Chinese digital platforms are on the brink of a fundamental shift in the monetization of artificial intelligence. In my analysis of market dynamics, a clear pattern emerges: over the next two to three years, leaders of the Chinese internet will begin accumulating the bulk of profits from AI technologies, but only if export restrictions on chips and computing infrastructure are relaxed.
Currently, margins in this sector are concentrated in the hands of hardware suppliers and related services—the market is squeezed in the grip of a capacity shortage. However, based on my estimates, grounded in deep research of the sector, the situation will change dramatically. Once barriers to the supply of advanced chips are lowered, pricing leverage will shift to platforms with massive user bases, rich data, and developed distribution channels. For China, this means a return to a full-fledged monetization phase.
Investment fever and market anxiety
The forecasts come amid unprecedented growth in capital expenditures by Tencent and Alibaba. The latter's figures are particularly telling: in the second quarter, capital expenditures nearly tripled, reaching 52.8 billion yuan ($7.86 billion). For the first time, the company's free cash flow turned negative—at 13.8 billion yuan ($2.05 billion). For the June quarter, the outflow amounted to 44.7 billion yuan ($6.65 billion), more than double last year's level, while total spending surged to 67.7 billion yuan ($10.07 billion).
Such investments raise justified concerns: the macroeconomic environment is slowing in the second half of the year, putting pressure on short-term profits. Notably, I estimate that Chinese tech companies' annual AI spending is equivalent to 18 months of their cash flow—an aggressive bet that leaves no room for error.
Efficiency as a trump card
Even with weak returns, these investments are critical for survival in the competitive race. The combined AI spending of Chinese players amounts to only one-seventh of similar expenditures by American giants. This gap is explained by limited access to advanced foreign chips and smaller business scale, which for now preserves the advantage for the infrastructure segment.
However, the Chinese AI sector has a hidden advantage—exceptional cost efficiency. The cost of training local models does not exceed 10% of the level of global leaders, and the average API price of China's largest models stays below 20% of international competitors. This factor will prove decisive: in the medium and long term, victory will go not to those who spend the most, but to those who most effectively leverage their audience, data, and distribution channels.
My expert assessment: the current phase is an inevitable "entry fee" into a new technological era. Chinese platforms are deliberately sacrificing short-term liquidity for strategic dominance. The question is not whether these investments will pay off, but who will first reach the break-even point in scaling AI services. Given Alibaba's aggressive expansion, which raised $10.2 billion through the placement of 710 million shares, the race has already entered its decisive stage.