Alibaba Is Sacrificing Near-Term Profit to Build an AI-and-Cloud Business

Alibaba is making an explicit trade: less profit and shareholder dilution today in exchange for the computing capacity, models and cloud products it expects will drive future growth. The strategy is coherent, but it raises the bar for execution. Investors now need to see AI-related cloud demand translate into sustained revenue, utilization and profitability.

The financing: equity capital for a full-stack AI push

Alibaba completed an HK$80 billion ($10.2 billion) placement of 710 million new shares at HK$112.70 each. The price was an 8.4% discount to the previous close, and the company said all net proceeds would go to its “full-stack” AI capabilities: chips, infrastructure, and AI-model development and deployment. 1721

That structure gives Alibaba capital to accelerate an expensive build-out without taking on additional debt. But it also dilutes existing shareholders. The market’s immediate reaction reflected that cost: Alibaba shares fell about 8% in early Hong Kong trading as investors weighed dilution against uncertain returns from AI investment. 24

The placement was not a sign that demand for the deal was absent. Reuters reported an order book of roughly $28 billion. Still, strong demand for a share sale is not evidence that the underlying AI investments will earn an attractive return. 2

Why profits fell while cloud demand accelerated

The June-quarter results show both sides of Alibaba’s bet. Net income fell 75% year over year to RMB10.4 billion, while revenue rose 9%. Capital expenditure rose 75% to RMB67.7 billion as the company expanded AI infrastructure and compute capacity. 3141

At the same time, AI Cloud and Compute Services revenue grew 45% to RMB48.4 billion. Alibaba reported that the segment’s adjusted EBITA margin reached 12%, an indication that the cloud business can show operating leverage even while group-level profits are pressured by the investment cycle. 3534

This is the central balancing act:

• Upfront cost: chips, data centers, compute capacity and model development require heavy spending before revenue fully catches up.
• Potential payoff: customers that use Alibaba’s models may also consume its cloud compute, storage and related services.
• Financial risk: if demand, pricing or capacity utilization disappoints, depreciation and continuing capital spending can weigh on earnings for longer than investors expect.

Alibaba is therefore not trying to preserve short-term margins. It is trying to build the infrastructure and product stack required to compete for enterprise AI workloads.

Wan3.0 is a product-led route to cloud monetization

Wan3.0, Alibaba’s AI video-generation model, matters because it is tied to a commercial cloud distribution channel rather than being only a technology demonstration. API pricing on Alibaba Cloud Model Studio starts at $0.05 per second for 480p video, rising to $0.10 for 720p and $0.20 for 1080p. 11

Usage-based pricing creates a direct revenue path: customers pay as they generate video. More broadly, a model product can increase demand for the surrounding cloud platform, including compute and storage. That is the strategic logic behind owning the stack from chips and infrastructure to models and developer-facing services.

The caveat is that low entry pricing can help adoption without guaranteeing attractive margins. Wan3.0’s commercial value will depend on customer volume, retention, inference costs and whether Alibaba can convert model users into broader cloud customers.

Why the market remains skeptical

The share-price reaction is understandable. The placement created immediate dilution, was priced at a discount, and followed a quarter in which profit fell sharply. Investors also face a timing problem: capital spending is visible now, while the return on AI infrastructure may take longer to prove. 232

The key issue is not whether AI is strategically important. It is whether Alibaba can earn enough incremental cloud and model revenue to justify the additional capital, ongoing depreciation and lower per-share ownership for existing investors.

That is why the company’s cloud metrics matter more than an AI launch headline. Investors will be watching for sustained AI Cloud and Compute Services growth, improving segment economics, evidence of capacity utilization and a stabilization in group profitability.

Why Goldman Sachs still sees a Buy case

Goldman Sachs’ positive view is a forward-looking earnings thesis. In June, the firm maintained a Buy rating and said it expected Alibaba’s earnings-per-share downgrade cycle to near a bottom, with profitability improving in the second half of the year. It cited Alibaba’s AI and cloud-computing leadership as part of its rationale. 23

That view depends on several things going right:

1. AI cloud growth must persist. The latest 45% growth rate is encouraging, but it must continue as Alibaba adds capacity. 3235
2. Cloud economics must scale. The reported 12% cloud EBITA margin offers some evidence of operating leverage, though it does not remove the pressure on consolidated earnings. 34
3. The new funding must produce productive capacity. Proceeds are earmarked for the full AI stack, so the eventual return depends on customer demand for the infrastructure and services that stack enables. 17
4. The core business must support the investment cycle. A recovery in overall profitability would make the AI spending easier to absorb and reduce concern that investment is permanently impairing earnings.

The bottom line

Alibaba is balancing AI expansion by choosing financing and product commercialization over near-term earnings protection. The HK$80 billion placement supplies capital for chips, infrastructure and models; the 45% expansion in AI Cloud and Compute Services shows real demand; and Wan3.0 offers a metered route to model revenue through Alibaba Cloud. 173511

But the strategy is not yet validated solely by rapid growth. The 75% profit decline, higher capital spending and shareholder dilution mean Alibaba must demonstrate that cloud and model revenue can scale into durable per-share earnings. Goldman Sachs’ Buy case is plausible if that conversion happens; the market’s caution reflects the risk that it may take longer or cost more than expected. 2332