What happened
AI cloud infrastructure provider Boost Run has entered into a $525.6 million contract to supply Cohere with dedicated GPU cloud computing services. According to Data Center Dynamics, the agreement spans a five-year term for each rack provided, with Cohere expected to begin accepting the infrastructure in the second quarter of 2027. The contract includes a provision allowing Cohere to terminate the agreement and receive a refund of any prepaid amounts if Boost Run fails to deliver a minimum amount of infrastructure by July 15, 2027.
Boost Run, an AI cloud provider founded in 2024, has secured a $525.6 million contract with Canadian AI startup Cohere. The agreement involves leasing access to Boost Run's dedicated GPU cloud computing infrastructure.
The contract term is set for approximately five years per rack. Cohere is scheduled to begin accepting the infrastructure in the second quarter of 2027.
A termination clause exists in the agreement: if Boost Run does not provide a minimum amount of infrastructure by July 15, 2027, Cohere is entitled to terminate the deal and receive a refund of any prepaid amounts.
Boost Run currently operates bare-metal GPU servers within top-tier data centers. While the company has previously utilized TierPoint facilities in cities including Seattle, Dallas, and Charlotte, the specific locations for the Cohere deployment have not been disclosed.
Source details: datacenterdynamics.com ↗
Why it matters
This deal highlights the ongoing capital-intensive expansion of AI infrastructure as startups like Cohere secure long-term, dedicated capacity to support their model training and workloads. By locking in five-year commitments for GPU-dense environments, Cohere is mitigating potential supply constraints. For Boost Run, the contract represents a significant scaling of its operations, which currently rely on colocation leases in various U.S. markets. The financial scale of this agreement underscores the high cost of entry for AI-focused cloud providers and the competitive necessity for AI developers to guarantee access to specialized hardware.
The agreement reflects the intense demand for dedicated GPU capacity among AI developers, who are increasingly moving away from generic cloud instances toward specialized, long-term infrastructure commitments to ensure performance and availability.
For Cohere, this deal complements previous infrastructure efforts, such as its $220 million agreement with Buzz HPC and its established use of AMD Instinct GPUs, signaling a multi-vendor strategy to secure resources.
The financial magnitude of the deal—over half a billion dollars—illustrates the significant capital requirements for AI cloud providers to build out and maintain high-performance computing environments.
Boost Run's recent financial performance, characterized by a 576 percent year-on-year increase in colocation lease costs and ongoing operating losses, highlights the high-risk, high-reward nature of the AI infrastructure market as firms race to scale capacity.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Which component of an AI application is the machine-learning model itself?
What to watch next
Observers should monitor whether Boost Run successfully meets its infrastructure deployment milestones by the July 2027 deadline, as the contract includes a termination clause tied to these targets. Additionally, the specific data center locations utilized for this capacity remain undisclosed, leaving questions about the geographic distribution of the resources. Future financial filings from Boost Run will be critical to track, particularly regarding how this large-scale contract impacts the company's operating losses and capital expenditure requirements as it scales its footprint beyond its current TierPoint-based operations.
The primary milestone to watch is the July 15, 2027, deadline for infrastructure delivery, which serves as a critical performance for Boost Run.
Future SEC filings from Boost Run will be essential to determine if the company can successfully manage the operational costs associated with this massive expansion while narrowing its operating losses.
Market analysts will be looking for further disclosures regarding the specific hardware configurations and data center locations, which will provide insight into the technical capabilities Boost Run is offering to support Cohere's AI workloads.