The smart money is rotating out of chips and into the platforms that actually monetize AI, and Beijing just made that bet a lot less risky.
The Summary
- Franklin Templeton's main emerging equity fund is cutting memory chipmaker positions and rotating into Chinese internet stocks like Alibaba and Tencent after chips rallied hard
- China's internet sector is staging a comeback on stronger earnings outlook, policy support signals, and optimism around domestic AI development
- The trade isn't just value rotation. It's a bet that Beijing's regulatory freeze is thawing and Chinese platforms will capture AI upside without the chip premium
The Signal
Franklin Templeton is making a clear call: the AI trade has moved too far into hardware. Memory chips ran up on inference demand and data center buildouts. Now the fund is taking profits and moving capital into the layer that actually talks to users and captures recurring revenue.
The targets are Alibaba and Tencent, the two giants that own Chinese consumer internet. These aren't speculative AI plays. They're profitable platform businesses with distribution, data moats, and the compute partnerships to deploy models at scale without building the infrastructure themselves.
"The rotation is from picks-and-shovels into the companies that own the towns where the gold rush is happening."
What changed? Three things converged. First, earnings expectations for Chinese internet stocks are improving. The sector lagged the broader Chinese market during the regulatory crackdown. Now analysts see a path to growth again.
Second, policy signals suggest Beijing is warming to tech again. The government needs these platforms healthy to compete globally in AI. Crushing your best distribution networks while trying to win a technology race doesn't work. The thaw isn't loud or formal, but capital markets are reading the room.
Third, optimism around China's AI progress is giving internet platforms an AI premium without requiring them to be pure-play AI companies. They're infrastructure for deploying Chinese models to Chinese users at scale.
Key dynamics at play:
- Chip valuations reflected scarcity and buildout urgency. That trade is mature.
- Platform valuations still reflect regulatory overhang and skepticism. Room to run.
- Beijing's strategic interest in AI superiority means protecting, not punishing, distribution platforms.
This isn't a contrarian bet. It's recognition that the value chain in AI isn't linear. Chips enable models. Models need platforms. Platforms own users. If you believe Chinese AI development continues and regulatory pressure eases, the platforms are underpriced relative to the infrastructure layer.
The Implication
Watch how other institutional capital follows this rotation. If Franklin Templeton is early, others will pile in once earnings confirm the thesis. If Alibaba and Tencent start reporting AI-driven engagement or revenue growth in the next two quarters, the trade gets crowded fast.
For builders, this signals where deployment capital is flowing. Chinese platforms will spend on models and inference, not custom chips. That shapes what gets built and who gets funded in the Chinese AI stack. The winners won't be the companies trying to replicate Nvidia. They'll be the ones plugging models into super apps with 500 million users.