China's technology sector stands on the brink of a fundamental shift. Contrary to the widespread belief that profits from artificial intelligence settle exclusively in the pockets of semiconductor manufacturers, my calculations and industry analysis indicate that over the next two to three years, the situation will change dramatically. The internet platforms of the Middle Kingdom, possessing colossal user bases and unique data, will begin to accumulate the lion's share of AI revenue, but only under one condition—the easing of restrictions on chip and infrastructure component supplies.
Currently, the market is in a phase where margins are flowing to hardware suppliers due to an acute shortage of computing power. However, this is a temporary phenomenon. Once supply chains recover, the balance of power will shift toward platforms with multi-million audiences and developed distribution channels. It is they who will be able to effectively monetize AI services, returning to the sustainable revenue growth model that characterized the Chinese internet over the previous decade.
Investment frenzy: stakes are higher than ever
The market is warily watching the unprecedented growth in capital expenditures by sector leaders. Alibaba nearly tripled its capital expenditures in the second quarter, bringing them to 52.8 billion yuan ($7.86 billion). Significantly, the company's free cash flow turned negative for the first time, amounting to -13.8 billion yuan ($2.05 billion). The total outflow for the quarter reached 44.7 billion yuan ($6.65 billion), more than double the figures from the previous year. The e-commerce giant's total quarterly expenditures amounted to 67.7 billion yuan ($10.07 billion).
Such figures inevitably cause nervousness among investors, especially against the backdrop of slowing macroeconomic dynamics in the second half of the year. By my estimates, the annual volume of AI investments by Chinese tech companies is equivalent to their combined cash flow over 18 months. This is an aggressive but necessary measure: stopping now would mean guaranteed exclusion from the race in the future.
Efficiency as the main trump card
The key advantage of the Chinese approach is anomalously high cost efficiency. The cost of training local models is less than 10% of the level of global leaders, and the average API price of major Chinese models does not exceed 20% of international counterparts. At the same time, the total AI expenditures of Chinese companies are only one-seventh of the budget of their American competitors.
This gap is explained both by limited access to advanced foreign chips and by the smaller scale of business. However, in the medium term, the decisive factor will not be the volume of investment, but the ability of platforms to convert their data and audience into dominant AI products. It is here that China has every chance not just to catch up, but to redefine the rules of the game.
My analysis: The market underestimates the potential of Chinese platforms. While attention is focused on chip bans, it is overlooked that the PRC is building an alternative, more cost-effective AI ecosystem. If restrictions are eased, we will witness explosive monetization growth that could overturn the global balance of power in the technology sector.