Apple’s new BFF, Broadcom, reveals that three hyperscalers aim to deploy 1,000,000 GPUs, or XPUs, by 2027; something that will make Nvidia shudder
- Broadcom is rumored to have an ongoing partnership with Apple to help it build its own AI chip
- TikTok parent company ByteDance and OpenAI are also reportedly in the picture
- The move comes as hyperscalers look to reduce their dependence on Nvidia AI chips
Nvidia has driven the generative AI boom to record-breaking revenues and profits over the past two years, and while it remains well ahead of its competitors, the company is facing increasing pressure — not just from rival AMD, but also from hyperscalers that traditionally rely on on Nvidia GPUs, but now want to reduce their dependence on the hardware.
If The next platform notes: “Nvidia’s biggest problem is that its largest customers have large enough IT expenses that they can afford to compete with Nvidia and AMD and design their own XPUs for serial and parallel computing. And if they do, it will be the chip design and manufacturing houses Broadcom and Marvell, which have vast expertise in running chipmaking through the foundries of Taiwan Semiconductor Manufacturing Co, that will benefit.”
In its most recent earnings conference callHock Tan, president and CEO of Broadcom, told investors: “Specific hyperscalers have embarked on their respective journeys to develop their own custom AI accelerators, or XPUs, and to network these XPUs with open and scalable Ethernet connectivity. As you know, we currently have three hyperscale customers who have developed their own multi-generational AI XPU roadmap, which will be deployed at varying rates over the next three years. We believe each of them plans to deploy one million XPU clusters in a single fabric by 2027.”
Getting a fair share
Without naming specific companies, Tan added: “To compound this further, we have been selected by two additional hyperscalers and are in advanced development for their proprietary next-generation AI XPUs.”
Broadcom is widely believed to be working with Google and Meta, and as we previously reported, ByteDance and OpenAI on custom AI chips.
Apple is also thought to be developing its first artificial intelligence server chip, codenamed ‘Baltra’, with Broadcom providing the advanced networking technologies essential for AI processing.
When asked about market share during the Q&A portion of the earnings call, Tan replied: “All we are going to do is capture our fair share. Today we are simply very well positioned and have the best technology, which is very relevant in this area. We have by far one of the best combination technologies out there for creating XPUs and connecting those XPUs together. The silicon technology that makes this possible is something we have in abundance here at Broadcom, which is why we are very well positioned with these three customers of ours.”