Aakib Ansari.
Back to articles
News Brief

Anthropic Signs a $10 Billion, Six-Year Compute Deal With Norwegian AI Cloud Startup Volta

Md Aakib Ansari
Md Aakib AnsariWeb Developer & AI Tools Reviewer
2 min readCompany: Anthropic
Anthropic Signs a $10 Billion, Six-Year Compute Deal With Norwegian AI Cloud Startup Volta

Anthropic has signed a $10 billion, six-year computing services agreement with Volta Infra Holdings, an AI infrastructure startup that emerged from stealth on August 4, 2026, according to reporting by Bloomberg. Volta did not name the counterparty in its public announcement, referring only to a "leading AI lab," but Bloomberg and multiple subsequent reports identified Anthropic as the customer. Anthropic declined to comment.

The deal provides Anthropic with access to 133 megawatts of capacity at a data center in Tydal, Norway — a site chosen in part because it runs on 100% renewable hydroelectric power and was built in partnership with Bitdeer Technologies, a former Bitcoin mining company that has been pivoting its infrastructure toward AI workloads. The facility will be equipped with Nvidia's Vera Rubin chips, Nvidia's successor to the Blackwell generation and the chip architecture Anthropic will use for its next compute cycle.

Volta Infra itself launched with a $2.4 billion valuation and a notable investor lineup: Andreessen Horowitz, Altimeter Capital, Nvidia, and the family office of Michael Dell all backed the firm. The founding team came from Brookfield Asset Management, which has experience structuring long-duration infrastructure deals at scale — a background that maps onto what is, in practice, a six-year compute reservation rather than a traditional cloud services contract.

The deal follows a broader pattern of frontier AI labs locking in dedicated compute capacity outside of the major hyperscalers. Anthropic has separately established agreements with AWS and Google Cloud. The Volta deal adds a purpose-built, Nvidia-stocked facility in a jurisdiction with stable power and a favorable regulatory posture — and, notably, no exposure to the US grid congestion that prompted Texas to halt new data center approvals this week.

Frequently Asked Questions

What is the Anthropic Volta cloud deal?
Anthropic reportedly signed a $10 billion, six-year compute agreement with Volta Infra Holdings, giving it access to 133 megawatts of AI compute capacity at a data center in Tydal, Norway, equipped with Nvidia Vera Rubin chips and powered by 100% renewable hydroelectric energy.
Who is Volta Infra Holdings?
Volta Infra is an AI infrastructure startup that came out of stealth in August 2026 with a $2.4 billion valuation. It is backed by Andreessen Horowitz, Nvidia, Altimeter Capital, and Michael Dell's family office, and was founded by former Brookfield Asset Management executives.

Related Articles

Meta Releases Muse Glimmer: Apache 2.0 Licensed 30B Local Agent Model
News Brief2 min read
Meta Releases Muse Glimmer: Apache 2.0 Licensed 30B Local Agent Model

Meta Superintelligence Labs has released Muse Glimmer, a 30-billion parameter open-weight model distilled from its proprietary Muse Spark flagship. Published under an Apache 2.0 license, Glimmer is purpose-built for offline, on-device agentic workloads like coding, debugging, and file management on consumer hardware.

Ant Group's inclusionAI Team Releases Ling 3.0 Flash FP8 Under MIT License
News Brief2 min read
Ant Group's inclusionAI Team Releases Ling 3.0 Flash FP8 Under MIT License

Ant Group's inclusionAI team has released Ling 3.0 Flash FP8, a highly efficient 124-billion parameter Mixture-of-Experts (MoE) model. Featuring an MIT license and a custom hybrid attention architecture, the model reduces active parameters to 5.1 billion per token, matching the performance of much larger models while dramatically lowering operational costs.

Liquid AI's LFM2.5-2.6B Matches Models Three Times Its Size on Agentic Tasks
News Brief2 min read
Liquid AI's LFM2.5-2.6B Matches Models Three Times Its Size on Agentic Tasks

Liquid AI released LFM2.5-2.6B on August 4, a 2.69B-parameter on-device model purpose-built for agentic workloads. Using a hybrid architecture of short convolution blocks and grouped query attention, it runs under 2.5 GB of memory and reaches approximately 220 tokens/s on Apple M5 Max — while matching or exceeding Qwen3.5-9B on tool use and instruction-following benchmarks.