{"id":38561,"date":"2025-07-01T11:11:36","date_gmt":"2025-07-01T03:11:36","guid":{"rendered":"https:\/\/www.1ai.net\/?p=38561"},"modified":"2025-07-01T11:11:36","modified_gmt":"2025-07-01T03:11:36","slug":"openai%ef%bc%9a%e6%ad%a3%e5%af%b9%e8%b0%b7%e6%ad%8c-tpu-%e8%bf%9b%e8%a1%8c%e6%97%a9%e6%9c%9f%e6%b5%8b%e8%af%95%ef%bc%8c%e6%9a%82%e6%97%a0%e5%a4%a7%e8%a7%84%e6%a8%a1%e9%83%a8%e7%bd%b2%e8%ae%a1%e5%88%92","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/38561.html","title":{"rendered":"OpenAI: Early testing of Google's TPUs underway, no plans for large-scale deployment"},"content":{"rendered":"

July 1 News.OpenAI<\/a> The spokesperson responded on June 30 local time to previous reports that the company had leasedGoogle<\/a> TPU<\/a> The report clarified that the company is conducting early testing of Google TPUs and has no plans for large-scale deployment.<\/p>\n

\"U.S.<\/p>\n

OpenAI currently relies heavily on NVIDIA's AI GPUs, and at AMD's Advanced AI 2025 event Sam Altman announced that he would be using AMD's MI-series GPU gas pedals. If Google's TPUs are added to its computing hardware \"arsenal\", it's likely that the company's TPUs will be used in the next few years.Helping OpenAI Reduce Operating Expenses<\/strong>.<\/p>\n

but,AI hardware still has a way to go between lab testing and large-scale commercial deployment<\/strong>, in between, it needs to go through the adaptation of the software stack and the tuning of the model, which will consume a not-so-short period of time.<\/p>\n

In addition to buying and renting, OpenAI is also exploring self-developed AI chips. The design of its first chip is expected to be completed this year, using TSMC's 3nm process with HBM memory.<\/p>","protected":false},"excerpt":{"rendered":"

On 1 July, according to information received, on 30 June, OpenAI spokesperson local time clarified, in response to a previous report on the company's renting Google TPU, that the company was conducting early testing of Google TPU without plans for large-scale deployment. OpenAI is currently highly dependent on AI GPU in Inverda, and Sam Altman announced at the AMD Advanced AI 2025 event that he would use the MID ' s MI series GPU accelerator. If Google's TPU is added to its calculator hardware's ammunition depot, it will help OpenAI to reduce operating expenses. However, AI hardware still has a distance between laboratory testing and large commercial deployment<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[190,7090,281],"collection":[],"class_list":["post-38561","post","type-post","status-publish","format-standard","hentry","category-news","tag-openai","tag-tpu","tag-281"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/38561","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=38561"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/38561\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=38561"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=38561"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=38561"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=38561"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}