The Anthropic website and mobile app are seen in this photo in New York, Friday, July 5, 2024.
Richard Drew/AP
hide title
Richard Drew/AP
Stay up to date with our up first newsletter, sent every weekday morning.
Anthropic stock is about to make its market debut after the company raised more than $100 billion in a record initial public offering that valued it at around $1.5 trillion.
It’s a remarkably rapid rise for the five-year-old AI powerhouse and a sign of strong investor appetite despite growing concerns about a potential AI bubble.
Anthropic sold 956 million shares, much more than originally planned, at an initial public offering price of $83 per share, it announced Wednesday. That price exceeded its target range of $70 to $78 per share.
The company’s seven founders also sold more than 23 million shares as part of the IPO. On top of that, Anthropic’s prospectus included plans to put an additional $26 billion worth of shares on the market through private placements with companies like Google, Microsoft and Nvidia.
The total loot is more than 107 billion dollars.
Anthropo’s valuation could rise or fall later on Thursday, when its shares begin trading publicly on the Nasdaq exchange under the symbol ANT.
The amount of money raised makes Anthropic’s IPO the largest in history, after Elon Musk’s megaconglomerate, SpaceX. landed $75 billion when it went public in June.
A separatist beginning
Anthropic’s rise has been meteoric.
“We have never seen a company that is capable of growing so quickly [and] to be that dominant,” said Andrew Rocco, a stock analyst at financial research firm Zacks.
Anthropic began in 2021 as a splinter group of researchers who once worked for OpenAI, creator of ChatGPT. Led by now-CEO Dario Amodei, Anthropic’s founders were concerned that OpenAI, led by Sam Altman, was not on the path to safely developing AI.
Amodei and the others quickly raised hundreds of millions of dollars and set out to develop AI models in a way that the company said focused on safety and “alignment,” or the idea that AI should be harmless and act in ways that are consistent with human values and intent.
Anthropic launched Claude in early 2023, just months after OpenAI launched ChatGPT. Since then, the two companies have been neck and neck competitors in cutting-edge artificial intelligence, known as “frontier AI.” The Claude family of AI models, including the Claude chatbot and Claude Code, are among the most popular and powerful in the world and have helped transform everything from online search to software engineering.
They are also the San Francisco-based company’s main source of revenue, according to its initial public offering prospectus, which said revenue in the first half of the year soared 1.3% to $16.3 billion, while its adjusted operating income – a measure of potential profitability – hit $1.6 billion during the period, up from a loss of $1.8 billion a year earlier.
Opportunities and risks
The IPO comes at a time when shares of technology companies linked to the rise of AI have been volatile. Investors have vacillated between enthusiasm for what many believe will be a revolutionary technology and concern that the huge sums of money being poured into it may be difficult to justify.
“It’s going to be a test of investor sentiment toward AI trading,” said Janice Quek, an equity analyst at investment research firm CFRA. “We’ve seen in recent weeks that there’s been a lot of market pressure and sell-off on the AI ecosystem simply because a lot of people are worried about the AI bubble, for example, or concerns about AI funding and whether there’s a return on investment in AI.”
The tech-heavy Nasdaq has lost more than 11% since the beginning of October.
Altimeter Capital CEO Brad Gerstner, an early investor in Anthropic, said at a Semafor conference on Monday that the IPO is one of the best ways to invest in AI applications. He said his company had cut its exposure to AI infrastructure (companies involved in things like semiconductor manufacturing and data center construction), but that Altimeter had requested an allocation of Anthropic shares worth $5 billion.
“I was waiting for it,” he said of the IPO. “We submitted the largest application in our company’s 20-year history.”
There is indirect evidence in Anthropic’s prospectus that companies, rather than individual subscribers, are its largest source of income. Malik Ahmed Khan, an equity analyst at Morningstar, says the company’s long-term success may depend on the pace of adoption of AI technologies by those companies, but it’s unclear how fast it will be.
“Radios went from 10% to 90% penetration in the United States in a decade. Washing machines, to reach that level of penetration, needed a period of several decades,” he said. “So when we think about anthropic AI and generally in the business environment, what does that adoption curve look like?”
Conflict with the Pentagon and “walking the border”
Anthropic has not been without controversy. At the beginning of this year he was involved in a conflict with the Pentagon on how their models can be used. The Pentagon wanted the military to be able to deploy Anthropic’s models without restrictions, but Amodei drew a line between fully autonomous weapons and mass American surveillance. The Pentagon later labeled Anthropic a “supply chain risk,” a designation previously reserved for foreign adversaries.
And skeptics have criticized Amodei’s rapid push to develop AI despite its own Dire warnings about the risks of technology.. In September, Amodei wrote an essay urging his AI competitors to “rhythm the border“, or slow down the development of the most advanced models and ensure that security incidents are reported and analyzed by external auditors.
That trial followed a summer in which it was revealed that AI agents from Anthropic’s lab, as well as OpenAI and other competitors, had hacked other companiesunleashing a frenzy of concern the risks of dishonest agents. This week, California sued OpenAI about its agents hacking Hugging Face, an open source software platform.
Anthropic’s IPO prospectus devoted dozens of pages to potential risks, including the possibility that people could lose control of AI or that AI could take harmful actions on its own.
The company also flagged the risk of increased scrutiny or political backlash. In fact, the IPO comes just weeks after the US midterm elections, which analysts say likely heralds a period of increased scrutiny and oversight for AI companies. Data centers became a Hot topic in this electoral cycleand technology stocks It sank after the election as investors anticipated that the incoming Congress would crack down on the industry.
Still, Anthropic’s prospectus was optimistic about the development and spread of AI. “We believe this abundance of AI will rapidly transform all sectors of the global economy in ways greater than the impact of industrialization, electricity, semiconductors and the Internet,” he said.
OpenAI was on track to also list shares this year, but CEO Sam Altman postponed those plans, likely until next year, after the Hugging Face incident. At Monday’s Semafor conference, Altman promised: “At some point we will go public.”
“The world as a whole would like to be exposed to this,” he continued. “And I think that’s, you know, directionally a good thing to do.”
Meanwhile, shares of SpaceX, the first of this year’s AI mega-IPOs, are trading near their initial price of $135. After the company’s IPO in June, shares jumped to over $200, then fell to around $108 before regaining ground.
Anthropic financially supports NPR.