Off-Prem

Alibaba and Tencent clouds see demand for CPUs level off, GPUs accelerate

Lenovo also cashes in on AI demand, without being able to turn it into profit


Demand for cloudy CPUs has levelled out at top Chinese clouds Alibaba and Tencent, whose customers increasingly want GPUs instead.

Both businesses released quarterly results last week. Alibaba Cloud posted six percent year-over-year revenue growth to deliver $3.64 billion for the quarter. Tencent doesn't break out numbers for its hyperscale operation, but attributed growing demand for cloud as a reason its Business Services line item delivered year-over-year growth in the "teens."

Tencent chief strategy officer James Mitchell told investors the rent-a-GPU business is booming – but modestly compared to the US market, as China has fewer well-funded AI startups. Local AI outfits, he observed, may have $1 billion or $2 billion capital, compared to US startups that could have between $10 billion and $90 billion in the kitty.

Mitchell said "percentage growth rates are very fast" for GPU rental, but admitted that's partly because of a low base rate of business. "Some of that demand for renting GPUs in the cloud is incremental, some of it is replacing demands that would otherwise have existed anyway for renting CPUs in the cloud," he added. "And so, while the business in a GPU provision is doing very well, the business of CPU processing is more flat because the incremental demand is for GPU, not CPU."

Alibaba Cloud offered investors similar analysis.

"There's very, very robust demand among our customers for AI and AI-relevant products," CEO Eddie Wu told investors, adding that "that demand is still far from being satisfied."

Wu predicted double-digit growth for Alibaba Cloud in in the second half of 2024 and noted "probably most of that growth will be driven by AI products."

"If you look at the industry as a whole, demand for CPU-based traditional cloud computing is relatively limited, where most of the growth is now focused on GPU-based AI product development," the CEO added.

Neither Chinese cloud mentioned difficulties obtaining GPUs.

Nor did Lenovo in its quarterly earnings call, which saw its Infrastructure Solutions Group (ISG) post record revenue of $3.2 billion – up 65 percent year-on-year – thanks in part to AI creating demand for both conventional servers and GPU-equipped machines.

Investors were told of "buoyant investments in AI infrastructure and an improved supply of AI GPUs" helping Lenovo to grow.

Profit remained elusive, though: ISG posted a $37 million loss, which was at least better than the $60.4 million of red ink from the same quarter last year and the $97 million loss last quarter. On the earnings call, financial analysts were keen to understand how a 65 percent revenue jump didn't result in a profit.

ISG senior vice president Vladimir Rozanovich promised "deeper engagement with that customer base to drive further profit enhancement for ISG."

"But we're also going to optimize our business model, continuing to look at improvements in simplifying portfolio, improving operations, and making sure that we are putting the right R&D into our portfolio to ensure we are ready for this hybrid AI era," he added.

Which sounds like jobs and/or products could be on the chopping block.

Lenovo was in good health overall. Q1 revenue of $15.45 billion represented 20 percent year-on-year growth, and the company's overall $859.6 million profit was a jump from the $643.1 million recorded in the same period last year. ®

Send us news
Post a comment

Buying a PC for local AI? These are the specs that actually matter

If you guessed TOPS and FLOPS, that's only half right

Canadian artist wants Anthropic AI lawsuit corrected

Tim Boucher objects to the mischaracterization of his work in authors' copyright claim

AI firms propose 'personhood credentials' … to fight AI

It's going to take more than CAPTCHA to prove you're real

AMD's Victor Peng: AI thirst for power underscores the need for efficient silicon

Moore's Law may be running out of steam, but there are still knobs to turn and levers to pull

Copper's reach is shrinking so Broadcom is strapping optics directly to GPUs

What good is going fast if you can't get past the next rack?

DoJ reportedly advances Nvidia antitrust probe

Uncle Sam apparently worried GPU giant may be punishing customers who shop around

The future of AI/ML depends on the reality of today – and it's not pretty

The return of Windows Recall is more than a bad flashback

Nvidia admits Blackwell defect, but Jensen Huang pledges Q4 shipments as promised

The setback won't stop us from banking billions, CFO insists

Dell's all-in bet on AI pays off in latest earnings

The term was mentioned over 140 times during the earnings call

AI's thirst for water is alarming, but may solve itself

Its energy addiction, on the other hand, only seems to get worse

Nvidia's growth slows to a mere 122 percent but it’s still topping expectations

Still growing in China, ramping Hopper prods and predicting Blackwell billions soon

Benchmarks show even an old Nvidia RTX 3090 is enough to serve LLMs to thousands

For 100 concurrent users, the card delivered 12.88 tokens per second—just slightly faster than average human reading speed