DeepSeek's Shockwave: A New Dawn for European AI?

Over the past 2 weeks the AI landscape has been dramatically reshaped. The entry of Chinese AI company DeepSeek has effectively democratised access to powerful AI models. By offering their models for free using an open-source approach, DeepSeek has levelled the playing field for European startups.

This move has shifted the global AI investment focus away from foundational models - where US companies have so far been dominant - to AI applications where Europe can excel. This sudden change in the emerging AI value chain has been likened to a "Sputnik moment", highlighting the need for an urgent recalibration of where value will now accrue. 

DeepSeek's sudden entry not only raises questions about the U.S. competitive edge in the AI race but potentially enables Europe to develop its own competitive edge faster. Investors are eagerly trying to assess how this new paradigm will play out.

DeepSeek: A Game Changer

DeepSeek has disrupted the AI industry by leveraging an innovative approach to AI development. Its models, like the R1 reasoning model, match the capabilities of leading US models like GPT-4 but at a significantly lower cost and, apparently, with less reliance on expensive hardware. As Tomasz Tunguz, general partner at Theory Ventures, optimistically states, "Startup margins will surge. As AI performance per dollar skyrockets, startup economics will fundamentally improve."

This lower cost approach has been achieved through a combination of efficient algorithms, optimised training methods, and a focus on readily available hardware. As a result, DeepSeek's chatbot application has rapidly gained popularity, even surpassing ChatGPT in download numbers in early 2025.

DeepSeek's R1 reasoning model, which thinks through problems step-by-step before answering, matches the capabilities of OpenAI's o1 model while offering it at a significantly lower cost. This suggests that DeepSeek may be leading in reasoning efficiency, although direct comparisons remain difficult due to the lack of public pricing information from other companies like Google.

Beyond cost efficiency, DeepSeek's technology is also more energy-efficient than traditional AI models. This could offer a path towards more sustainable AI scaling, reducing the environmental impact of AI development as demand for the technology increases.

DeepSeek's open-source approach allows anyone to download, modify, and deploy its models, further lowering the barriers to entry for AI development. This open-source approach has the potential to drive advancements in AI technology more broadly, fostering a collaborative environment where developers worldwide can contribute to AI innovation.

CB Insights notes that this open-source trend is particularly prevalent among smaller AI models, where it "dominates". Pitchbook echoes this sentiment, highlighting that DeepSeek's open-source nature allows for "greater data customisation for specific use cases", particularly beneficial for vertical AI startups.

Where Europe Can Shine

Europe has several advantages that position it well in this new era of AI applications:

• Industry Expertise: Europe has a diverse industrial base, with strengths in sectors like manufacturing, healthcare, and energy. This deep industry knowledge can be leveraged to develop AI solutions tailored to specific needs and challenges. Pitchbook emphasises this, stating that "with a large industrial base, European startups focusing on integrating AI have a competitive edge in areas such as defence, manufacturing and supply chain optimisation".
• Strong Data Protection Frameworks: Europe has a long-standing commitment to data privacy, with regulations like GDPR setting a global standard. This focus on ethical and responsible AI development could be a key differentiator for European startups, especially as concerns grow about DeepSeek's data security practices and potential ties to the Chinese government.
• Talent Pool: Europe boasts a highly skilled workforce, with a strong emphasis on research and development. This talent pool is essential for building innovative AI applications that address real-world problems. This was highlighted in our recent piece: Are Europe’s AI Ambitions Falling Short?

What VCs Will Be Looking For

With the rise of AI applications, VCs are increasingly seeking startups that can demonstrate:

• Real-World Value: The focus is shifting away from "AI for AI's sake" towards solutions that address tangible business challenges and deliver measurable impact. This sentiment is echoed by investors interviewed by Snowflake, who emphasise the need for startups with "real value propositions" that use AI to solve fundamental business challenges.
• Strong Go-to-Market Strategies: Startups need to demonstrate a clear understanding of their target market, a compelling value proposition, and a robust plan for customer acquisition and retention. This includes identifying early adopter segments, building strong sales and marketing channels, and demonstrating a clear path to profitability.
• Data Security and Ethical Considerations: In light of concerns surrounding DeepSeek, VCs will favour startups that prioritise data privacy and adhere to ethical AI principles. This includes demonstrating compliance with regulations like GDPR and the EU AI Act, ensuring transparency in data usage, and implementing robust security measures to protect user data.
• Competitive Differentiation: With the democratisation of AI, VCs are acutely aware of the increasing competition in the AI application space. They will be looking for startups that have a clear competitive edge, whether it's through unique technology, proprietary data, strong partnerships, or a deep understanding of a specific niche. This differentiation is crucial to avoid being overshadowed by larger players like OpenAI (who are quickly becoming much more than a foundation model company) or a wave of similar startups.

Key Application Areas

European startups are well-positioned to capitalise on the growing demand for AI applications that deliver real-world value and address specific challenges within European markets. By focusing on areas where they have a competitive edge, such as data privacy, sustainability, and industry-specific solutions, European startups can attract investment and establish themselves as leaders in the global AI landscape. Some key application areas include:

• Agentic AI: This rapidly evolving field focuses on AI systems that can make autonomous decisions and execute tasks without constant human input. European startups are already making strides in this area, with companies like Maki People (France) and Spleenlab (Germany) developing innovative solutions for HR and autonomous mobility. Pitchbook also highlights the potential of AI agents in customer relationship management and enterprise resource planning.
• AI for Industry: Europe's strong industrial base provides fertile ground for AI applications in manufacturing, energy, and logistics. Startups like DeepOpinion (Austria) are developing AI agents to automate complex business processes, while others are focusing on robotics, quality control, and industrial process optimisation.
• AI for Healthcare: Europe has a long history of innovation in healthcare, and AI is poised to play a key role in the future of medicine. Startups like Huma are developing AI-powered solutions for drug discovery, diagnostics, and personalised medicine.
• AI for Sustainability: As the world grapples with climate change, AI is emerging as a powerful tool for developing sustainable solutions. European startups like Sinay, Overstory, and Mitiga are leveraging AI to optimise energy consumption, reduce waste, and develop innovative clean technologies.
• AI for SpaceTech: The burgeoning space economy presents exciting opportunities for European startups. With the cost of space launches decreasing, AI is being used to enhance Earth observation, space manufacturing, and inter-satellite communications. European startups like Isar Aerospace, ClearSpace and Iceye, are developing innovative solutions in this sector.

The Road Ahead

While the opportunities for European AI startups are vast, the landscape is also becoming increasingly competitive. The democratisation of AI access brought about by DeepSeek means that more players are entering the field, creating a more complex environment for both startups and investors to navigate. This increased competition could make it "difficult for startups to build a defensible business," as Elizabeth Yin, general partner at Hustle Fund, points out in her latest blog.

In addition, concerns about DeepSeek's security and potential for data misuse cannot be ignored. As Rasmus Rothe, General Partner at Merantix Capital, cautions, "There's still a risk that the model somehow gives an answer that is clearly pointing back to China". Pitchbook adds that "An evaluation by AI compliance company LatticeFlow AI found that the model has significant cyber attack vulnerabilities." This highlights the need for European startups to prioritise data security and ethical AI development, not only to comply with regulations like the EU AI Act but also to build trust with customers and investors.

Despite these challenges, the future of European AI is bright. With a strong foundation in research, a commitment to ethical development, and a growing ecosystem of startups and investors, Europe has the potential to become a global leader in AI innovation, especially at the application layer. The key will be to seize this moment, address the challenges head-on, and build a thriving AI ecosystem that benefits both businesses and society as a whole.

Let's talk.