DistantNews
Support us
France's OVHcloud plans frontier AI models to become Europe's second LLM player
🇸🇬 Singapore /Technology

France's OVHcloud plans frontier AI models to become Europe's second LLM player

From CNA · () English

Summarized and contextualized by DistantNews.

At a glance

News Named sources Context piece
  • OVHcloud plans to train its own advanced AI models, aiming to become a significant European competitor in the large language model (LLM) market.
  • The company's CEO, Octave Klaba, stated that developing these models is crucial for OVHcloud's future, citing advancements in technology that have reduced development costs.
  • OVHcloud intends to release a family of specialized models and eventually open-source them once they reach a high performance level, differentiating itself from U.S. and Chinese AI systems.

OVHcloud, Europe's largest cloud provider, is embarking on a strategic initiative to develop its own frontier artificial intelligence models, positioning itself as a potential European challenger to established players like Mistral. This move is driven by a growing global demand for alternatives to U.S. and Chinese AI systems, a need amplified by recent disruptions in AI model availability.

It became quite clear to us that if we don't master this technology, we can't guarantee our future.

— Octave KlabaOVHcloud CEO Octave Klaba explaining the strategic importance of developing frontier AI models.

CEO Octave Klaba emphasized the critical importance of mastering this technology for the company's future. He noted that significant advancements in chips, training methodologies, and synthetic data have made the development of cutting-edge AI models more economically feasible. Klaba estimates that a project previously costing around 1 billion euros could now be undertaken for approximately 150 to 200 million euros.

The company views the AI industry as entering a "second wave," where new entrants can build upon the foundational work of pioneers like OpenAI and Anthropic. OVHcloud plans to develop a range of models, recognizing that different models are optimized for specific tasks, rather than a single, all-encompassing system. Crucially, Klaba assured that client data would not be used for training these proprietary models.

We can clearly see that the major players release multiple models, because each model is built for something specific. There's no one model that does all the magic alone.

— Octave KlabaOVHcloud CEO discussing the company's strategy to develop a family of AI models.

OVHcloud has already made progress, including pre-training a model using Jupiter, Europe's fastest supercomputer, via the recently acquired startup DragonLLM. While performance claims are pending, the company's long-term goal is to open-source its models once they achieve sufficient performance levels. This strategy aims to foster innovation and provide European technological sovereignty in the rapidly evolving AI landscape.

We'll see when we're good enough to open source them, but that is indeed the goal.

— Octave KlabaOVHcloud CEO outlining the plan to eventually open-source their AI models.
DistantNews Editorial

Originally published by CNA. Summarized and contextualized by our editorial team with added local perspective. Read our editorial standards.

More Stories

From Our Blog