SpaceX plans to deploy 1GW of space computing power next year; the server cabinet and optical communication industries are "moving in response to the news."

SpaceX plans to deploy 1GW of space computing power next year; the server cabinet and optical communication industries are "moving in response to the news."

```

The listing of SpaceX not only reshaped the global tech giants’ market cap landscape, but its strategies to enter the space AI field have directly ignited a new wave in the hardware manufacturing and optical communication industry chain.

On the 12th local time, SpaceX officially landed on the Nasdaq Stock Exchange, raising about $75 billion. On its first trading day, it closed up nearly 20%, with its market value soaring to $2.11 trillion. On the same day as its listing, Musk responded to Nvidia’s congratulations on the social platform X, explicitly stating that he will push cooperation between the two sides to a "new level."

Behind this statement is SpaceX’s massive blueprint for space computing power. WallStreetCN previously reported that on the eve of SpaceX's listing, Musk for the first time disclosed the core technical specifications of its first-generation dedicated AI compute satellite "AI1", planning to directly deploy AI datacenters in space orbit, and has already signed multi-billion dollar cloud computing orders with companies such as Google and Anthropic.

The massive compute deployment plan quickly translated to capital markets and the supply chain. Industry analysis believes that with the clarity of demand for space server cabinets and inter-satellite laser communication, server assembly giants such as Foxconn and Quanta, as well as key optical communication component suppliers like United Fiber and LumenTech, are ushering in substantial order expectations and business growth.

AI1 Satellite Unveiled: 1GW Space Compute Power to Deploy Next Year

The first dedicated AI compute satellite, AI1, launched by SpaceX aims to overcome the limitations of ground infrastructure.

Each satellite supports up to 150 kW peak compute power, equipped with liquid-cooled radiators, meteoroid protective layer, centralized computing modules, and deployable solar panels. These units will be mass-produced in SpaceX’s Gigasat factory in Texas.

According to SpaceX’s plans, the company will fully enter the space AI market and realize an annualized deployment of 1GW (1 billion watts) space AI compute power each year by the end of 2027, gradually promoting commercialization in 2028, and ultimately boosting the scale of compute power to 100GW by 2030.

Annual Demand for 7,400 Cabinets: Server OEM Giants Benefit

To achieve such a scale of space compute network, creating AI server cabinets suitable for the space environment is the foremost task.

According to media citing industry evaluation, for each launch of an AI1 satellite, SpaceX will send a Nvidia GB300 AI server cabinet into space. Based on the approximate power consumption per GB300 cabinet of 135 kW, deploying 1GW of compute power per year equals the annual demand for about 7,400 GB300 AI server cabinets.

Driven by huge hardware demand, the related supply chain has begun operating. The industry expects that preliminary assembly of related servers will continue to be handled by ODM partners such as Foxconn, Quanta, and Wistron. After completing the basic assembly, these OEMs deliver it to SpaceX, which performs final satellite integration and space adaptation adjustments.

1Tbps Bandwidth Interconnection: Optical Communication Industry Chain Sees New Growth

Aside from the core server compute units, data transmission between AI1 satellites likewise relies on high-end hardware support.

SpaceX plans to achieve direct interconnection between satellites via advanced laser optical paths, constructing a high-transmission, low-latency data network in space.

This design allows large-scale AI computing tasks to collaborate efficiently and exchange data among different satellite nodes, speeding up computation. Currently, the latest generation of SpaceX’s Starlink satellites already supports single orbit bandwidth of 1Tbps via laser optical links.

According to media citing industry analysis, the main device suppliers for this feature are international laser giants Coherent and Lumentum. Upstream supply chain prosperity is also driven directly, including indium phosphide (InP) epitaxial wafer supplier United Fiber, and LumenTech, which is Coherent’s laser packaging OEM; both will benefit from this expansion of space compute power.

Monetizing Compute Power: From Rocket Launches to Cloud Service Provider Transformation

While promoting hardware deployment, SpaceX has begun to convert its computing power reserves into commercial orders, accelerating its transition to basic AI compute power suppliers.

On the eve of its listing, SpaceX signed a cloud service agreement with Google to provide compute power equivalent to 110,000 Nvidia GPUs and related components, with a monthly fee of $920 million.

Additionally, SpaceX offers AI startup Anthropic access to 220,000 Nvidia GPUs, including H100, H200, and GB200, with the contract charging $1.25 billion per month, totaling $15 billion per year.

In addition to compute power leasing, SpaceX is also expanding its ecosystem horizontally. The company is collaborating with Tesla and Intel to quickly push forward the “Terafab” project, trying to build a multi-layer commercial ecosystem across space, compute power, and manufacturing.

Risk Warning and DisclaimerThe market has risks, investment needs caution. This article does not constitute personal investment advice, and does not take into account the individual user's specific investment objectives, financial status, or needs. Users should consider whether any opinions, views, or conclusions in this article fit their particular circumstances. Investment based on this is at your own risk. ```