Huawei Cloud CEO: AI progress demands scale, cloud is essential

Huawei Cloud CEO Zhou Yuefeng delivered a blunt assessment of the artificial intelligence landscape this week: the technology’s progress now demands a scale that most enterprises cannot achieve on their own. Speaking at an industry event, he argued that the sheer computational and data requirements of modern AI have pushed the technology beyond the reach of all but the largest players. The message was clear—companies that want to compete in AI must rethink their infrastructure strategies.

Zhou noted that between 80 and 90 percent of AI innovation now takes place on cloud platforms. This statistic underscores a fundamental shift. A few years ago, many enterprises believed they could build proprietary AI systems in-house, using their own data centers and custom hardware. That approach is increasingly untenable.

The cost of training large language models, for instance, runs into the tens of millions of dollars, and the necessary computing clusters require specialized chips, cooling systems, and energy supplies that few organizations can justify. Huawei is betting that this scale barrier will drive a wave of cloud adoption. The company’s cloud division has been aggressively expanding its AI services, offering everything from pre-trained models to custom chip access.

The strategy is not subtle: position Huawei Cloud as the only viable infrastructure for enterprise AI in China. Competitors like Alibaba Cloud and Tencent Cloud are also racing to capture this market, but Huawei is leaning heavily on its hardware roots. The company designs its own Ascend AI chips and integrates them into its cloud offerings, creating a vertically integrated stack that rivals argue is difficult to replicate.

What a casual observer might miss is how this shift reshapes the competitive dynamics of China’s tech sector. For years, the cloud market was dominated by e-commerce and social media giants. Huawei entered late but has been gaining ground by targeting government and state-owned enterprise clients, which value security and local supply chains. The AI boom amplifies that advantage.

If Zhou’s thesis holds—that AI demands cloud-scale infrastructure—then Huawei’s deep ties to China’s industrial and public sectors could give it a durable edge over more consumer-focused rivals. The implications extend beyond cloud market share. As enterprises migrate their AI workloads to cloud platforms, they also hand over control of their data and algorithms. This concentration of AI capability in a handful of cloud providers raises questions about innovation, competition, and even national security.

In China, where the government already exerts significant oversight over AI development, the cloud layer becomes a chokepoint. Zhou’s remarks also hint at a longer-term vision. He did not frame cloud as a mere utility but as the engine room for the next wave of AI breakthroughs. That suggests Huawei is preparing for a future where even mid-sized companies will rely on cloud-based AI services rather than building their own.

The question is whether those companies will accept the trade-off: access to cutting-edge AI in exchange for dependence on a single infrastructure provider. For now, Huawei is betting they will have no choice.