One day in November 2020, somewhere in San Francisco, a woman named Daniela Amodei called a friend for advice.

At the time, Daniela was responsible for security and policy at OpenAI. OpenAI is one of the world's most high-profile artificial intelligence (AI) companies.

But Daniela and her brother Dario Amodei had concerns about OpenAI’s safety standards. Dario also worked at the company, and the two began to think it might be time to leave.

So she did what many people do when faced with a big decision: called a friend.

My friend’s answer was succinct: “You don’t need to talk to me on the phone. You already know the answer.”

The Amodei siblings and several colleagues later resigned to co-found a rival company, Anthropic.

Today, six years later, according to analyst estimates, Anthropic, which owns the leading AI model Claude, is expected to have revenue of US$52 billion in 2026 and US$135 billion in 2027.

Today, the company is selling shares to the public for the first time, an initial public offering (IPO), with a target valuation of about US$1.4 trillion (10.22 trillion yuan; NT$45.64 trillion). If it successfully meets the criteria, Anthropic will directly rank among the top ten companies with the highest market value in the United States.

End of Hot Reading

End of content

The listing comes at a critical time. As external concerns about the prospects of artificial intelligence rise, many people have begun to wonder whether the AI ​​industry bubble has shown signs of bursting.

In the past, investors were almost entirely betting on a technological revolution that was seen as rivaling the revolutions in railroads, cars, electricity and the Internet.

Now, however, some investors are beginning to take a more conservative stance. Some people believe that as AI technology becomes more popular, prices will gradually decline; others worry that the current large amount of AI investment is actually just a cycle formed by cross-investment between technology giants.

The stock prices of companies that provide AI computing infrastructure, such as CoreWeave and Nebius, have halved from their highs. Even Nvidia (Nvidia/Huida), the former global market capitalization king, has lost its throne as the world's most valuable company. Its stock price fell by 30% and its market value evaporated by as much as US$2 trillion.

So, will Anthropic’s IPO re-energize the AI ​​bubble? Or will it be viewed in retrospect as the last gasp of prosperity before its collapse?

Valuation Benchmarks for the AI ​​Industry

"My fundamental rule is never to let a company be worth a dollar more than it's really worth," Anthropic CEO Dario Amodei told JPMorgan Chase CEO Jamie Dimon at an event last week.

Anthropic will sell approximately $81 billion in shares to general investors this time, and another $26 billion in shares will be sold to other large technology companies.

This is the largest fundraising event in U.S. corporate history.

There is an old saying that in the gold rush, the most valuable investment is not the gold diggers, but the people selling shovels and pickaxes.

In the AI ​​industry, those "shovel sellers" are chip manufacturers, data center builders, and energy suppliers.

But Anthropic itself can be said to be the "gold".

Fiona Cincotta, an analyst at StoneX Trading Company, pointed out that this IPO will allow the market to truly test for the first time how much value people are willing to give to AI itself.

She said: "The reason why this IPO is so important is that it gives ordinary investors the opportunity to directly invest in pure AI concept stocks for the first time."

She added that previous listing targets were mainly other links in the AI ​​ecosystem, such as semiconductors, chips, and data centers. This is the first time investors have been able to invest directly in the AI ​​model providers themselves.

image copyrightBloomberg via Getty Images

Business value depends almost entirely on expectations of future profitability.

It is this valuation logic that many people regard as the sharpest needle that may puncture the AI ​​bubble.

The question is: Are users really willing to pay enough to support the trillions of dollars currently being spent on chips, data centers, energy and talent?

After all, the annual salary of some AI talents even exceeds 100 million US dollars.

Gary Marcus, professor emeritus at New York University and author of Taming Silicon Valley, said earlier this year: "There is no doubt that AI is in a financial bubble. While large language models (LLM) do have practical uses and will certainly continue to exist, from an economic perspective, this account does not make sense."

He said: "As more and more companies build AI models, the technology will gradually be commoditized, competitive advantages will be eliminated, and prices will continue to be driven down."

So far, Anthropic's revenue growth rate is indeed amazing.

However, these explosive growth also hide potential problems.

About a quarter of its revenue comes from a handful of companies, including Google and Amazon.

In addition, its revenue is highly concentrated in the U.S. market, which accounts for approximately two-thirds of total revenue.

What is even more noteworthy is that there is a huge circular financing network within the AI ​​industry.

For example, Anthropic has signed agreements with a handful of companies, including Broadcom, Google, and Amazon, committing to purchase or lease hundreds of billions of dollars worth of chips and computing power.

Many of these companies are also investors and major customers of Anthropic.

Rising public concerns

Wall Street investors aren't the only ones doubting whether the AI ​​craze is real.

There is also growing concern among ordinary Americans.

Since the United States is the main engine of global AI growth, opposition to AI has been rising recently. Many people worry that AI will replace millions of jobs and consume large amounts of energy and water. These issues even affected the recent U.S. midterm elections.

Many regions in the United States have adopted data center development bans or moratoriums.

Many candidates who are considered too close to big technology companies are finding that these connections are no longer political assets but instead become political liabilities. Several Democratic candidates who openly opposed data center development successfully won the election.

In addition, Huida CEO Jensen Huang, who is regarded as the spiritual leader of the AI ​​industry, resigned last week.

Earlier reports accused him of having an extramarital affair with Alice Huang, a woman suspected of spying for China, and an extramarital relationship with a company executive who reported directly to him. Huang later admitted that he had been "unfaithful" to his wife.

image copyrightEPA

Jen-Hsun Huang's dramatic reputational collapse and Huida's stock price falling 30% from its peak have left AI investors looking for a new AI industry leader.

After all, it is the AI ​​craze that has supported the U.S. stock market and even, to some extent, the overall U.S. economy over the past two years.

If Anthropic and its CEO become the new AI leaders, the tune they play will be completely different from the leather-clad Jen-Hsun Huang who is always optimistic about the future of AI.

Huang Renxun advocates the full speed development of AI.

Anthropic was born precisely because of the concerns of retired OpenAI employees about AI safety and liability issues.

Amodei has warned that AI capabilities may bring about hidden worries such as the development of biological weapons, mass surveillance, authoritarian rule, and the disappearance of millions of white-collar jobs. He called for slowing down the development of AI because its capabilities are increasing at an exponential rate and may soon exceed human supervision capabilities.

However, on the one hand, he warned about the risks of AI and called on the industry to slow down its development. On the other hand, Anthropic's pace of launching more powerful new models has not slowed down, so its claims have also been questioned. Critics say he asked other AI companies to slow down, but Anthropic itself hasn't shown the same restraint.

Fanaticism and FOMO

Despite many concerns from the outside world, just like SpaceX back then, the market is still enthusiastic about this revolution that may profoundly change technology and even human society.

Investors fear of missing out (FOMO, fear of missing out).

According to reports, the number of applications to subscribe to Anthropic shares currently exceeds the supply by more than six times.

When demand far exceeds supply, it is almost expected that the stock price will rise sharply on the first day of listing.

But that doesn't speak to its long-term return on investment.

Taking SpaceX as an example, its IPO price was US$135, and its stock price rose nearly 70% to US$225 after listing.

Currently, SpaceX stock price has fallen back to around $145.

Anthropic's IPO may be a success and breathe a breath of fresh air into what many see as an AI bubble.

Eileen Burbidge, founder of venture capital firm Passion Capital, does not believe the AI ​​bubble is about to burst.

"Maybe the bubble has slowed down a bit, but I'm not willing to think it's about to burst yet," she said.

She added: "This IPO may even make the entire market take off again."

StoneX analyst Fiona Chincotta believes that the importance of Anthropic's listing goes far beyond Anthropic itself. "I think the most interesting thing is that it may become the valuation benchmark for the entire AI ecosystem."

She also said: "So this will not only affect Anthropic stock price. I think it will have a broader impact on the entire AI investment market."

Finally, both things may actually be true at the same time.

Anthropic's IPO could be extremely successful. After all, the fear of missing out is often powerful in financial markets.

However, even if the listing performance is outstanding, it may not eliminate the doubts in the minds of some investors and the public: not every company, every investor or every citizen believes that AI is the answer to solve human problems.

This article was originally written in English. We used artificial intelligence to assist with translation and was reviewed by BBC journalists before publication. Learn more about how we use artificial intelligence.

Explore more

More short films

More short films end

On the occasion of Trump's visit to China, the United States blocked the "back door" for Chinese companies to rent AI computing power in Southeast Asia.

China hosts world leaders in an AI wonderland, but what is the other side that they can’t see?

In the context of the militarization of AI between China and the United States, the BBC learned that the Pentagon has officially stopped using the Anthropic tool

"Bring Back Jen-Hsun Huang" Movement: Huida refutes, Trump supports reinstating his CEO position

Why do AI companies want to scare you?

China and the United States are leading on different AI tracks, but either one may eventually get ahead.

headlines

The trilateral summit between China, Russia and the United States is more spectacular than substantive results - Analysis by BBC North America correspondent

Changing from Taiwan’s Cangjie to Mainland China’s Hanyu Pinyin: Will the computer input method affect the next generation’s learning of Chinese?

The "Big Three" summit between China, the United States and Russia ends: involving Ukraine and Iran issues, issuing a peace declaration and committing to nuclear dialogue

Special recommendation

Popular content