Off-Prem

PaaS + IaaS

Washington plans to block Chinese access to AI cloud services

Can't buy advanced GPUs... and might be blocked from renting them soon


The US is preparing to escalate its campaign to block Chinese access to AI by including cloud services among the technologies that require government permission before they can be provided to Chinese customers.

According to Wall Street Journal sources, Washington is considering enacting rules that would require US cloud providers such as AWS and Microsoft Azure to obtain explicit approval before allowing customers in the Middle Kingdom access to services for training AI models.

The US Department of Commerce is said to be preparing to unveil these new rules within the coming weeks. However, the department was not immediately available to comment, owing to the July 4 holiday.

If enacted, these rules would make cloud services subject to similar export restrictions that the US has already imposed on technologies used in AI processing, such as the GPU hardware used to accelerate key functions when training AI models.

It emerged last week that the Biden administration was mulling further restrictions on the export to China of advanced chips used for AI processing. It may be that these rules on cloud services are the form the new measures will take, instead of blocks on the sale of less powerful AMD and Nvidia GPUs as was expected, or they may well be in addition to those measures.

These latest restrictions would seem to be aimed at closing off a loophole that has seen Chinese outfits banned from buying advanced compute hardware instead renting access via the cloud, as The Register reported earlier this year.

Whether these measures will have any significant effect remains unanswered. China's cloud market is dominated by local companies, with Alibaba Cloud the lead provider, accounting for 36 percent of total customer spending on cloud infrastructure services during 2022, according to figures from Canalys.

Behind it was Huawei Cloud with 19 percent, Tencent Cloud with 16 percent, and Baidu AI Cloud with 9 percent. All other providers made up the remaining 21 percent.

While AWS and Azure both have cloud regions inside China, the infrastructure for these is operated by Chinese partner companies, as mandated by Beijing, and they are effectively walled gardens separate from either company's global infrastructure.

"I would say that this would have little impact as China spends as much, if not more, on AI than the US does," Omdia chief analyst Roy Illsley told us. "While the infrastructure may not be as advanced and fast as that available on US clouds in China, they can still train models."

Illsley did concede, however, that the restrictions could be a way of preventing foundational AI models developed in the US from being copied or cloned by Chinese agencies.

Such a move could be seen as an escalation in what is now becoming a tit-for-tat exchange of trade restrictions between the two countries. China has already banned memory products from US chipmaker Micron, labeling them a security risk, and just this week announced export restrictions on gallium and germanium, two materials used in semiconductors and other electronics. ®

Send us news
6 Comments

AMD slaps together a silicon sandwich with MI300-series APUs, GPUs to challenge Nvidia’s AI empire

Chips boast 1.3x lead in AI, 1.8x in HPC over Nv's H100

Tech world forms AI Alliance to promote open and responsible AI

Everyone from Linux Foundation to NASA and Intel ... but some big names in AI are MIA

Creating a single AI-generated image needs as much power as charging your smartphone

PLUS: Microsoft to invest £2.5B in UK datacenters to power AI, and more

Don't be fooled: Google faked its Gemini AI voice demo

PLUS: The AI companies that will use AMD's latest GPUs, and more

Trust us, says EU, our AI Act will make AI trustworthy by banning the nasty ones

Big Tech plays the 'this might hurt innovation' card for rules that bar predictive policing, workplace emotion assessments

Nvidia’s China-market H20 chips hit another speed bump

Integration woes delay Nvidia's hopes of maintaining grip on Middle Kingdom

AI chip outfit Graphcore's sales to China hit by US export rules

Company hopes more folks needing AI compute globally will offset pain

The AI everything show continues at AWS: Generate SQL from text, vector search, and more

Invisible watermarks on AI-generated images? Sure. But major tools in the stack matter most

HPE targets enterprises with Nvidia-powered platform for tuning AI

'We feel like enterprises are either going to become AI powered, or they're going to become obsolete'

After bashing Nvidia for ‘arming’ China, Cerebras's backer G42 alarms US govt with suspected Beijing ties

What was it they say about folks in silicon houses?

Mere minority of orgs put GenAI in production after year of hype

Folks are dipping their toes in without a full commitment

EU running in circles trying to get AI Act out the door

Bloc risks missing out on first-to-legislate status if timetable slips