OpenAI unveiled its first custom artificial intelligence (AI) chip developed in partnership with Broadcom, part of a bid by the ChatGPT maker to gain an edge by tailoring hardware to better run its AI products.
The companies said on Wednesday that OpenAI has received the first samples of the chip, called Jalapeno, and is testing how the silicon handles running AI workloads. So far, the accelerator is showing cost savings of roughly 50 per cent compared with typical AI graphics processing units, Broadcom Chief Executive Officer Hock Tan said.
The finalised chips will be integrated into large data centres from OpenAI-backer Microsoft and other partners starting later this year. Tan expects OpenAI and Broadcom should be able to exceed his prior forecast for deploying 1.3 gigawatts’ worth of chips next year. “We like to think we can do better because there is a lot of demand,” he said.
Though OpenAI continues to rely heavily on chips from market leader Nvidia, it has worked to broaden its mix of suppliers, and has struck multibillion-dollar deals with chipmakers like Advanced Micro Devices and Cerebras Systems.
Qualcomm to buy Modular for $3.9 bn to help AI push
Qualcomm said it is acquiring Modular, for about $3.9 billion in stock, adding artificial intelligence (AI) software capabilities that will aid its push into data centres.
The two plan to close the transaction in the second half of 2026, when Qualcomm will give Modular’s owners 19.2 million shares, the San Diego-based chipmaker said Wednesday in a regulatory filing.
“Modular provides an open, AI-native software stack that enables AI to run efficiently across hardware architectures,” Qualcomm said in an accompanying statement.
“Built by engineers who helped create much of today’s AI infrastructure, Modular’s unified platform runs models with industry-leading performance across” multiple types of chips.
Shares of Qualcomm fell as much as 4.6% to $194.77 after markets opened in New York. The stock is still up about 16% so far this year.
Chipmakers of all sizes are looking to seize on technology and talent at the cutting edge of AI. Qualcomm is trying to pad out its product lineup and forge alliances that will make it a competitor to market leader Nvidia Corp.
Modular’s technology offers customers versatility in how they offer AI services, according to Qualcomm, with the ability to process queries via the cloud or on smartphones and other devices. With that kind of setup, software will have to work on multiple chips from different suppliers requiring an open system, the chipmaker said. In contrast, some of Nvidia’s rivals have argued that its offerings are closed off and designed to keep customers locked in.
Modular was founded in 2022 in Silicon Valley by Chris Lattner and Tim Davis, who met at Google and were “frustrated by AI’s fragmented infrastructure,” according to its website.
Qualcomm, which is holding an investor day Wednesday, has been pursuing tuck-in acquisitions in recent years after its deal for NXP Semiconductors NV was scrapped when it didn’t receive regulatory approval. It agreed to buy London-listed semiconductor company Alphawave IP Group Plc last year for about $2.4 billion in cash.
