TSMC worries about a chip surplus? Musk: They’re right—power supply and liquid cooling can’t keep up.

```

Recently, Musk stated that TSMC's concern about excess chip production capacity is "correct." He predicts that the limiting factor for the AI industry will shift from chip manufacturing to "getting the chips up and running," and the core bottleneck of this transition lies in power supply, transformer configuration, and cooling system deployment.

On January 6, Musk held an in-depth discussion on the future of AI with Peter Diamandis, Executive Chairman of Singularity University, and David Blundin, founder of Link Ventures. He pointed out that current chip production capacity is growing exponentially, while the energy infrastructure that supports their operation can only expand linearly. When the two curves intersect, a large number of high-performance AI chips will be unable to be put to use due to lack of supporting power conversion equipment and cooling systems.

This judgment directly targets the serious underestimation of current AI infrastructure needs. For investors, this means that the focus of the AI computing power race is shifting from chip procurement to the ability to build energy infrastructure.

Electricity Becomes AI Deployment’s “Rate Limiting” Factor

In the conversation, Musk further revealed the specific form of the power bottleneck in AI infrastructure. He emphasized that deploying AI chips is by no means "just shipping GPUs to a power plant"; rather, it requires simultaneously addressing three core issues: gigawatt-level power supply, high-voltage power conversion, and efficient heat dissipation systems.

Musk specifically pointed out that the entire data center industry is undergoing a critical transformation from air cooling to liquid cooling, and he warned that this process contains huge risks. He said:

“This is a fundamental change for data centers, which have relied on air cooling for a long time. If a liquid cooling pipe bursts—a water pipe exploding in a data center, for example—it could result in losses of up to $1 billion.”

He used xAI’s "Colossus 2" project in Memphis as an example to illustrate the practical challenges: although the project location is adjacent to several 300kV high-voltage lines, it still takes about a year to complete integration. To launch a 1-gigawatt training cluster by mid-January 2025, the team had to temporarily combine multiple gas turbines ranging from 10 to 50 megawatts as interim power sources and use a large number of Megapack battery packs for power balancing.

Curve Intersection of Chip Production Capacity and Power Supply

When asked whether he agreed with TSMC's worries about excess production capacity, Musk gave an affirmative answer: "I’m not sure about their reasons, but the conclusion is correct."

He pointed out that the key is to identify the “limiting factor at each stage,” and he predicts that by Q3 2026 (about 9-12 months later), the core bottleneck will shift from chip manufacturing to the ability to “get chips up and running.”

This judgment stems from a mismatch in two development trajectories: AI chip production capacity is expanding exponentially, while the power infrastructure that supports their operation can only grow linearly. Musk emphasized: “If chip output grows exponentially but power supply can only increase slowly and linearly, the two curves will eventually cross.” This means chip manufacturing speed may far outpace the speed at which they can actually be deployed and energized.

In response, David Blundin offered a different viewpoint, believing that even if TSMC increases GPU output from 20 million to 40 million units, the market will find a way to solve the power supply problem. However, Musk insisted that any missing link in power conversion or cooling systems will prevent the chips from truly being activated, fundamentally suppressing actual demand and procurement behavior.

Risk Warning and DisclaimerThe market has risks, investment must be cautious. This article does not constitute personal investment advice and does not consider individual users’ special investment objectives, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article suit their own specific circumstances. Investments made based on these opinions are at your own risk. ```