Alibaba’s $10.2 Billion AI Bet: How Wan3.0 Could Turn Cloud Spending Into Growth

Alibaba is making an explicit trade: accept near-term pressure on profit and per-share value to build AI infrastructure and cloud demand quickly. The August 2026 launch of Wan3.0—a commercial video-generation model—shows the revenue side of that strategy. The HK$80 billion ($10.2 billion) equity placement shows how Alibaba is financing it. 12

The strategy: pair AI products with cloud infrastructure

Wan3.0 can create video clips of up to 30 seconds from inputs including documents, spreadsheets, presentation slides and web pages, according to Alibaba Cloud. Alibaba said the model had already been used during its public beta for short dramas and film production, advertising and marketing, tourism promotion, and music videos. 1

The product creates an immediate, metered revenue opportunity. Reported API prices range from $0.05 per second at 480p to $0.20 per second at 1080p. 3 But the strategic objective is likely broader than video-generation fees alone: a business using a model through Alibaba Cloud also consumes inference capacity and may require storage, data tools and other cloud services. That makes Wan3.0 a potential entry point for higher-value enterprise cloud workloads.

The important caveat is economics. Low usage prices can help adoption, but revenue must scale enough to cover the cost of chips, data centers and model development. The launch demonstrates a path to monetization; it does not yet prove that the return on the infrastructure investment will be attractive.

Why Alibaba raised equity instead of prioritizing buybacks

Alibaba completed a placement of 710 million newly issued ordinary shares at HK$112.70 each, raising HK$80 billion. The company said net proceeds would be used to expand its full-stack AI capabilities, including infrastructure, chips and models. 234

Equity financing supplies committed capital without adding debt and interest expense. In exchange, existing shareholders own a smaller percentage of the company. The placement price was an 8.4% discount to Alibaba’s prior Hong Kong closing price, intensifying the immediate dilution concern. 2

This changes the emphasis of capital allocation. Rather than using available resources chiefly to reduce the share count through repurchases, Alibaba is directing capital toward a multi-year AI buildout. The company is effectively betting that returns from AI and cloud investment will exceed the benefit of retiring shares now.

That is a reasonable strategic choice only if the investment earns its cost of capital. If cloud demand and margins do not rise sufficiently, shareholders will have borne dilution without receiving a matching increase in earnings power.

The near-term cost: weaker earnings, heavier capex and a stock sell-off

The financing came after Alibaba reported a 75% year-on-year fall in quarterly earnings, which Reuters linked to higher AI-related capital expenditure. 17 Large AI deployments typically impose costs before the associated revenue matures: hardware and data-center investment comes first, followed by depreciation, operating expense and the work of acquiring customers.

Investors responded to both that earnings pressure and the equity raise. Alibaba’s Hong Kong-listed shares fell as much as 10% after the placement was announced, with reporting pointing to concerns about dilution and execution risk. 192

The share-price reaction does not establish that the strategy will fail. It does show that the market wants evidence that AI spending can translate into durable cloud revenue and improved profitability, rather than simply larger capital expenditure.

Why Wan3.0 matters to Alibaba Cloud

For Alibaba, Wan3.0 is not just a consumer-facing generative-AI feature. It is a cloud service that can be sold on a consumption basis. Businesses creating marketing videos, product demonstrations, training material or media content can pay for output generation while potentially becoming users of Alibaba Cloud’s wider AI stack.

That creates a possible flywheel:

1. A useful model attracts customers and generates API usage.
2. Usage increases demand for cloud compute and related services.
3. Greater utilization can support better infrastructure economics.
4. Higher cloud revenue and improved margins can fund additional AI investment.

This is the bull case, not a reported outcome. The critical measures will be AI-related cloud revenue, utilization, cloud margins and the relationship between incremental earnings and the dilution created by the placement.

Why Goldman Sachs remains bullish

The constructive view is that Alibaba’s spending has a defined payoff horizon. Reuters reported that Alibaba expects a three-year payback period on its AI capital expenditure. 17 Separately, reporting on Goldman Sachs research said the bank sees potential for a virtuous cycle in which AI demand accelerates cloud growth and improving margins help recover infrastructure costs. 20

Goldman’s reported fiscal 2027 and fiscal 2028 outlook calls for EPS growth of 64% and 33%, respectively. 21 That forecast depends on cloud and AI monetization accelerating while Alibaba’s core commerce operations contribute to profit recovery. It is therefore a forward-looking thesis—not confirmation that the current spending wave has already generated adequate returns.

The bull and bear cases can both be true at this stage. Bulls see a scaled cloud platform building a new earnings engine; skeptics see upfront capex, discounted dilution and an uncertain timeline for returns.

What will decide whether the bet worked

Alibaba has chosen to fund AI aggressively and make products such as Wan3.0 commercially available now. The next phase is execution. Investors should watch whether Alibaba Cloud’s AI-driven revenue grows quickly enough to improve infrastructure utilization and margins, whether capex intensity moderates over time, and whether EPS growth ultimately outpaces the effect of issuing new shares.

If those measures improve, Wan3.0 could become more than a video tool: it could help Alibaba Cloud acquire and monetize enterprise AI workloads. If they do not, the discounted share sale will be remembered as an expensive way to fund an AI race with insufficient returns.