Canopy managed the provider evaluation, performance testing and contract negotiation, securing GPU capacity at 35% below market after pricing conditions had tightened.

"When we first approached Canopy we weren’t sure whether we wanted spot capacity, a lease, or to buy outright. They ran a full evaluation in three weeks and priced all three options for us. The project then paused on our side, and by the time we restarted the market had moved against us, with rates rising and suppliers pushing for larger three-year commitments. 

Canopy picked it straight back up, held firm on the rate, and secured us capacity around 35% below the market at the time. They handled the contracts and the behind-the-scenes work so the whole process was seamless, and we’re already working with them on our next, larger round." - Lindley Lentati, Ph.D, Co-Founder, Cambridge Inference

About the Client

Cambridge Inference is a UK-based AI and machine learning company preparing to scale its GPU infrastructure during 2026.

Sector: AI and machine learning

Region: United Kingdom

Engagement: GPU sourcing, evaluation and contract negotiation 

Operational Summary: 

Cambridge Inference engaged Canopy to source, evaluate and negotiate GPU infrastructure for its 2026 growth plans. Canopy compared spot, lease and purchase options, coordinated provider testing and modelled three-year costs before securing GPU capacity with Civo, approximately 35% below the prevailing market rate.

Challenge:

Cambridge Inference needed GPU capacity quickly, but did not want to commit to the wrong supplier, commercial structure or contract length.

The team needed to compare several procurement models, validate technical performance and understand the long-term financial impact before signing.

As Lindley Lentati explained, the market had moved against the company during the pause, with rates rising and suppliers applying pressure for larger commitments. This meant the original commercial assumptions could no longer be taken for granted.

Solution:

Canopy returned to the market and managed the process through to contract completion.

The team coordinated supplier engagement, commercial terms, scope, testing, onboarding and contracting. Canopy also pushed back against rising prices and larger commitment requirements to keep the agreement aligned with Cambridge Inference’s actual needs. The final agreement secured GPU capacity approximately 35% below the prevailing market rate at the time. 

Results: 

Canopy gave Cambridge Inference a clearer route to GPU capacity, combining competitive pricing with a fast, managed procurement process.

• Negotiated Rate Advantage: Approximately 35% below the prevailing market rate

• Speed to Capacity: Three weeks from evaluation to completed agreement

Ready to reduce your GPU costs? Visit Canopy Compute to compare options and start securing better rates.