China's AI spending spree just hit the quarterly earnings wall, and investors want their money back.
The Summary
- Alibaba and Tencent both missed revenue estimates despite ramping up AI infrastructure spending over the past year
- Chinese investors are demanding evidence that billions poured into AI development are generating actual returns, not just press releases
- The disconnect: massive capex on models and compute, minimal revenue attribution from AI products
- This mirrors the pressure facing US tech giants, but China's growth slowdown makes the scrutiny sharper
The Signal
Both Alibaba and Tencent entered earnings season carrying investor expectations that their AI investments would show up as revenue growth. They didn't. The companies have spent billions building out large language models, AI-powered commerce tools, and cloud infrastructure. None of it moved the revenue needle enough to hit Street estimates.
This isn't a story about bad technology. It's a story about monetization lag. China's tech giants built the models. They launched the products. What they haven't figured out is how to charge for AI in ways that customers will actually pay. Free tier wars and promotional pricing dominated 2025. Now the bills are due.
"Chinese investors are heading into earnings season demanding proof that billions in AI spending are paying off."
The timing makes this worse. China's domestic economy is grinding through a slower growth phase. Consumer spending is cautious. Enterprise budgets are tight. This is exactly the wrong environment to try selling expensive new AI capabilities to companies that are watching every yuan. The US market has enterprise customers with budgets to experiment. China's market is asking for ROI spreadsheets before the pilot program.
Here's what the earnings miss signals:
- AI infrastructure spending is real and recurring, but revenue is lumpy and promotional
- Cloud margins are getting squeezed as AI compute costs rise faster than pricing power
- Consumer AI features aren't yet differentiated enough to drive platform stickiness or premium pricing
The contrast with their US counterparts is stark. Microsoft and Google can point to GitHub Copilot subscriptions and Workspace AI add-ons. They have B2B pricing models that enterprises are actually buying. Alibaba and Tencent are still figuring out whether to charge for AI features at all, or bundle them into existing products and hope for halo effects.
The Implication
Expect both companies to shift from "look what our models can do" messaging to "here's how customers are paying for it" storytelling. That means clearer product packaging, tiered pricing, and probably some writedowns on infrastructure that got built ahead of demand.
For the broader Web4 build-out, this is a reminder that foundation models alone don't generate revenue. The value is in the application layer, the workflow integration, the thing people actually pay to solve. Alibaba and Tencent have the models. Now they need to find the business model that makes investors stop asking when the payoff arrives.