Key Takeaways
- From Underperforming to Underwriting: Despite X (formerly Twitter) ad revenue being cut in half, Elon Musk rolled its data and the lagging Grok AI project into a deal that pleased investors by creating a massive, unexpected revenue stream.
- The 'Colossus' Power Play: Musk built the world's largest GPU cluster, dubbed 'Colossus,' to train Grok. When Grok fell behind competitors like ChatGPT and Anthropic, he didn't scrap the investment.
- Rent to Your Rivals: Instead of a sunken cost, Musk's XAI became an Airbnb for AI compute, renting out Colossus to competitors like Anthropic and, astonishingly, Google. Sam Parr notes Google signed up to spend “a billion dollars a month” with them.
- Failing Forward with Big Assets: This strategy exemplifies a unique approach to entrepreneurial failure: instead of cutting losses on a struggling project, identify its underlying valuable components and monetize them aggressively, even if it means serving rivals.
The AI Compute Airbnb
Imagine building the world's most powerful data center for your own ambitious AI project. Then, your project starts to lag behind the competition. Most founders would see a colossal sunk cost, a monument to a misstep. Not Elon Musk. As Shaan Puri explained on My First Million, Musk turned this potential setback into a “billion-dollar-a-month” opportunity.
Musk's AI venture, Grok, was created to compete with OpenAI's ChatGPT and Anthropic. To train Grok, he built what Puri calls "Colossus," the largest GPU cluster globally. But Grok hasn't dominated the market. Puri observed, "he used his Grock asset where he's trying to build a chat GPT anthropic competitor even though it's way way behind he built the largest data center for training AI Colossus this."
Instead of letting those powerful, expensive GPUs sit idle or only serve a lagging product, Musk made a truly unexpected move. “He just turned it into Airbnb,” Puri remarked. “He started renting out Colossus to Anthropic and to Google.” This wasn't a small side hustle. Sam Parr jumped in with the stunning detail: “Google just signed up to spend a million dollars, a billion dollars a month with us to become a customer.”
Monetizing the 'Failure' of X
The story of Colossus isn't isolated. It fits into a broader pattern where Musk re-architects the value proposition of his assets. Consider X, the platform formerly known as Twitter. Since Musk bought it, its traditional ad revenue has declined significantly. Puri noted, “Twitter revenue is like half of what it was when he bought it.” Yet, Musk packaged X's data as a crucial training input for Grok, and by extension, for the larger XAI ecosystem.
This meant that even as a traditional business, X was struggling, its data became a raw material for a new, potentially massive industry. Puri summarized this ingenious approach: “he rolled it together with all this other great stuff so that the investors did well even though that business didn't do so hot uh since then.” The value wasn't just in the original ad platform, but in its potential as a data source and as a part of a much larger, more diverse portfolio of ventures that includes not only AI training but also, potentially, future data centers in space.
What to Do With This
This week, identify your company's biggest internal "cost center" or an underperforming project that consumed significant resources. Brainstorm five non-obvious ways to monetize its underlying components, even if it means working with current competitors or completely pivoting from its original intent.