AI · SEMICONDUCTOR · FACT CHECK

SKD WIRE

EN WIRE · 2026-10-28 12:11

OpenAI Moves to Purchase Google AI Chips, Reducing Reliance on Nvidia

Mostly True

OpenAI Moves to Purchase Google AI Chips, Reducing Reliance on Nvidia
Image source: 조사 출처

OpenAI is pushing to purchase AI semiconductors from Google, a move that would lower its dependence on Nvidia chips, according to the claim examined by this desk's fact-checking review, which drew on 11 primary sources and found the core assertion largely accurate.

원문 주장 (KR)
오픈AI, 구글 AI 반도체 구매 추진…엔비디아 의존도 낮춘다

A Shift in AI Chip Procurement

The claim states that OpenAI is seeking to buy Google's AI semiconductors as part of an effort to diversify its hardware supply. Nvidia currently dominates the market for AI accelerators, and companies developing large language models have been looking for alternatives to secure computing capacity and manage costs.

Google designs its own AI chips, and making them available to OpenAI would mark a notable change in how the two companies — long positioned as competitors in AI — interact on infrastructure.

What the Sources Show

The fact-checking review examined 11 primary sources covering the procurement plans. The reporting on the proposed purchases was found to be consistent with the available evidence, supporting the claim that OpenAI is pursuing Google AI semiconductors to reduce its reliance on Nvidia.

Uncertainty remains over the scale and timing of any purchases, as well as the terms under which Google's chips might be supplied to OpenAI. Neither company's commercial arrangements have been fully disclosed, and the available sources do not establish final agreement details.

Implications for the Chip Market

A deal of this kind would reshape the competitive landscape of the AI semiconductor market, where Nvidia has held a commanding position. For chip suppliers and foundry partners, including Korean players such as Samsung and SK hynix, diversified demand from AI developers could influence future orders for advanced chips and high-bandwidth memory used in AI systems.

The overall finding of the review was that the claim is Mostly True.

Sources — primary documents (11)
  1. https://www.theinformation.com/articles/openai-longtime-nvidia-customer-plans-buy-googles-ai-chips
  2. https://openai.com/careers/search/?q=TPU
  3. https://it.chosun.com/news/articleView.html?idxno=2023092172950
  4. https://www.googlecloudpresscorner.com/2026-10-27-Oracle-Brings-Google-TPUs-to-Oracle-Cloud-Infrastructure
  5. https://cloud.google.com/blog/products/compute/ironwood-tpus-and-new-axion-based-vms-for-your-ai-workloads
  6. https://cloud.google.com/blog/products/ai-machine-learning/google-cloud-brings-openai-models-to-customers
  7. https://nvidianews.nvidia.com/news/openai-and-nvidia-announce-strategic-partnership-to-deploy-10gw-of-nvidia-systems
  8. https://www.cnbc.com/2026/08/17/nvidia-financing-open-ai-data-center-ohio.html
  9. https://www.cnbc.com/2025/11/07/googles-decade-long-bet-on-tpus-companys-secret-weapon-in-ai-race.html
  10. https://storage.googleapis.com/gweb-cloudblog-publish/images/Ironwood.max-2200x2200.jpg
  11. https://storage.googleapis.com/gweb-cloudblog-publish/images/3_WZEo7he.max-2000x2000.png

KR: /news/20261028-9c718a · 판정: 대체로 사실