The real reason why the AI bubble hasn't burst is Anthropic

Source: **WhiteLine; Translation: Wu Shuo Blockchain

This episode focuses on the most critical theme in the current AI market: Anthropic's rapid revenue growth is reshaping the market's pricing logic for AI infrastructure investment (AI CapEx).

Host Minta believes that the reason AI infrastructure hasn't collapsed amid skepticism about a “bubble” isn't just chips, semiconductors, or model iterations, but rather that Anthropic has proven that large model commercialization is quickly materializing. The article combines data on revenue, valuation, and enterprise client payments to show that Wall Street is shifting from doubting whether AI can make money to a new framework asking whether computing power limits revenue ceilings. It further breaks down three layers of trading pathways around Anthropic, including primary market pre-IPO shares, secondary market shadow stocks, and the entire AI CapEx supply chain.

Minta believes SK Telecom is a relatively purer Anthropic shadow asset but also highlights risks such as stake percentage, mapping discounts, and fluctuations in its main business. The overall conclusion is: Anthropic is responsible for validating demand, and AI CapEx is responsible for expanding production. Together, they form the core logic of this AI market cycle.

Below is a summary of this episode's video transcript:

I. The Core Logic of AI CapEx

AI CapEx refers to the infrastructure costs large companies invest upfront to develop AI, including GPUs, data centers, electricity, liquid cooling, and networking. These investments are difficult to recoup in the short term, essentially preparing for future capacity.

The reason the AI bubble hasn't burst isn't just because semiconductors are strong and AI iteration is fast, but more critically because demand is starting to be validated. White Line believes the key company making this logic work is Anthropic. It is precisely Anthropic's explosive revenue growth that has shifted AI CapEx from a pure money-burning narrative to a production capacity narrative that can be realized.

II. Why Anthropic is the AI Main Theme

Anthropic is the main theme of this AI cycle because it has been the first to prove that AI can be truly commercialized, not just a money-burning narrative.

Three key points:

1. Revenue explosion. Annualized revenue grew from approximately $1 billion in early 2025 to over $47 billion by May 2026, expanding about 47 times in a year, indicating it has deeply penetrated enterprise workflows.

2. Valuation revaluation. The valuation rose from $183 billion to $965 billion, reflecting the market's systematic repricing.

3. Enterprise payments substantiated. Claude Code's revenue is growing rapidly, and the number of enterprise clients paying over $1 million annually has increased significantly, showing that growth is mainly driven by deep enterprise spending rather than consumer-side experimentation.

Thus, Anthropic's significance lies not in “fast growth” but in validating AI's commercial closed loop: enterprises are willing to pay continuously, and AI CapEx thus has a recovery logic.

III. Gavin Baker's Judgment Framework

Whether Anthropic is expensive depends not on static valuation but on where the revenue bottleneck lies: is it insufficient demand or insufficient computing power?

If the bottleneck is demand, growth space is near its limit; if the bottleneck is computing power, current revenue is not the ceiling. According to Gavin Baker, once computing power is unleashed, Anthropic's revenue could rise from $50B to $1 trillion or even $2 trillion.

Seen this way, a nearly $1 trillion valuation corresponds to about 20x PS for $500 billion revenue; but if revenue rises to $1-2 trillion, the valuation multiple would drop to 5-10x.

So the core issue is not “is it expensive now?” but rather: how much additional revenue can Anthropic generate for each additional unit of computing power?

IV. How to Trade Around the Main Theme When Anthropic is Not Public

Trading around Anthropic can be divided into three layers: buying pre-IPO shares, buying shadow stocks, and buying the AI CapEx supply chain.

1.Pre-IPO

The most direct way is to buy Anthropic's pre-IPO shares, offering the greatest leverage and closest proximity to the company itself.

But the core risk is whether the equity transfer is officially recognized. If the platform and shares are not compliant, they may not truly capture subsequent valuation upgrades.

2. Shadow Stocks

The criterion for a shadow stock is not high correlation, but actual ownership of Anthropic shares, with the value sufficient to affect its own stock price.

Amazon, Google, Fidelity, Blackstone, Sequoia, as well as Micron, Samsung, and SK Hynix, all have ties to Anthropic, but their stock prices are more driven by their own core businesses, making the trading logic less pure.

In comparison, SK Telecom (SKM) is closer to a pure shadow stock: first, it is publicly listed and directly tradable; second, it holds Anthropic shares; third, its own market cap is relatively small and its core business is stable, making Anthropic's valuation changes more easily transmitted to its stock price.

Based on public estimates, SKM holds approximately 0.54% to 0.58% of Anthropic. At Anthropic's $965 billion valuation, the corresponding equity value is about $52 billion, while SKM's market cap is around $14 billion, accounting for over 30%. Therefore, its logic can be summarized as “telecom cash flow + Anthropic valuation option.”

The risks are equally clear: the stake percentage has not been officially confirmed, subsequent financing and options may dilute it, SKM is essentially a telecom stock, and shadow stocks cannot map one-to-one to Anthropic's valuation changes.

3. AI CapEx Supply Chain

Expand outward from Anthropic to the AI CapEx supply chain. As long as Anthropic continues to validate demand, the market will keep betting on AI infrastructure expansion, with trading spilling over into cloud computing power, chips, advanced manufacturing, as well as liquid cooling, optoelectronics, electricity, networking, and other links.

In other words, Anthropic validates demand, while AI CapEx trades on production expansion.