- Nvidia is pitching its upcoming Vera CPUs to Chinese customers while announcing availability soon
- Move into data center CPUs pits Nvidia against traditional rivals Intel and AMD, which currently control the majority of the market in China and other regions
- Nvidia’s move comes at a time when the Chinese government continues to push for home-grown chip solutions and US chip control has effectively reduced its share to 0% of a lucrative Chinese data center market
Nvidia is apparently pushing to win over Chinese customers for what its CEO views as the next multibillion-dollar frontier for the company: data center CPUs.
The firm has spent the past two years watching the world’s second-largest chip market effectively carve it out with a mix of consumer-grade chips and homegrown solutions such as Huawei’s Ascend offering, backed by a Chinese government push for confidence.
While Chinese officials have held the line with soft barriers—no official restrictions on Nvidia’s chip exports to China exist on the mainland—Nvidia is apparently aiming to reset the relationship when it comes to its Vera CPUs for the data center.
Why CPUs and why now to Nvidia?
Nvidia’s Vera CPU is more than just another competitor in the market. It threatens to change the existing status quo, where Intel and AMD chips dominate the market, by adopting an AI-first approach to its design.
Nvidia touts Vera as a CPU that is up to 1.8x faster than current x86 CPUs from Intel and AMD in certain workloads, offers 4x the memory bandwidth and delivers up to a 50% increase in performance over traditional rack-scale CPUs.
According to a Pakinomist report, at least one major unnamed Chinese cloud company intends to buy 300 servers, each containing 2 Vera CPUs. The CPUs themselves are estimated to cost upwards of $20,000 before bulk discounts kick in.
Whether that results in an order, however, remains to be seen. Chinese regulators appear to be making a stronger push for self-sufficiency in their chip sector, prompting many of its AI startups and giants to opt for local chip options, such as Huawei’s Ascend and T-Head’s Hanguang.
Nvidia’s salvation may come from an unexpected place, however, as the battle lines may be different this time: the same AI export controls that crushed its Chinese business may work in its favor now.
Not only are CPUs significantly less regulated by US export regulations, but the Chinese market is also being squeezed by server CPUs, with Intel pushing delivery times to as much as 6 months in some cases, even as AMD noted that the CPU market remains tight, with demand exceeding supply.
If Nvidia can navigate past the politics and software advantage of the established x86 architecture, as well as the ecosystem built around it, it could carve out an important part of the lucrative Chinese data center market, even without the stickiness of CUDA that makes its GPUs so desirable in this segment.
Whether that’s enough to overturn the obvious ecosystem advantage both established chipmakers (Intel and AMD) have, as well as the advantage domestic champion Huawei enjoys in terms of government-level backing, remains to be seen.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds.
