OpenAI reportedly plans to purchase Google (GOOGL.US) TPU chips, expanding computing power through multiple approaches

According to Jinse Finance APP, a person familiar with the matter revealed that OpenAI intends to purchase artificial intelligence (AI) computing chips from Google, a subsidiary of Alphabet (GOOGL.US). This move is part of the ChatGPT developer's broader efforts to build its AI infrastructure.

This individual, who requested anonymity because the discussions are private, stated that the final terms of any potential deal remain unclear. Both OpenAI and Google declined to comment.

This move reflects OpenAI's multi-pronged strategy to rapidly expand its AI computing capacity. The company had already announced plans to begin using self-developed chips to support its AI models later this year. Currently, a large number of technology companies are developing custom AI chips to break NVIDIA's (NVDA.US) dominance in this field, and OpenAI is among them.

OpenAI's self-developed Jalapeno chip, created in partnership with Broadcom (AVGO.US), was built with the assistance of former Google experts. Richard Ho, who leads OpenAI's internal chip team, previously worked on the development of Google's Tensor Processing Unit (TPU).

Google began developing its TPU more than a decade ago. Initially, these processors were used to accelerate the company's web search engine and improve efficiency. However, very quickly the value of these chips in handling AI tasks was verified, and in recent years they have become an alternative to NVIDIA's highly sought-after graphics processing units (GPU).

Oracle (ORCL.US) also announced on Tuesday that it will offer TPUs through its cloud computing infrastructure. “The next generation of AI will require diversified computing capabilities,” said Google executive Amin Vahdat in a statement announcing the partnership.