ARTICLE AD BOX
ChatGPT-maker OpenAI is reportedly planning to produce its first artificial intelligence (AI) chip next year. According to a Financial Times report, the chip will be co-designed with US semiconductor giant
Broadcom
and is likely to chip in 2026. OpenAI, it says, plans to put the chip to use internally. The move follows the strategy of other tech giants such as Google, Amazon and Meta who have already designed their own chips to power AI-related workloads.The FT report cites Broadcom’s chief executive Hock Tan who earlier this week referred to a mystery new customer committing to $10bn in orders. The deal marks the fourth major client for Broadcom in this executive, he told analysts during the company’s earnings call.Tan said the agreement has boosted Broadcom’s growth outlook, bringing “immediate and fairly substantial demand,” with chip shipments for the customer set to begin next year.However, Broadcom did not disclose customer names. Reports last year suggested the two companies had already begun initial collaboration, though the timeline for large-scale production had been unclear.
Nvidia
still leads the AI hardware market, with big tech companies as its main customers. OpenAI has long been one of Nvidia’s biggest buyers. But OpenAI CEO Sam Altman has repeatedly highlighted the need for more computing power to support products like
ChatGPT
and train large AI models. Last month, he said OpenAI would double its computing capacity “over the next 5 months” to meet demand for its new
GPT-5
model.HSBC analysts recently said Broadcom’s custom chip business could see faster growth than Nvidia’s by 2026. Broadcom’s shares are up more than 30 percent this year and rose nearly 9 percent in pre-market trading in New York on Friday after the earnings update.
Realme 15T First Look: Key Specs and Other Things You Should Know