Clichmont has launched a customer-facing compute platform that opens its GPU infrastructure to Web3 participants, making rentals available through app.clichmont.ai and bringing the CLAI utility token into active service.

The company developed its physical compute infrastructure before releasing a customer interface. The new portal isolates active compute onboarding from the rest of Clichmont's digital footprint, separating the transactional layer from its corporate operations at clichmont.com and its general project documentation at clichmont.ai.

The infrastructure behind the platform

Behind the interface, Clichmont reports 12 GPUs at its Billingstad facility: eight Nvidia H20 units and four Nvidia RTX PRO 6000 units. Its existing Bodø Harbor deployment comprises 20 GPUs — 10 RTX 4090 and 10 RTX 5090 units — with a further 40 RTX 5090 GPUs in the pipeline. The fleet spans data-center, professional, and consumer GPU classes, giving the platform a spread of hardware tiers rather than a single category of accelerator.

Beyond its current GPU deployments, the company plans to build additional data centers. Alicante is in development, with construction scheduled to begin on May 1, 2027. A separate, purpose-built Bodø data is targeted to become operational in Q4 2027.

Power availability and deployment timelines guide Clichmont's choice of sites. The company ranks available power and time to power delivery ahead of cooling conditions and fiber connectivity, with land cost following those requirements. Each factor affects when capacity can enter service and what it costs to operate over time. That ordering mirrors a constraint familiar across the data center industry, where securing power is often the long-lead step in bringing new compute online.

Alexis Cathalifaud, Clichmont's CEO, said the company chose to own its infrastructure in order to control the entire compute stack, managing both the physical environment and the hardware inside it. “Renting capacity means depending on someone else's availability, costs, and priorities.”

Where CLAI enters the platform

Compute on the platform is priced and paid for in stablecoin, meaning CLAI's market price does not directly determine customers' rental bills. The compute business depends on customers finding enough value in its available capacity to pay for it, which makes hardware availability, service delivery, and operating costs central to its commercial viability regardless of how much attention the token attracts.

CLAI serves as the access and utility layer for the compute infrastructure. Its functions include staking for access benefits and discounts, alongside eligibility tied to a buyback mechanism.

Clichmont has also identified the operating measure it wants to prioritize. “The primary metric we want to be judged on is utilization of deployed compute capacity,” the company said. Contracted capacity and the cost of delivering compute provide further measures of whether the hardware is turning into a sustainable service.

The launch serves as a test for a broader question in the AI and crypto space: can a utility token attached to real infrastructure demonstrate real function on its own? What is worth watching next is onboarding activity at app.clichmont.ai, and how the CLAI layer holds up once customers are actively using the platform. The expansion calendar offers further checkpoints: the additional 40 RTX 5090 units in the pipeline, the scheduled start of construction in Alicante on May 1, 2027, and the targeted Q4 2027 operational date for the purpose-built Bodø data center.

This article was first published by Metaverse Post.