It all happened in one day.
Three of the hottest startups in history moved closer Wednesday to becoming trillion-dollar companies. With a fresh filing, SpaceX showed the path to a landmark IPO likely to mint Elon Musk as the world’s first trillionaire. Anthropic showed potential to turn a profit far sooner than expected, and OpenAI prepared to file for its public debut as soon as Friday, The Wall Street Journal reported.
Meanwhile, two of the world’s biggest companies faced the challenging realities of the AI age, with Nvidia meeting investor apathy after posting record sales of $82 billion and Meta Platforms starting to lay off 8,000 employees.
Meta Chief Executive Mark Zuckerberg in a memo to staff Wednesday called AI “the most consequential technology of our lifetimes.”
“I’m optimistic about everything we’re building,” he said. “But success is not a given.”
Such is the mind-boggling nature of the AI boom, with its vast capital needs, a constantly evolving cast of winners and losers and the ability to quickly make or break whole industries and reset economic sentiment.
“There’s no status quo. Everything is evolving so quickly,” said Jeffrey Bernardo, chief executive of Augustine Asset Management, which holds shares in a number of large global technology companies. “Demand for these products is growing and exceeding supply, and it’s all driving improvements in productivity.”
Bernardo’s firm is “open-minded” about investing in the startups when they go public, but he plans to study IPO filings carefully before making any investment decisions.
In the first phase of the AI boom, Nvidia used its edge in making cutting-edge chips to become the world’s most valuable company. Its ability to deliver the most essential ingredient for AI development—computing firepower—drew comparisons with the makers of “picks and shovels” in the gold rush who made a fortune catering to miners.
As frontier AI companies continue to unleash powerful “agents” that can work on tasks for hours with limited human involvement, especially in coding, demand for computing power and the chips that enable it is expected to continue to grow.
While Nvidia’s dominance is expected to continue, Wednesday’s results showed how investor sentiment is changing, with expectations for sales and margins constantly rising. Nvidia beat Wall Street expectations, raised guidance, announced plans to buy back $80 billion in stock and saw its shares drop Wednesday afternoon.
Investors have so far been more patient with startups like SpaceX, OpenAI and Anthropic, which have established track records of losing billions of dollars. In their paths to IPO glory, SpaceX, OpenAI and Anthropic have each defied a host of expectations and long-held beliefs about the possibilities of the new technologies they are building.
The fates of all three companies are intertwined, as their leaders—including Musk, OpenAI’s Sam Altman and Anthropic’s Dario Amodei—have sparred with one another over a technology they helped launch together.
As a lawsuit brought by Musk against Altman and OpenAI recently showed, the two were once allies in the AI race. Amodei was also a top executive at OpenAI, which launched more than a decade ago as a nonprofit dedicated to improving humanity.
As OpenAI developed, its leaders realized the company would need potentially billions of dollars to pay for the computing firepower required to train and improve AI models. The company’s move to convert into a for-profit business led to a falling-out with Musk, who alleged in his lawsuit that OpenAI “stole a charity.” A jury and a federal judge sided with OpenAI; Musk has vowed to appeal.
Musk launched xAI to compete, but its chatbot Grok has fallen behind in the race after failing to gain a wide user following, outside of Musk’s social-media platform X. Ahead of the SpaceX IPO, Musk struck deals with several OpenAI competitors to level the playing field, including Cursor, a startup that specializes in coding tools, and Anthropic.
Fresh off a $45 billion deal with Musk and SpaceX that helped save it from a computing crisis, Anthropic told investors that it expects to generate its first-ever quarterly operating profit in the three months ending in June. While it still expects to lose money this year overall, the company’s forecast surprised many close observers of its finances, as most AI companies are still on track to spend far more than they take in for years because of high computing costs.
Although expectations and investor enthusiasm remain high for all three companies, some have warned that the novel technology comes with new risks.
“The market has priced AI’s potential but has not factored in the cost of the vulnerabilities these systems could create at scale, especially with agentic AI,” said Navrina Singh, chief executive of Credo AI, a startup focused on making AI systems reliable and trustworthy. “Capital is racing toward AI at full speed, but none of that is priced into a single one of these valuations.”
Write to Bradley Olson at bradley.olson@wsj.com