Recently, OpenAI CEO Sam Altman (Sam Altman) has considered entering the space field, planning to acquire or cooperate to establish a rocket company to compete directly with Musk's SpaceX.

According to the Wall Street Journal, Altman reached out to at least one rocket startup called Stoke Space this summer to discuss speeding up the pace of cooperation in the fall. His proposals include: OpenAI's multiple investment rounds in Stoke Space; finally achieve control; The total investment could be in the billions.

Space Data Center: Altman's Bigger Ambitions

Altman's proposal is: OpenAI will invest in Stoke in stages, eventually achieving control, with a total investment that could reach billions of dollars.

Although discussions heated up last fall, people familiar with the matter said that negotiations have largely stalled.

If the deal is finalized, Altman will be able to participate in Stoke's ongoing Nova rocket project. However, building rockets is extremely difficult - not only is the technical threshold high and the regulation is complex, but it usually takes more than ten years to start from scratch, so it is difficult for new players to break through quickly. Currently, in addition to Stoke, companies such as Blue Origin and Rocket Lab are also challenging SpaceX's position.

Altman has repeatedly publicly stated that the computing power and power required for AI in the future will be extremely large, and the energy and environment on Earth may be unbearable, so he is optimistic about building data centers into space.

Proponents argue that data centers in orbit can be continuously powered by solar energy, which is both clean and efficient. Altman even half-jokingly said on the podcast: "Maybe in the future we will build a giant Dyson sphere around the solar system and find out that it doesn't make sense to put a data center on Earth."

While it sounds like science fiction, there is real action: Google and satellite company Planet Labs have signed a contract to launch two prototype satellites equipped with Google's AI chips in 2027; Several tech leaders, including Bezos, Musk, and Google CEO Pichai, have publicly recognized the potential of "AI computing in space."

| OpenAI is too ambitious, where does the money come from?

In September and October last year, when the AI boom was in full swing, OpenAI CEO Sam Altman announced a series of blockbuster collaborations involving Oracle (Oracle), planning to build AI chips and data centers on a large scale.

The market was excited at the time - Oracle and Nvidia's stock prices soared. But enthusiasm quickly cooled: Oracle shares fell about 19% and Nvidia fell 13% in the past month; Nvidia's CFO admitted this week that the $100 billion partnership with OpenAI has not yet been finalized.

More worryingly, OpenAI has committed nearly $600 billion in computing power purchases in the past few months, while the company expects revenue of only $13 billion this year. The huge funding gap has raised questions from the outside world: how does it pay these sky-high bills?

At the same time, competitors are catching up at an accelerated pace: AI company Anthropic, which is rapidly expanding among programmers and enterprise customers, poses a direct threat to OpenAI. Under pressure, OpenAI announced a "red alert" this week — as its core product, ChatGPT, is being robbed of users by Google's Gemini chatbot.

In order to fully respond, OpenAI has decided: to postpone the launch of new businesses (such as advertising); Let some employees temporarily transfer jobs and focus resources on improving ChatGPT.

|Launch an "all-out competition" with Musk

If Altman does cooperate with rocket company Stoke, his competition with Musk will extend from AI to space, brain-computer interfaces, social networks and other fields, turning into an "all-out showdown":

Altman is not only the CEO of OpenAI but also a veteran investor. He has been at the helm of Y Combinator, a well-known startup incubator (which invested in Stoke), and manages a large, low-key portfolio of over 400 companies.

Although he now has less personal investment, he does not shy away from using OpenAI's funds to support big projects - for example, earlier this year, he pushed OpenAI and SoftBank to jointly invest $18 billion in data center project Stargate.

In June last year, he even seriously thought about it on a podcast with his brother: "Should I build my own rocket company?" He also explained: "I hope that in the future, humans will use far more energy than the earth can provide - so we may eventually have to go to space. ”

Jingtai's view: Controlling an immature rocket company is equivalent to "betting the life of aerospace with AI money". Even if Altman is rich (OpenAI has more than $10 billion in cash on its books), it has to weigh the ten-year return cycle.

In the real predicament, OpenAI itself is in a "red alert": ChatGPT's market share has been eroded by Google Gemini, and it has launched an internal "red alert" to fight back; postponing new businesses such as advertising, and focusing on chatbots; 600 billion computing power commitment vs 13 billion annual revenue, the financial model has been questioned; Nvidia's CFO admitted this week that the $100 billion cooperation "has not yet been finalized."

In other words: the battle on the ground has not been won, so you are in a hurry to go to heaven? This may explain why the rocket negotiations are "suspended" - stabilize the core business first, and then talk about the sea of stars.