Elon Musk and his investors are betting $6 billion that bigger AI will continue to be better AI.
The big picture: Musk long ago parted ways with the other co-founders of OpenAI — but his AI startup, xAI, is borrowing OpenAI’s “if you build it larger, it will get smarter” strategy.
- That game plan has worked impressively so far, as OpenAI has rolled out one eye-popping demo after another, and new advances are expected from the firm’s next big model, GPT-5, by year’s end.
Yes, but: There’s no guarantee such progress will continue, and if it doesn’t, the massive influx of cash to these companies, and the stratospheric market valuations they hold, could abruptly end.
State of play: Very few existing companies — perhaps only Google, Meta and Anthropic — are credibly taking on OpenAI when it comes to making the biggest large language models, known as frontier models.
- Those competitors have been at it for years. xAI is less than a year old.
- Hiring AI talent costs a fortune right now, and the tools those developers need — chiefly in the form of advanced AI chips — are pricey and scarce.
$6 billion in a single venture round is a ridiculously large sum. But this year, OpenAI CEO Sam Altman has been talking up efforts to raise trillions of dollars to build the chip-making infrastructure he thinks will be needed to support advanced AI.
- The Information recently reported that xAI plans to build a massive new supercomputer — the “Gigafactory of compute” — possibly in partnership with Oracle.
- Such a project could easily burn through this investment round before ever coming online.
Catch up quick: Musk has a reputation for taking on giant missions and delivering results.
- With Tesla, he popularized electric cars in the U.S., and with SpaceX, he built private-sector rockets that now dominate the American market.
- But Musk’s record is also full of failures, especially when it comes to software projects — like Tesla’s ill-starred autonomous-driving features or Musk’s still-faltering effort to revitalize Twitter’s business under its new name, X.
- More recently, Musk’s choices — particularly his tolerance and even encouragement of neo-Nazis, white supremacists and antisemites on X — have cast a pall over his reputation.
- This deal shows that none of this has yet crimped his capacity to raise large sums of capital.
The intrigue: One key advantage xAI will have is its access to user data from X.
- Musk is known for blurring the boundaries between the companies he owns and runs.
- Real-time access to X’s data could help xAI make its chatbot newsier and more plugged in.
- Alternately, the faster xAI metabolizes X users’ posts, the more likely it is to recycle extremist misinformation as gospel.
Between the lines: Some observers — including OpenAI itself, in its defense against a lawsuit by Musk — have suggested Musk envies OpenAI’s success with ChatGPT and is building xAI to try to seize back the limelight.
- Musk and his supporters say that both OpenAI and Google are jeopardizing the future of AI and humanity by attempting to build the technology with guardrails against hate speech.
- The pro-guardrails camp believes they’re just preventing the new chatbots from saying horrifying things, but Musk denounces these efforts as a “woke” and “deadly” form of lying.
Fun fact: The company’s funding announcement said that “xAI is primarily focused on the development of advanced AI systems that are truthful, competent, and maximally beneficial for all of humanity” — but also that “the company’s mission is to understand the true nature of the universe.”
The bottom line: Musk has always overpromised, and recently he has underdelivered enough to raise big questions about xAI — but the big money is still lining up for him.
