OVHcloud to Launch First AI Model in 2027 With 110 MW of Available Cloud Capacity

OVHcloud Announces First AI Model Launch With Significant Spare Capacity

OVHcloud, the French cloud computing company, will launch its first AI model in early 2027, CEO Octave Klaba announced following the release of the company's annual earnings report1

1

. The development positions OVHcloud as a potential European challenger in the AI landscape, particularly against established players like Mistral AI. Klaba revealed that approximately 110 MW of the company's 250 MW connected cloud capacity remains available for new clients, giving OVHcloud a distinct advantage as it can onboard customers without immediate infrastructure spending2

2

.

Of the company's total 250 megawatts of connected capacity, 140 MW is currently in operation, leaving substantial room for expansion1

1

. This available cloud capacity means OVHcloud can serve regional demand without the capital expenditure rivals face when building additional data centers. As AI computing demand surges across Europe, this operational flexibility could prove decisive in capturing market share.

Mixture-of-Experts Design Built on Jupiter Supercomputer

The upcoming OVHcloud AI model will employ a mixture-of-experts design, which routes specific tasks to specialized sub-models rather than processing every request through a single large model1

1

. This architectural approach typically offers improved efficiency and performance for diverse workloads. Development work has already begun on Jupiter, which Octave Klaba describes as Europe's fastest supercomputer2

2

. Klaba had previously told Reuters in June that OVH started training its AI models on Jupiter, signaling the company's ambition to compete directly in the European AI landscape1

1

.

Before the full AI model launches, OVHcloud plans to pilot OVHai Workspace, a productivity platform designed for workplace applications such as email1

1

. This workspace tool is tuned on OVH's internal data and represents the company's first step toward delivering AI-powered business applications2

2

. The pilot release is expected shortly, giving OVHcloud an opportunity to refine its AI offerings ahead of the main model launch.

Sharp Criticism of EU-Backed Computing Plans

Octave Klaba reiterated his criticism of EU-backed computing plans, arguing that their financing structure—half from the EU, half from member states—naturally pushes projects toward fragmented national initiatives1

1

. According to Klaba, this logic has produced a race of national data center projects, none supported by a customer base sufficient to justify the investment1

1

. Last month, he stated that national strategies for the European Union's planned AI gigafactories were "doomed to failure"1

1

.

Klaba also questioned whether a proposed €200 million public commitment to purchase AI services over five years would be sufficient to attract the private investment needed for Europe's AI infrastructure ambitions1

1

2

2

. His comments highlight concerns that Europe's scattered approach could leave the continent with fragmented systems and inadequate economies of scale to compete globally. As European firms increasingly turn to AI, the debate over centralized versus national infrastructure strategies will likely intensify, with implications for both public funding priorities and private sector competitiveness in the emerging AI economy.