China-based DeepSeek has disrupted the artificial intelligence sector by claiming it developed a high-performing AI model at a fraction of the cost seen with U.S. tech giants. Experts suggest this breakthrough highlights untapped potential for Europe, long considered lagging in the AI race.
DeepSeek’s large language model (LLM) “makes a mockery of the (idea that) we need a trillion dollars to train the next level of AGI (artificial general intelligence),” said Neil Lawrence, a machine learning professor at the University of Cambridge.
Lawrence referenced the recent unveiling of the U.S. “Stargate” project, a $500 billion initiative led by OpenAI—the creator of ChatGPT—that aims to bolster AI infrastructure. The project’s massive funding, largely earmarked for data centers equipped with state-of-the-art AI chips, highlights how few European firms can muster such resources.
Yet DeepSeek’s reported success in producing an AI model comparable to OpenAI’s for just $5.6 million challenges these assumptions.
The technology demonstrates the promise of creating “models that are more efficient and less hungry for GPUs (graphic processing units), for energy, and for cash,” said Laurent Daudet, CEO of French generative AI firm LightOn. “It’s interesting for Europe to see that we don’t need a Stargate project to do something innovative… you don’t need $500 billion,” he added.
Impact on Markets and AI Development
DeepSeek’s innovation has already rattled tech markets, with AI-linked stocks, including key chipmaker Nvidia, taking hits in recent days.
“It shows that competition is very, very strong and that there’ll be a price war too,” said Nicolas Gaudemet, AI lead at consultancy Onepoint. “An additional provider will bring prices down, accelerating the integration of generative AI within companies.”
A Wake-Up Call for Europe
For Lawrence, DeepSeek’s achievement underscores a missed opportunity for Europe, which boasts rich AI talent in both academia and industry.
DeepSeek’s approach, he said, reveals “a small change in the recipe” that hints at further innovation. “It is very encouraging for Europe and reflective of what we should expect going forward… there will be more than one DeepSeek,” Lawrence remarked.
He praised DeepSeek’s use of open-source methodology, where its designs are made accessible to the global AI community, enabling other organizations to refine and expand upon them.
This strategy presents opportunities for European AI players like France’s Mistral, said Gaudemet. “They can reuse (DeepSeek’s models) to train their own and stay in the race,” though he cautioned that “competition isn’t just between the U.S. and Europe—China is showing that it’s capable.”
The Future of Cheaper, Decentralized AI
Cheaper and more efficient AI technologies could pave the way for AI solutions tailored to local markets and smaller businesses, rather than centralized, resource-heavy models.
“There is a future for more frugal models that perform just as well, particularly for business needs,” said Daudet of LightOn. He likened his company’s role to “building the ‘chassis’ of a usable vehicle around the ‘motor’ of an AI model” and ensuring secure, customized solutions for clients.
Gaudemet highlighted security and data sovereignty as key advantages for European companies, particularly as businesses grow concerned about where and how their data is managed. “There’s obviously a card to be played for (European) companies, which is security: ensuring your data will stay in Europe,” he said.
Lawrence emphasized the importance of leveraging Europe’s strong AI research base and prioritizing locally-driven innovation. “We don’t need massive amounts of investment, but we do need to focus on strengths within our own continent—without being distracted by the latest narrative from figures like (OpenAI chief) Sam Altman.”
Europe’s potential in AI remains formidable, he added, though realizing it requires more attention to homegrown opportunities and less fixation on global tech giants.
