Chinese internet giants are on the verge of transitioning to an active phase of monetizing artificial intelligence. Based on my estimates, derived from an analysis of market trends and data from the investment bank UBS, these companies will begin generating a significant share of their profits from AI within two to three years. A key condition is the easing of restrictions on chip supplies and the development of the necessary infrastructure.
Currently, the main margin in the AI sector is captured by suppliers of "hardware" and related services. The market is hitting a severe shortage of computing power, which temporarily shifts the balance of power in favor of equipment manufacturers. However, as Kenneth Fong, head of UBS's China internet sector research, emphasizes, if restrictions ease, "pricing power" will inevitably shift to platforms with large audiences, unique data, and developed distribution channels.
Giants' Costs: A Bet on the Future or a Risk to Profits?
These forecasts come amid unprecedented growth in AI spending by Tencent and Alibaba. The latter nearly tripled its capital expenditures in the second quarter, reaching 52.8 billion yuan ($7.86 billion). Notably, the company's free cash flow turned negative for the first time, amounting to minus 13.8 billion yuan ($2.05 billion). For the quarter ending in June, the free cash flow outflow reached 44.7 billion yuan ($6.65 billion)—more than double the figure from a year earlier. The company's total quarterly expenses soared to 67.7 billion yuan ($10.07 billion).
UBS rightly notes that the market views such investments with caution, especially amid a slowing macro environment in the second half of the year and pressure on short-term profits. Fong estimates that Chinese tech companies' annual AI spending is equivalent to their cash flow over one and a half years. This is a colossal burden, but it is necessary to maintain competitiveness.
Efficiency as the Main Weapon
Even if the return on investment proves weaker than expected, these expenditures are critically important to avoid falling out of the race. Notably, the total AI spending of Chinese technology companies amounts to only about one-seventh of that of their American competitors. This gap is directly linked to limited access to advanced foreign chips and the smaller scale of their businesses. In the short term, this maintains the advantage for hardware manufacturers.
However, the Chinese AI sector also has an undeniable advantage—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 is less than 20% of international competitors. In the medium and long term, the key factor will not be the volume of capital expenditures, but the dominance of platforms with large audiences, data, and service distribution channels.
My view: Chinese platforms are in a unique position. Their main trump card is not an arms race in computing power, but the ability to extract profits from data and their user base. If infrastructure constraints are lifted, we will witness explosive growth in the monetization of AI services in China, which could reshape the global balance of power in the technology sector.