Musk’s $60 Billion Bet on Cursor

After the news broke that SpaceX secured a $60 billion option to acquire Cursor, discussions erupted about whether this deal is worth it. After all, a company with an annual revenue of only $2 billion is asking for 30 times its yearly income. However, one detail that has been rarely discussed is that over half of Cursor’s revenue comes from Fortune 500 companies, with corporate clients accounting for 60%. What truly attracts Musk is not the 7 million individual users, but the continuous stream of real interaction data from large enterprises in development scenarios. This seemingly crazy deal is fundamentally about the dual pursuit of computing power and data, but can the underlying capital logic support a trillion-dollar valuation?

Not Just About Products, But Real Development Scenario Data

Many commentators suggest that Musk is rushing to fill xAI’s programming gap by acquiring an existing AI programming product. However, if he merely needed a product, SpaceX could easily develop one from scratch with its computing power and funds, without needing to pay a $10 billion “entry fee” for an acquisition.

What is truly valuable is the real developer interaction data that Cursor has accumulated over four years. Unlike training samples from public datasets, this data represents the complete behavior chain of developers in their daily work, from reviewing and modifying AI-generated code to final adoption.

This real data, which includes human feedback loops, cannot be replaced by any public dataset. For training programming large models, knowing where developers made changes and why is far more valuable than having a thousand lines of well-annotated example code.

Cursor holds this data goldmine but cannot fully exploit it due to a lack of computing power. Relying on large models from OpenAI and Anthropic not only incurs a fee but also limits access to complete data training autonomy, let alone independent iteration of its own models.

It’s like having a key to open a treasure but lacking the strength to pry open the stone door. SpaceX, with its Colossus supercomputer and 200,000 H100 equivalent GPUs waiting to process data, is a perfect match for Cursor.

Closed-Loop Evolution as the Core Barrier of This Era

This transaction shatters a common industry misconception: in today’s AI entrepreneurship, the competition is no longer about who can produce a product first, but who can effectively run a data-model-product self-evolution closed loop.

Having a product does not necessarily create a barrier, but a continuously self-evolving loop is a moat that cannot be taken away by others.

Most companies currently using AI are still at the stage of “using AI to improve efficiency”: using AI to write code, create plans, and generate content, with no systematic collection of these interaction data to feed back into model optimization.

Even many established AI products only represent a “half-closed loop”: products generate data, but that data does not continuously feed back into model iteration, either due to insufficient computing power or a lack of awareness of the long-term value of data.

The collaboration between Cursor and xAI completes this loop: users generate interaction data on Cursor, xAI uses supercomputing to clean and train the model, and the optimized model is then returned to Cursor to serve users. This entire process is a continuous cycle, enhancing model capabilities, improving user experience, and increasing data volume.

Once this positive cycle is established, the cost for later entrants to catch up will increase. This is also the core reason Musk is willing to spend $60 billion—he is not buying existing users but an entry point into an already functioning evolution system.

Capital Puzzle More Urgent Than Technical Loop

It must be acknowledged that, compared to filling the AI technology gap, the more pressing mission of this transaction is to help SpaceX complete the valuation puzzle for its IPO. The entire capital market is aware that SpaceX aims for the largest IPO in history, and the $17.5 trillion valuation requires a new growth story to support it.

The old stories of rocket launches and Starlink are already familiar to the market, and the growth ceiling is clear. Even if profits reach $8 billion by 2025, it won’t support a $17.5 trillion valuation. A sufficiently attractive AI story must be added to encourage investors to buy into the future.

In February 2026, SpaceX completed its acquisition of xAI, integrating AI business into the listed entity. Now, it just needs a tangible AI asset with users and revenue to fill the narrative. Cursor perfectly meets all requirements:

• It has real annual revenue of $2 billion, not just a concept of burning money without profit.
• It covers more than half of the Fortune 500 companies, with sufficient B-end recognition.
• It has a clear synergy with xAI’s computing power, enabling a complete narrative of “space + AI.”

From this perspective, even if the $60 billion valuation is indeed inflated, it is still a bargain for Musk. As long as Cursor is integrated into the listed entity, SpaceX’s valuation can be pushed up another notch, with the increased market value far exceeding the cost of the acquisition.

Industry Signals Behind the Bubble Worth Watching for All Entrepreneurs

Many criticize Musk for exploiting the capital market, inflating a bubble with an AI story. However, from another angle, this transaction provides a clear direction for all AI entrepreneurs.

The future valuation of AI companies will no longer depend on user numbers or revenue growth rates, but on whether they can construct a continuously self-evolving data closed loop. As long as you possess real, high-frequency, continuously updated interaction data, finding buyers willing to pay a premium will not be a problem.

Conversely, if your AI product merely wraps around someone else’s model without accumulating its own exclusive data closed loop, no matter how fast your current revenue grows, there is no real barrier in the long run. Once the model provider decides to enter the product space, you may find yourself in a very passive situation, just like Cursor being squeezed by Claude Code.

Cursor recognized this and chose to collaborate with SpaceX to secure stable computing power and free itself from dependence on competitors’ models. This is actually a life-saving move for Cursor.

The entire AI industry is shifting from “competing on model parameters” to “competing on data closed loops.” Computing power is becoming easier to acquire, while high-quality real interaction data will become increasingly valuable. Musk’s move is essentially a bet on this industry turning point.

As for whether the $60 billion valuation is a bubble, the answer is no longer important. When all players begin to scramble for data and compete on closed loops, the next round of competition in the AI industry has already begun. Those still relying on wrapping products for quick profits have little time left.