Where is the US$10.2 billion going?

On 23 August, Alibaba priced 710 million newly issued shares at HK$112.70 each, raising HK$80 billion. The placement price represented an 8.4% discount to the previous Hong Kong close. Reuters described it as the largest-ever primary follow-on offering by a Hong Kong-listed company.

The destination of the money is unusually straightforward. Alibaba says 100% of the net proceeds will be invested in its ‘full-stack AI capabilities’, including AI infrastructure. In plain English: Alibaba wants more computing power, and it is willing to dilute shareholders to get it.

The new shares represent roughly 3.7% of Alibaba’s pre-deal issued share capital. The placement is expected to close on 26 August 2026, subject to customary closing conditions.

What sparked the market reaction?

The capital raise did not happen in isolation. Just days earlier, Alibaba reported numbers that showed both sides of its AI transformation.

For the June quarter, group revenue increased 9% YoY to RMB268.95 billion, or US$39.6 billion. Its AI Cloud and Compute Services business was the standout. Revenue increased 45% YoY to US$7.1 billion, the fastest growth rate Alibaba has reported for the business in 22 quarters.

AI-related product revenue also recorded its 12th consecutive quarter of triple-digit YoY growth. Demand is showing up. The question is how much Alibaba needs to spend to meet it.

Can Alibaba fund its AI spending?

Alibaba spent nearly US$10 billion on capex in a single quarter, up 75% YoY. Meanwhile, GAAP net income dropped 75% to RMB10.44 billion, or roughly US$1.54 billion. Non-GAAP net income fell 38%.

Free cash flow (FCF) was a RMB44.7 billion outflow, compared with an RMB18.8 billion outflow in the same quarter a year earlier. That makes the market’s reaction easier to understand. The question is becoming less about whether AI demand is showing up and more about how much the build-out will cost before it pays off.

Alibaba has previously committed at least RMB380 billion, around US$56 billion, over three years to cloud and AI infrastructure. The new equity raise shows that commitment is not getting smaller.

How is Alibaba restructuring for AI?

The organisational chart is changing too. Beginning this quarter, Alibaba reports four distinct operating segments:

The new reporting structure separates Alibaba’s AI infrastructure and applications businesses, giving investors a clearer view of the revenue and adjusted EBITA generated by each.

Cloud is beginning to show operating leverage. AI Cloud and Compute adjusted EBITA jumped 133% to US$830 million during the quarter. But AI Labs and Applications recorded an adjusted EBITA loss of about US$2.04 billion.

That contrast may be the most interesting number in the entire result. Alibaba’s AI infrastructure business is generating more adjusted earnings. Its model and application layer is still recording substantially larger losses.

Is Qwen driving real growth?

This is where Alibaba’s AI story gets more interesting. Alibaba says its Qwen family of open-weight models has passed 3 billion global downloads over six months. The company has released more than 460 open-weight models, while developers have created more than 300,000 derivative models across the ecosystem.

Downloads do not equal profit. But they do matter. For cloud companies, widespread developer adoption can become distribution. Developers build on the models, businesses need inference and compute, and some of that workload may ultimately land back on Alibaba Cloud.

Alibaba is also trying to connect Qwen directly to its enormous commerce ecosystem. Its Qwen Shopping Assistant has been integrated into Taobao, while Qwen App introduced paid productivity and professional subscriptions in August.

The strategy appears increasingly clear: distribute the models widely. Then seek to monetise the compute, enterprise workloads, subscriptions, and commerce around them.

Why are executives buying shares?

Alibaba shares were falling. Its top two executives were buying.

On 24 August, chairman Joe Tsai purchased 720,000 shares, worth roughly HK$81 million. CEO Eddie Wu bought another 350,000 shares, worth about HK$39 million. Combined, that is approximately HK$120 million, or US$15.3 million.

Joe Tsai Chairman

720,000 shares~HK$81 million

Eddie Wu Chief Executive Officer

350,000 shares~HK$39 million

Insider transactions reflect public disclosure filings and do not constitute a trade recommendation or an indicator of future share price performance.

It is an eye-catching detail, but it needs context. Insider buying is public information, not a trading signal. Executives may purchase shares for many reasons, and their purchases tell investors nothing certain about where a share price goes next. Still, amid a US$10.2 billion equity raise and a sharp sell-off, the timing is worth knowing.

Which Alibaba story wins?

There are really two Alibaba stories happening at once.

Story one: Alibaba’s AI strategy appears to be gaining commercial traction. Cloud growth is accelerating, Qwen downloads are high, and the cloud segment is becoming substantially more profitable on an adjusted EBITA basis.

Story two: Getting there is capital-intensive. Alibaba is spending billions on infrastructure, its AI applications division remains deeply loss-making, FCF is under pressure, and existing shareholders are being diluted to fund the next phase.

Will the future earnings from Alibaba’s AI ecosystem justify the capital being poured into it today? Nobody has that answer yet. That uncertainty is one reason the stock has been volatile.

What metrics to watch next?

The next phase is less about AI headlines and more about execution. Key areas include whether cloud growth can remain around current levels, whether AI Cloud margins continue to expand, whether the AI Labs loss begins to narrow, and whether Alibaba’s huge capex program starts producing stronger FCF.

Alibaba has put a very large number on its AI ambitions. Now investors get to watch the payback clock.

The story is loud. The numbers matter more.