Google will pay SpaceX $920M per month for compute

TL;DR

Google has agreed to pay SpaceX $920 million monthly from October 2026 through June 2029 for access to around 110,000 NVIDIA GPUs and related hardware. This deal signals a major move in AI compute capacity ahead of SpaceX’s IPO, with Google seeking additional capacity to meet surging demand.

Google will pay SpaceX $920 million per month from October 2026 through June 2029 for access to approximately 110,000 NVIDIA GPUs, CPUs, and related hardware, according to a regulatory filing. This deal underscores Google’s effort to expand its AI compute capacity amid rising demand and coincides with SpaceX’s preparations for its historic IPO.

The agreement, disclosed in a filing with the Securities and Exchange Commission, stipulates that Google will pay SpaceX $920 million monthly for nearly three years. The deal covers access to a significant amount of AI hardware, including NVIDIA GPUs and CPUs, although SpaceX did not specify which of its data centers Google will utilize. It appears to provide roughly half the compute capacity that SpaceX has allocated to another client, Anthropic, which pays $1.25 billion monthly for exclusive access to its Colossus 1 data center near Memphis.

The deal is structured with a flexible termination clause, allowing either party to end the agreement with 90 days’ notice after December 31, 2026. Google’s access will ramp up through September 2026 at a reduced fee, with provisions for immediate termination if SpaceX fails to deliver the agreed compute capacity by then. The announcement comes just a week before SpaceX’s planned IPO, which aims to raise approximately $75 billion at a valuation of around $1.75 trillion, making it the largest in history. Google, a longstanding investor in SpaceX, is expected to see its stake worth over $100 billion post-IPO.

Why It Matters

This development highlights the increasing importance of AI compute infrastructure in the tech industry, with Google seeking to bolster its capacity to meet surging demand for its AI products. The deal also signals a strategic partnership that could influence the future of orbital data centers, a space both companies are reportedly exploring. For SpaceX, the deal provides a substantial revenue stream ahead of its IPO, potentially increasing its valuation and market influence. For readers, it underscores the growing scale of AI infrastructure investments and the strategic moves by major tech firms to secure hardware capacity in a competitive landscape.

HPE NVIDIA Tesla V100 32GB HBM2 PCIe 3.0 x16 Passive GPU Computational Accelerator for AI Machine Learning HPC Deep Learning 699-2G500-0216-400 (Renewed)

HPE NVIDIA Tesla V100 32GB HBM2 PCIe 3.0 x16 Passive GPU Computational Accelerator for AI Machine Learning HPC Deep Learning 699-2G500-0216-400 (Renewed)

NVIDIA Volta GV100 Architecture — 5,120 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112…

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Background

In recent months, SpaceX has been positioning itself as a major player in AI compute infrastructure, with previous deals such as the one with Anthropic, which pays $1.25 billion monthly for exclusive access to its Colossus 1 data center. Elon Musk has indicated plans to reserve the Colossus 2 data center for xAI, SpaceX’s AI venture. Google’s involvement reflects its need to expand AI hardware capacity, especially following the launch of its Gemini Enterprise platform, which has experienced higher-than-expected demand. The deal with Google is part of a broader trend of large corporations investing heavily in AI infrastructure, with Alphabet committing over $180 billion in capital expenditures this year alone.

“This is a short-term, timely agreement to ensure we have bridge capacity to meet surging customer demand for our agent platform, Gemini Enterprise.”

— a Google representative

GDSTIME 120mm 48V Fan, 120x120x25mm Dual Ball Bearings DC Brushless Cooling Fans 2Pin for Data Center Servers, Electronics Cabinets

GDSTIME 120mm 48V Fan, 120x120x25mm Dual Ball Bearings DC Brushless Cooling Fans 2Pin for Data Center Servers, Electronics Cabinets

High Performance Cooling: 120 mm 48v fan widely used in data center servers, electronics cabinets, inverters, distribution box,…

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

What Remains Unclear

Details about which specific data centers Google will use remain undisclosed, and it is unclear how the deal will impact other existing agreements or the broader AI compute market. The future of the partnership beyond the initial term and the potential for further collaborations are still uncertain. Additionally, the impact of this deal on SpaceX’s IPO plans and its long-term data infrastructure strategy is yet to be clarified.

Tecmojo 12U Open Frame Network Rack for IT & AV Gear, AV Rack Floor Standing or Wall Mounted,with 2 PCS 1U Rack Shelves & Mounting Hardware,Network Rack for 19" Networking,Audio and Video Device

Tecmojo 12U Open Frame Network Rack for IT & AV Gear, AV Rack Floor Standing or Wall Mounted,with 2 PCS 1U Rack Shelves & Mounting Hardware,Network Rack for 19" Networking,Audio and Video Device

【Powerful Load-bearing】12U Network Rack Open Frame is constructed from durable cold rolled steel; Rack shelf supports enhance stability,…

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

What’s Next

Next steps include the ramp-up of Google’s access through September 2026, with potential adjustments if capacity delivery falls short. The deal’s execution will be closely monitored, especially as SpaceX approaches its IPO, scheduled for later in 2026. Further announcements may clarify the specific data centers involved and any additional collaborations or infrastructure projects, including orbital data centers.

Kubernetes for AI Infrastructure: The Engineer’s Guide to GPU Orchestration and Reducing MLOps Overhead in Production (Architecting Enterprise Agents Series)

Kubernetes for AI Infrastructure: The Engineer’s Guide to GPU Orchestration and Reducing MLOps Overhead in Production (Architecting Enterprise Agents Series)

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

Why is Google paying such a high monthly fee to SpaceX?

Google is seeking to expand its AI compute capacity rapidly to meet increasing demand for its AI products and services, such as Gemini Enterprise, and to ensure it has sufficient infrastructure during a period of surging customer interest.

How does this deal compare to SpaceX’s other compute agreements?

It is similar in scope and duration to a previous deal with Anthropic, which pays $1.25 billion monthly for exclusive access to a dedicated data center. Google’s deal appears to be for roughly half that capacity, reflecting different client needs and capacity allocations.

What role does this deal play in SpaceX’s IPO plans?

The deal provides SpaceX with a significant revenue stream ahead of its IPO, which aims to raise approximately $75 billion and achieve a valuation of around $1.75 trillion. It demonstrates the company’s ability to monetize its infrastructure before going public.

Will Google’s access to SpaceX’s data centers be permanent?

No, the agreement includes a cancellation clause allowing either party to terminate with 90 days’ notice after December 31, 2026, providing flexibility based on evolving needs and capacity delivery.

Source: Hacker News

You May Also Like

Intel Core Ultra 5 250K Plus vs AMD Ryzen 5 7600X3D faceoff — Battle for the fastest mid-range gaming CPU

Intel’s Core Ultra 5 250K Plus and AMD Ryzen 5 7600X3D compete in gaming and productivity. Here’s what is confirmed and what remains uncertain in their showdown.

PostHog will train AI models with your data (opted-in by default)

PostHog announces plans to train AI models on user data, with default opt-in for US users, aiming to enhance product capabilities and automate analysis.

AI is a technology not a product

Experts emphasize AI as a technology that enhances products, not a standalone product itself, challenging hype around AI-driven innovations.

Revamped Siri will reportedly offer auto-deleting chats

Apple’s upcoming Siri update in iOS 27 will include an option for users to auto-delete chat histories, emphasizing privacy over data retention.