This column is from Trendlines, my business newsletter that covers the forces shaping the economy in Boston and beyond. If you’d like to receive it via email on Mondays and Thursdays, sign up here.
In 2004, when Google sold stock to the public for the first time, it famously told investors its motto was “Don’t be evil,” and that it aimed to be “a company that does good things for the world even if we forgo some short-term gains.”
We know how that went. (See: a string of antitrust lawsuit losses, a surveillance-based business model, and its power to devastate businesses with a single tweak to its algorithm.)
But the company, which later rebranded as Alphabet, has delivered phenomenal gains for shareholders, posting an average annual return of 25.5 percent over 22 years.
Now comes artificial intelligence developer Anthropic, which is channeling Google’s founding ethos as it launches what could be the largest initial public offering ever.
In IPO paperwork filed Monday, the maker of the Claude chatbot said that “for a technology as consequential as AI, entrepreneurial capitalism must be balanced and augmented by mechanisms that protect the common good and share the upside as widely as possible.”
Anthropic’s stock market debut, which is expected before Thanksgiving, could raise as much as $100 billion, topping SpaceX’s record-setting deal in June. It will be the first IPO by a standalone AI “frontier lab” producing the most advanced AI systems.
The windfall will help fund Anthropic’s race against OpenAI, Alphabet, and others to artificial general intelligence, an AI powerful enough to learn, reason, and perform any knowledge-based task as well as a person can.
Whether Anthropic can uphold its ideals any better than Google did remains to be seen. But its prospectus offers the fullest look yet at its ambitions, finances, and the pressure it will face to keep Wall Street happy. Here are some of the key disclosures.
Remarkable growth
Anthropic, founded just five years ago by seven former employees of rival OpenAI, said its revenue soared to $16.3 billion in the first half of this year from $1.16 billion in the year-ago period.
As of Sept. 30, the company said, about 11,900 of its customers were spending more than $100,000 a year, up from 1,500 a year ago.
The company recorded a first-half net loss of $80.7 billion, largely because of paper losses tied to debt that can be exchanged for stock.
Using a nonstandard accounting measure, the company reported adjusted income of $2.67 billion in the first six months of the year compared with an adjusted loss of $1.74 billion a year earlier.
Cash cache
Anthropic is set to go public with more money in the bank than any other startup in history.
It was sitting on a surprisingly large $131.7 billion in cash, cash equivalents, and short-term investments as of June 30, more than any other US company except Berkshire Hathaway and Alphabet, according to CNBC and S&P Capital IQ.
Huge commitments
The company is spending massively on the computing resources it needs to develop and run its models.
So-called compute and infrastructure expenditures were almost $9 billion in the first half, nearly four times the figure from a year earlier.
As of June 30, Anthropic had committed to more than half a trillion dollars in future spending: about $321 billion in leases (at face value) and another $266 billion in other contracts with fixed or guaranteed minimum payments. Since then, it has signed or revised agreements that add at least $245 billion more.
Grand ambitions, daunting risks
Anthropic believes AI will radically change the world.
“If managed well, it could rapidly accelerate scientific innovation, greatly increase economic growth rates, bring cures for major diseases within reach, and solve problems that have resisted generations of human effort,” it says in the IPO filing.
But the document also includes more than 90 pages of risks to its business and dangers AI poses to the world. Among them, AI could:
- Break free of human control and act independently, leading to “catastrophic or existential risks to humanity.”
- Enable the production and use of chemical, biological, and malicious cyber weapons.
- Cause rapid job losses and significant economic dislocations.
These are more than cover-your-butt disclosures. AI lab insiders and outside experts have issued the same warnings for years.
Unusual corporate governance
Anthropic is a public benefit corporation, a structure that requires its board of directors to balance shareholders’ financial interests with the company’s mission of AI safety. There are fewer than 20 publicly traded PBCs, most notably eyewear company Warby Parker.
The company also created a unique “long-term benefit trust,” an independent body that will appoint a majority of its board. But Anthropic’s founders — including CEO Dario Amodei — retain significant control through a special class of stock that holds 50.1 percent of total voting power.
Final thought
Analysts estimate Anthropic’s initial public market value at $1.5 trillion to $2 trillion. At the midpoint of that range, it would rank among the country’s 10 most valuable companies.
That’s an astounding feat for a company that didn’t exist six years ago.
A key question: Can Anthropic grow revenue and earnings fast enough to satisfy investors without sacrificing its commitment to AI safety?
Or will it end up like Google, which eventually let profit trump ethics?
Larry Edelman can be reached at larry.edelman@globe.com.