Live Nov 18 · Cohort 3Nov 18, 19:00 MSK — free hands-on AI agents live workshop · Live Cohort 3 starts Nov 23Join Nov 18 Live →Cohort 3 (Nov 23)
IndustryInsight · On RadarThe Verge→ original

SpaceX Launches $55 Billion Terafab AI Chip Factory

Elon Musk's SpaceX is launching Terafab, an ambitious project to build its own AI chip factory in Austin, Texas. The initial investment is at least $55 billion, and it could grow to $119 billion if the project is fully built out. The factory is designed to support compute workloads on the scale of 200 gigawatts per year.

AI-processed from The Verge; edited by Hamidun News
SpaceX Launches $55 Billion Terafab AI Chip Factory
Source: The Verge. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceThe Verge ↗
Editor-in-Chief
Key metric #1Mass production of specialized processors for AI acceleration
Key metric #2Scaling output to match fast-growing demand for chips used to train language models
Key metric #3Creating an independent source of chips, both for SpaceX's own needs and for sale to outside customers

SpaceX, led by Elon Musk, is preparing one of the largest investment projects in the company's history. According to The New York Times and CNBC, SpaceX will invest at least $55 billion in building a giant AI chip factory codenamed Terafab in Austin, Texas.

The scale of the investment: more than NASA

SpaceX has filed detailed tax incentive documents in Grimes County, Texas, laying out the full scale of its ambitious project. The initial construction phase will require an investment of $55 billion. However, the company states in the filing that if all additional phases are built, the total cost of the project could reach $119 billion.

For context: NASA's entire annual budget is roughly $25 billion. SpaceX's investment in a single factory is more than double that. It is comparable to the annual GDP of many countries and shows how seriously Musk takes the need for his own chip manufacturing capacity.

Funding on this scale reflects fierce competition for access to semiconductor fabrication in the age of the AI revolution. SpaceX makes no secret of its intention to become one of the key players in the AI accelerator market.

What Terafab will produce

Musk first publicly announced Terafab in March 2026, immediately laying out ambitious goals. The factory will produce specialized chips capable of supporting compute workloads totaling up to 200 gigawatts per year. This is not just a big number: it is a scale that could meet the needs of the largest AI models, cloud services and new generations of LLMs.

According to the plans:

  • Mass production of specialized processors for AI acceleration
  • Scaling output to match fast-growing demand for chips used to train language models
  • Creating an independent source of chips, both for SpaceX's own needs and for sale to outside customers
  • Potentially reducing dependence on NVIDIA and other traditional semiconductor suppliers

The factory will be located in a Texas technology hub, which offers convenient logistics and access to a skilled workforce.

Why Musk needs his own factory

SpaceX, Musk's xAI and his ties to OpenAI mean he needs enormous computing capacity. Training modern large language models requires thousands of specialized GPUs/TPUs. Right now, the main supplier of such chips, NVIDIA, cannot meet all of the demand, and prices remain high.

Building its own production solves several problems at once: SpaceX will be able to control prices, guarantee access to critical infrastructure and avoid supply delays. Amid geopolitical tension and trade wars, this becomes a strategic advantage.

Terafab fits a global trend in which large tech companies (Google, Meta*, Amazon, Apple) are starting to design their own chips instead of depending entirely on third-party manufacturers. Musk is simply moving faster and more decisively than the rest.

What this means

If SpaceX manages to deliver Terafab on the announced timeline, it will reshape the geopolitics of the semiconductor industry. The US would gain a major independent chipmaker, reducing the tech sector's vulnerability to sanctions, trade restrictions and disruptions in the global supply chain.

For NVIDIA's competitors, this could mean the start of a new era in which integrated companies (those with their own manufacturing) gain a serious advantage.

*Meta has been designated an extremist organization and is banned in Russia.

Story

Key Questions Answered

What are the key details on The scale of the investment: more than NASA?

SpaceX has filed detailed tax incentive documents in Grimes County, Texas, laying out the full scale of its ambitious project. The initial construction phase will require an investment of $55 billion. However, the company states in the filing that if all additional phases are built, the total cost of the project could reach $119 billion. For context: NASA's entire annual budget is roughly $25 billion. SpaceX's investment in a single factory is more than double that.

What Terafab will produce?

Musk first publicly announced Terafab in March 2026, immediately laying out ambitious goals. The factory will produce specialized chips capable of supporting compute workloads totaling up to 200 gigawatts per year. This is not just a big number: it is a scale that could meet the needs of the largest AI models, cloud services and new generations of LLMs.

Why Musk needs his own factory?

SpaceX, Musk's xAI and his ties to OpenAI mean he needs enormous computing capacity. Training modern large language models requires thousands of specialized GPUs/TPUs. Right now, the main supplier of such chips, NVIDIA, cannot meet all of the demand, and prices remain high. Building its own production solves several problems at once: SpaceX will be able to control prices, guarantee access to critical infrastructure and avoid supply delays.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…