This audio is automatically generated. Do us know if you have comments.
Brief of diving:
- Openai cut an agreement with AMD to deploy six Gigawatts of the AI computing capacity in a massive infrastructure thrust presented on Monday. Companies will share technical experience and align product route sheets to several generations of GPU AMD Instinct, according to a joint announcement.
- “This collaboration is an important step in constructing the calculation capacity needed to achieve the full potential of AI,” said Sam Altman, co -founder and CEO of Openai, in the statement. Companies are planning to activate a GPAWatt of the GPU of the AMD Instinct Mi450 series in the second half of next year.
- Openai has invested a lot in the calculation needed to feed large language models through the Stargate infrastructure initiative of $ 500 million. The company agreed to acquire up to 160 million ACTIONS from AMD shares as milestones are made specific to the Monday Agreement, producing billions of income by AMD, said the EVP of the Chip manufacturer, CFO and Treasurer Jean Hu in the announcement.
Divide vision:
As companies pursue evasive returns on the AI generative investments, Big Tech dumps resources in a boom in the building where capacity is measured in gigawatts and investments in the hundreds of billions of dollars.
Last year, data center spending reached almost half a trillion dollar, fed by hyperacle in scale in AI infrastructure, according to Dell’Oro Group Research. Expenditure continued during the first half of this year, increasing investments in the data center in the first quarter and in the first quarter 53% and 43%, respectively, to $ 134 billion and $ 158 billion.
Openai has been an important player everywhere.
The scope and cost of the models builder’s participation in the Stargate Dendorates company in the White House expanded last month, and the company pledged to bring seven online AI capacity gigawatts for the next three years in five places in the data center in collaboration with Oracle and Softbank. At the beginning of the year, Openai and Oracle started working at $ 300 million in construction to deliver additional 4.5 Gigawatts for the next five years.
GPU Coreweave’s cloud supplier is intertwined in the widest constructions with Openai. In the last six months, Openai has almost doubled his commitments to Coreweave’s calculation capacity to $ 22.4 billion, companies were announced on September 25.
The AMD agreement arrives two weeks after Openai entered a strategic alliance with the GPU NVIDIA giant to deploy 10 Gigawatts of AI processing capacity, which will also begin to enter online during the second half of 2026. Nvidia will invest up to $ 100 billion in Openai as part of the alliance.
AI constructions have been an advantage for Nvidia. The company largely overshadowed AMD as the main provider of the GPU, growing its revenue for almost eight years in three years and raised to the maximum of the Gartner’s semiconductor global market share rankings for the first time last year. AMD fell at number 8 last year, compared to 2023.
As the GPU demand increased, AMD invested almost $ 5 billion to acquire the ZT Systems cloud infrastructure provider last year. The chip manufacturer increased revenue by 32% year -on -year to a record of $ 7.7 billion in the 2nd 2025, partly increased by an increase of 14% year -on -year income from the data center segment, which reached $ 3.3 billion.
AMD hopes that Openai collaboration will “allow additional revenue from existing and new customers,” said President and CEO Lisa Su, on Monday, during a virtual press conference. The agreement has the potential to generate more than $ 100 billion in revenue over the coming years, added Su.
