Alibaba will place 710 million new shares to raise HK$80 billion ($10.2 billion). The company will direct all proceeds to artificial intelligence—from computing infrastructure and in-house chips to models and applications.
The placement price is HK$112.70 per share. Closing is expected on August 26, subject to customary conditions.
According to Reuters, the sale is the largest follow-on by any Hong Kong-listed company and the third-largest globally in 2026, after Alphabet and Intel.
All net proceeds will go to AI
The new shares represent about 3.6% of Alibaba’s enlarged share capital. The company priced them at an 8.4% discount to the Friday, August 21, closing level.
In the company’s official statement, Alibaba said the capital will expand its capabilities across the AI stack.
Funds will go in part to infrastructure for training and running models. Reuters also listed the company’s in-house chips among the areas.
Alibaba did not disclose how it will allocate funds across individual areas.
Demand for the placement significantly exceeded its size. According to the agency’s sources, orders totaled about $28 billion, of which roughly $6 billion came from long-term and sovereign investors.
Participants included Qatar Investment Authority, Norway’s Norges and Hillhouse, the sources said. The organizations and Alibaba did not confirm the information.
Alibaba shares fell more than 10%
Investors reacted negatively to the dilution of existing stakes. On August 24, Alibaba’s Hong Kong shares fell 10.5%, later trimming losses to roughly the size of the placement discount.
Sentiment is also influenced by questions about the payback of rapidly rising AI spending.
For April–June, Alibaba’s capital expenditures rose 75% year over year to 67.68 billion yuan (about $10 billion). The company attributed the increase primarily to AI infrastructure and processor purchases.
Net income fell 75% in the same quarter.
At the same time, revenue at the AI Cloud and Compute Services unit jumped 45% to 48.44 billion yuan ($7.1 billion). Revenue directly from AI products exceeded $1.8 billion and grew at triple-digit rates for the twelfth consecutive quarter, according to the company’s data.
Alibaba has already spent nearly half of its AI budget
The new placement complements Alibaba’s previously announced investment program.
In February 2025, the company pledged to invest at least 380 billion yuan in AI and cloud infrastructure over three years. At the current exchange rate, Reuters estimates the amount at about $56.5 billion.
By August, Alibaba had already deployed nearly half of the program’s capital.
CEO Eddie Wu explained the acceleration in spending by the need to build computing infrastructure ahead of future demand.
“To benefit from future growth, we first need to make these capital investments and build the necessary computing capacity,” he said on the quarterly earnings call.
Alibaba expects the investments to pay back in roughly three years, and with further margin expansion—in two and a half years.
One way to reduce costs, the company believes, is to shift from purchased processors to its in-house T-Head solutions. Wu said increasing their share in data centers should lift gross margin and the profitability of the AI business.
China lags the U.S. in investment scale
Despite the size of the new placement, China’s largest tech companies are still investing in AI far less than their U.S. rivals.
Capital Group estimates cumulative AI capex by Microsoft, Amazon, Alphabet, Meta and Oracle reached $791 billion by the end of July. For ByteDance, Alibaba, Tencent and Baidu, the figure was $118 billion.
One reason for the gap is U.S. restrictions on shipments to China of Nvidia’s most powerful accelerators. They are forcing local developers to build more of their own processors and optimize models for less compute.
Alibaba is betting on several layers: its T-Head chips, cloud infrastructure, the Qwen model family, and consumer AI products.
In August, company representatives said downloads of its open models surpassed 3 billion. The company has released more than 460 neural networks, on which developers have built hundreds of thousands of derivatives.
Later, Alibaba decided to sell its gaming unit Lingxi Games to the fund Trustar Capital. The potential deal value starts at $1.5 billion. The decision was attributed to a business shift toward AI and cloud technologies.
