Anthropic, maker of the Claude chatbot, filed Monday for what could be the largest stock offering in history. The company is expected to be valued at around $1.5 trillion, roughly eight times its worth a year ago.
Investors are being offered an extraordinary technology, but with little influence over its future.
1. AI is ‘more important than electricity,’ claims the seller
Anthropic takes in more revenue each day than it earned in the whole of 2023. The tech is powerful: NASA engineers have used its coding agent to plot a route for the Mars rover. Anthropic’s prospectus says that AI will transform the global economy “in ways greater than the impact of industrialization, electricity, semiconductors, and the internet.”
2. Powerful AI models might not pay off colossal bills
On paper, Anthropic lost $80.7 billion in the first half of this year. Most was an accounting charge, and the business itself still lost $693 million.
The IPO prospectus admits Anthropic must “grow our customer base and expand revenue rapidly” to pay its computing bills. Where it gets tough: Anthropic has promised Amazon and Google more than $220 billion over the next decade, largely whether it uses the computing power or not. Both own a slice of Anthropic and sell rival chatbots. Imagine Pepsi signing a ten-year lease on Coca-Cola’s bottling plants.
Most of Anthropic’s revenue comes from AI models that are less than six months old. Fall behind for a season, and the customers can leave, but Anthropic will still owe all that money.
3. The IPO invokes the end of the world
Most prospectuses warn that interest rates might rise. Anthropic acknowledges concerns, raised “both within our organization and by other researchers,” that advanced AI could “pose catastrophic or existential risks to humanity.” It then asks the public for money to build more of it.
The company also admits it may not know how fast it is going. Because each Claude AI helps build the next version of Claude, “we may not identify a significant acceleration in our rate of development until after it has occurred.” Would this be reasonable in any other industry? Carmakers, after all, rarely warn that the speedometer can’t keep up.
4. The Pentagon calls Anthropic’s safety feature a threat. The fight could cost the company
In February, Anthropic told the Pentagon it could not use Claude for mass surveillance of Americans or fully autonomous weapons – a safety feature. The Trump administration branded it a “supply-chain risk,” a label usually reserved for foreign adversaries.
A California judge ruled the Pentagon was punishing Anthropic for criticizing the government. In a separate case, a Washington appeals court let the Pentagon bar the company from its supply chain. The majority of judges reasoned that Anthropic could reprogram Claude to refuse tasks that the military might need. The company says the dispute could cost it serious money.
5. ‘Trust us,’ say the guardians
Anthropic is frank about what investors are buying. Ordinary shareholders, it says, will generally not affect any decision “regardless of how they vote.”
One special share will give the seven founders 50.1 percent of the vote; a trust of outsiders picks most of the board. Hopefully they will be wise guardians. If the end of the world ever comes up for a vote, the founders can decide on it, and everyone else will have no say.
Comments