SpaceX officially announced the acquisition of xAI, intending to build a "vertically integrated innovation engine" connecting the ground and space. In order to completely solve the bottleneck of energy and heat dissipation faced by the expansion of AI computing power, SpaceX proposed an ambitious "orbital data center" plan, relying on Starship's ultra-high capacity every hour to deploy millions of satellites in orbit and build a space computing network with an annual increase of 100GW computing power.
SpaceX officially announced the acquisition of xAI, aiming to build "the most ambitious vertically integrated innovation engine on earth (and in space)". This strategic merger is not a simple capital operation, but a deep breakthrough in the current physical bottleneck of the AI industry. With the exponential growth of AI model parameters, the ground data center is facing insurmountable challenges of power supply and heat dissipation, and it is difficult to support the large-scale expansion of AI in the future only by the existing energy network of the earth.
This integration will achieve multi-dimensional business collaboration:
Energy and Space Breakthrough: Using the endless solar energy and low temperature environment in space to solve the energy consumption pain point of xAI future supercomputing cluster.
Communication network convergence: combining Starlink's global coverage and Direct to Cell technology, it provides real-time access capability for AI applications at all times.
Data closed loop: integrating the real-time information flow of X platform with the sensor data of SpaceX, providing a unique physical world data source for xAI training.
SpaceX elaborated the technical path of "Orbital Data Centers" in the official update. The official pointed out that using only one millionth of the solar energy is enough to obtain millions of times more energy than the current total of human civilization through space facilities, and space is the only long-term logical destination for the expansion of AI computing power.
To achieve this vision, SpaceX has published a series of key engineering objectives and data:
| Key indicators | Planning objectives and data |
| Radio frequency | It is planned to launch Starship every hour and carry about 200 tons of payload at a time. |
| Calculation force increment | According to the calculation of 100 KW computing power per ton of satellite, 100 GW AI computing power can be added by launching one million tons of load every year. |
| Satellite performance | The upcoming V3 Starlink satellite will have a single launch capacity of more than 20 times that of V2. |
| Long term vision | Using lunar manufacturing and electromagnetic ejection, the long-term goal is to deploy 500-1000 TW AI computing clusters into deep space every year. |
SpaceX predicts that in 2-3 years, space will become the lowest cost path to generate AI computing power. This "orbital data center" constellation will be composed of millions of satellites, which will realize zero-carbon emission operation by continuous illumination in orbit without consuming precious water resources of the earth for cooling.
From an industry perspective, SpaceX's acquisition of xAI will have a far-reaching impact:
Recent integration: AI technology will quickly penetrate into spacecraft design and satellite chain network optimization, forming a cross-domain collaborative paradigm of "AI+ Aerospace".
Medium-term supply: If Starship's transportation capacity is fulfilled as scheduled, the rail data center will effectively alleviate the hard constraints of ground power on AI development and reshape the global computing power supply pattern.
Risk challenge: The project faces extremely high engineering complexity, including on-orbit maintenance, space debris management and complex international space regulatory game. The deployment of millions of satellites has put forward unprecedented requirements for space traffic managemen.
