UBS analysts forecast that Chinese internet platforms — armed with vast proprietary data and large user bases — will begin capturing a disproportionate share of artificial intelligence profits within two to three years, even as near-term macro headwinds and surging AI infrastructure spending weigh on free cash flows across the sector.

The cycle argument

Kenneth Fong, head of China internet research at UBS, outlined the thesis at a UBS event in Shenzhen on Tuesday, September 2, 2026. He argued that profit power currently sits with upstream hardware and infrastructure providers because capacity constraints give them pricing leverage. That balance, he said, is set to flip.

'This is all about the cycle. Now the capacity constraint is on the upstream... so [they] capture a big chunk of the whole profit pool,' Fong said. 'But two to three years down the road after the capacity constraint eases, the pricing power will shift to the downstream, where they have the distribution capability, data and users.'

Capex surge squeezes near-term cash flows

The spending ramp-up is already visible in quarterly filings. Tencent Holdings nearly tripled its second-quarter capital expenditure to 52.8 billion yuan (US$7.85 billion), posting a negative free cash flow of 13.8 billion yuan for the first time. Alibaba Group Holding saw its free cash outflow in the June quarter more than double year on year to 44.7 billion yuan, driven by a 67.7 billion yuan quarterly capital outlay.

According to Fong, investor caution in the second half of the year reflects concern that aggressive hardware and infrastructure spending will drag down short-term profitability — a rational response to a weak macroeconomic backdrop.

Why it matters

The UBS view reframes the current capex cycle not as value destruction but as positioning. Internet platforms that control distribution, user relationships, and proprietary data are, in this reading, building moats that will monetise once compute supply loosens and unit economics improve. The question is whether investors will hold through the trough.

Chinese tech giants are effectively betting that the AI infrastructure buildout — however costly today — will eventually compress model costs and allow downstream application layers to extract margin. That is a thesis familiar from the early cloud era, though the timeline and regulatory environment differ materially.

What's next

Markets will watch whether Tencent, Alibaba, and peers can demonstrate early AI-driven revenue acceleration in upcoming quarters to justify the capital outlays. Any easing of upstream chip or server capacity constraints — whether through domestic semiconductor progress or global supply normalisation — could accelerate the profit-shift timeline Fong describes. Investors most exposed to the current drag are those with short-horizon positions in names carrying elevated capex guidance through 2027.