Every now and again, something momentous happens that changes the world and it’s never the same again. As Lenin once said, “There are decades where nothing happens; and there are weeks where decades happen”.
Last week was such a week. The huge flappings of the biggest black swan reverberated around the tech world when China’s DeepSeek released its R1 model. What was this madness?
China was supposed to be lagging behind the US in the AI race and, indeed, as Marc Andreessen said, it was a Sputnik moment, referring to when the Russians beat the Americans in the first Space Race.
Naturally, stock markets and crypto went into meltdown. Nobody knew what was happening, chip companies such as Nvidia lost hundreds of billions and new-President Trump’s announcement of its $500 billion Stargate initiative was rendered as obsolete as Open AI’s business model.
And then everybody calmed down. Stock markets recovered and people put their heads back on the right way. However, it was a pivotal moment for tech, a small step for mankind and a giant leap for AI. Aside from tenuous Space analogies, what did it really mean?
Well, what it means is that it is obviously bad news for people selling AI, it’s very good news for those buying it. Apple’s price went up after DeepSeek’s release. That’s because when there are losers, there are always winners.
And so it may be for the state of European AI, it may be very good news indeed. Ten days ago, it seemed Europe had missed the boat and would never become an AI player. Now with prices slashed and the apparent lack of need for huge data centres and unattainable chips, Europe may have a once-in-a-lifetime opportunity to win the AI race.
So, what do AI companies across Europe think of DeepSeek's recent influence and does it offer an opportunity for Europe to catch up and overtake the US through innovation, and not funding? Perhaps it's an unique opportunity for them, so we asked some of them.
Some agree wholeheartedly. Elena Poughlia is the founder of Dataconomy and is working from Berlin with a 150-person, hand-picked contributors of AI mavens, developers and entrepreneurs to create an AI Ethics framework for release in March.
For European startups who have not built on ChatGPT, Perplexity and Claude models it’s great. It's cheaper, more sustainable because it needs less hardware and computing power… and an open source solution that changes the rules of the game.
The bubble was going to burst anyway and let’s see how that now pops. For existing startups, it could be the moment that turns off the lights or makes them throw their work to the bin. Now the bigger broader question is what will happen with our data and how will it be used - and how will this play out in the larger geopolitical game.
There are many precedents in the tech world where second movers have ‘piggy-backed’ on the shoulders of the tech giants who came before them.
The case of M-Pesa may be an African story, not a European one, but its release of a mobile money app ‘for the unbanked’ in Kenya almost 18 years ago created a platform that led the way for European FinTechs and banks to compare themselves to… and imitate.
In DeepSeek’s case, European AI startups will not ‘piggyback’, but rather use its release to springboard their businesses. To supercharge their businesses… that is if they get it right.
Alex Ball is the investment director at London-based Web3 and AI "Venture Builder" Block Dojo and he suggests that AI startups need to think about the end user, not the provider.
AI startups have been chasing the wrong trophy. DeepSeek just dropped, and for a moment, it looks like the game has changed. But by next month? There’ll be another. And another. The cycle never stops.
Meanwhile, a different kind of AI company has been playing a longer game—one that isn’t about who has the best model, but who owns the relationship with the user. ChatGPT didn’t dominate because it had the most advanced tech. It won because it became second nature.
Carol Constant is the founder and CEO of an AI HR company WhomLab and points out both geopolitical and regulatory risks for European AI companies that embrace DeepSeek.
There are red flags to be aware of in terms of geopolitical and regulatory risks. DeepSeek's reliance on Nvidia H800 chips, subject to US export controls, raises concerns about long-term access, especially under Trump’s presidency.
"Moreover, our EU Al Act may impose stricter compliance requirements, making it challenging for startups to use models trained in jurisdictions with looser regulations such as China.
Others such as Appu Shaji, the CEO and Co-Founder of Mobius Labs, which ‘delivers high accuracy with 10x costs across text, audio, and video for enterprise AI’, points out the funding gap between US and European AI companies.
European AI has always had the talent and the research breakthroughs, but historically lost out in commercialisation because of funding barriers. Now, with open-source models matching or even surpassing closed systems, the game shifts from billion-dollar war chests to execution speed, fine-tuning prowess, and, crucially, innovation.
“We’re seeing teams across Europe, replicate and refine cutting-edge models at a fraction of the traditional expense. The next chapter of AI won’t just be about money; it will be driven by original thought, creativity, and the ability to move fast - something Europe is structurally weaker in and needs to improve.
Others point to the budgetary situation when it comes to potential European AI innovation. Anders Ibsen is CEO at Copenhagen-based Savery, an AI platform that autonomously writes and tests codes for its customers.
With restrained budgets, Europe’s AI opportunity isn’t in more chips, but in better algorithms. True innovation comes from First Principles Thinking; rethinking fundamentals rather than scaling inefficiencies.
DeepSeek R1 proves that efficiency beats brute force, and those who master algorithmic breakthroughs will lead the next wave of AI, fundamentally disrupting entire industries.
Of course, DeepSeek will not be the only Chinese company creating springboards for European companies. There is likely to be a tranche of announcements and releases coming out of China that will springboard on themselves.
As much as OpenAI was only the first AI earthquake, there will be more tremors and others even greater on the Richter scale. In the end, it will all be based on investment and eventual exits.
Daniel Hulme is the co-founder of Conscium, the ‘world’s first applied AI consciousness research organisation’, who previously sold his AI enterprise business Satalia to WPP for $100m in 2021, one of the biggest UK AI sales after DeepMind sold out to Google. He sees agentic AI as an excellent opportunity for European AI startups.
The DeepSeek saga shows that AI is still at an early stage and very fluid. The competition is now moving towards agentic AI and the field is wide open for innovative players from Europe to stake their claim here.
Furthermore, consciousness in machines remains an under-represented element of AI development, and 2025 could be the year we see this change.
Anything that comes out of China, however, will come with caveats. DeepSeek’s R1 comes with censorship. Questions or prompts about Taiwan or Tiananmen Square bring up the usual Great Wall and this may be something that the European AI ecosystem may decide is a red line they don’t want to cross.
Barcelona-based Nikita Kaeshko is the CEO of Overwatch AI, a company that ‘develops secure, offline AI for compliance-heavy industries that enables transportation professionals to access critical intelligence without cloud dependence’.
"Despite criticism about Chinese origins, I believe this misses the core point. Open-source models provide far better transparency and data control than closed commercial ones, making them ideal for EU use under strict privacy regulations. For EU startups, open-source is the fastest path to building compliant AI applications," he says.
Meanwhile, moving from Barcelona and across to Ghent in Belgium, Xander Berkein is the Co-founder of Donna, an AI assistant for ‘sales reps on the go’:
The launch of Deepseek reinforces the growing power of open-source AI. For us, having the ability to self-host high-performing models at low cost is a game changer. This shift enables us to build more customisable and secure voice-powered AI solutions tailored to the unique needs of the sales teams of our enterprise customers.
And so say all of us, albeit in somewhat vanilla terms.
The opportunity here for AI companies in Europe, whether they are from Denmark, Germany, Belgium, Spain or the UK is clear to see. They have been handed a grenade from DeepSeek that may go off in their face or it may completely explode the global AI ecosystem in their favor.
Let’s see where it all goes off.
Would you like to write the first comment?
Login to post comments