
  • The extraordinary GPU from amusing to supercomputer
  • Local opposition, not ROI, will determine AI’s build-out pace
  • Begging for chips
  • Supply, the growth limiter, not circular investment
  • AI meets the Robotics & Autonomous Systems world
  • Data center build-out in China
• Q1’26
  • Ainekko merges with Veevx after Esperanto pickup
  • Alibaba’s RISC-V CPU runs AI natively
  • Ambiq puts edge AI on a power diet
  • AMD: AI will be everywhere, and is for everyone
  • AMD and Meta announce $60B, 6 GW AI partnership
  • AMD to deliver 6 GW of AIPs to Meta
  • AMD and Samsung partnership on AI memory
  • Apple
  • Axelera real-world edge AI
  • Axelera secures $250M
  • Broadcom’s $100B custom silicon bet
  • CoreLab’s Allen Wu’s open hardware playbook
  • DMP expands its AI IP processor portfolio
  • Fujitsu joins the AIP crowd with an NPU
  • Furiosa—the proof is in the shipping
  • GlobalFoundries carries on the storied ARC processor line
  • Google Coral NPU used by Synaptics
  • Google TPU roadmap & custom silicon scaling
  • Interview with Jake Kochnowicz, Imagination Technologies
  • Innatera seeks to build a neuromorphic ecosystem
  • Interview with Innatera CEO Sumeet Kumar
  • OK, no wedding, but can we still date?
  • Meta’s AI chips epitomize hybrid hyperscalers
  • Microsoft’s Maia 200 inference AIP
  • Mythics—AI processor start-ups, the money continues to flow in
  • Neurophos CIM promises super high performance at a fraction of the power
  • Nvidia may now sell to China
  • Nvidia is (re)-proving the Jevons paradox
  • Nvidia leads and shapes the AI industry
  • Nvidia GTC 2026 keynote
  • Nvidia’s Groq tie-in
  • Nvidia Nemotron 3 Super
  • An AI processor by any other name
  • Positron jumps up to the big-league investment circle
  • Remember optical computing? It’s here
  • Quadric—$30 million more invested in AIP start-ups
  • Interview with Steve Roddy, Chief Marketing Officer, Quadric Inc.
  • Rebellions’ Rebel100 NPU
  • Renesas joins the AI process club
  • Semidynamics announces 3 mm inference tape-out and a full-stack systems roadmap
  • Semifive NPU plus SoC design
  • Sunrise releases next-generation inference GPU, Qiwang S3
  • TI integrates TinyEngine NPU across MCU portfolio
• Appendix
  • Autonomous systems
  • Tensors, vectors, and matrices
  • The who’s who of AI processors