The race isn’t over for European AI startups | Equity Podcast

European AI Startup Ecosystem Growth and Investment

• The European AI startup ecosystem has made significant progress in the past decade, with the region now producing unicorns and successful IPOs, such as UiPath's $1 billion IPO in 2021, the largest software IPO globally that year (1m49s).
• The European ecosystem has grown proportionally in terms of venture capital investment in software and cloud companies, from 1/10 of the equivalent in the US in 2016 to 1/3 today (2m14s).
• A recent report by Exel, called the Euros Scape report, highlights the impact of AI on the software industry, with AI "eating software" and rewriting the rules of the ecosystem (2m29s).
• The report finds that AI is driving the recovery of the software ecosystem, with AI companies creating significant value in the public markets, including $5 billion of the $8 billion in value created in the past 12 months on the NASDAQ (3m13s).
• The private market is also seeing significant investment in AI, with an estimated $80 billion invested in private cloud companies this year, representing 65% growth versus 2020, with the entire growth coming from investments in AI (3m45s).
• The amount invested in AI has been significant, with around $25-30 billion invested per year over the past couple of years, totaling around $56 billion (4m2s).
• The European ecosystem has come a long way since 2011, when journalists were asking when Europe would generate its first unicorn, and now the region is producing successful startups and attracting significant investment (1m34s).

AI's Impact on the Software Industry and Investment Trends

• The potential of AI technology is highlighted in terms of productivity improvement and redefining software capabilities, with investments in AI startups reaching nearly $4 billion in the third quarter of 2024. However, growth outside of AI is slowing. (4m14s)
• The AI sector is experiencing a frenzy of investments and growth, while traditional software companies face challenges, with their growth rates dropping from 47% in early 2021 to 15% in a recent quarter. These companies are now focusing more on profitability and incorporating AI models to boost growth. (4m42s)
• Factors affecting traditional software companies include the digestion of previous software purchases, consolidation of software tools, and a shift in IT budgets towards AI investments, which are growing at a rate of 30%, while other areas are decreasing. Additionally, macroeconomic uncertainties are slowing sales processes. (5m45s)

European AI Talent and Investment Landscape

• Europe is known for its enterprise software, but the US currently has an advantage in AI investments, with 80% of AI funding going to the US and only 20% to Europe. Despite this, Europe has a strong talent pool in AI, with significant contributions from France, the UK, and Germany. (6m41s)
• European AI talent is exemplified by institutions like Meta's AI lab in Paris, DeepMind in the UK, and the development of Stable Diffusion in Munich. Although Europe started later in foundational AI models, companies like Mistral, H, and Poolside 11 Labs are emerging in the field. (7m11s)
• European AI startups have raised hundreds of millions, but not billions, and the next step for them would be to invest in foundational models to not rely on US models (8m6s).
• There is room for these companies to focus on more specialized models that can provide value, particularly in Enterprise applications where inference cost and latency are crucial (8m31s).
• Some European AI startups may end up raising a lot of capital and changing leagues to compete with companies like Entropic and OpenI, which are currently raising billions (9m4s).
• The question remains whether the best models will require tens of billions of investment in infrastructure, making it challenging for smaller companies to catch up (9m32s).

Infrastructure and Competition in AI

• Owning infrastructure seems to be very important, and companies like Microsoft are investing in energy production, such as nuclear power, to secure their energy needs (10m5s).
• Startups, or "challengers," may struggle to compete with larger companies that own infrastructure, data centers, and energy production, and may need to focus on smaller use cases (10m43s).
• Companies like Alphafold, which raised over $500 million, have pivoted from building their own foundational models to AI support and developing advanced generative AI models, due to insufficient funding (10m53s).
• Some companies, like Mistral AI, have enough capital to build their foundational models, but others may need to adapt their strategies to remain competitive (11m35s).

Funding and Future of European AI Startups

• The European AI startup race is not over, and it's too early to call it, as the industry is still in its early stages (11m42s).
• European startups can access funding from various sources, including private funds, strategic investors like Amazon, Google, and Microsoft, and sovereign funds from the Middle East (12m2s).
• The key to securing funding is to demonstrate progress and show that there's a big opportunity for the company, making it attractive to investors (13m1s).
• To promote portfolio companies on the international stage, venture capitalists focus on the team and product, which can speak for themselves and generate interest from investors (13m34s).

Talent Acquisition and Acquisition of European Startups

• European startups have an advantage in terms of talent, but big tech companies have consolidated AI talent by acquiring or hiring top researchers and engineers (14m39s).
• European startups are vulnerable to being acquired by larger tech giants, with examples including the founder of Character AI returning to Google and Microsoft's acquisition of Inflection AI (15m7s).
• The agentic workflow revolution is a promising area, with companies like H, Microsoft, and Tropic working on it, but it's still in the early stages and tangible results are yet to be seen (14m6s).
• European AI startups that have successfully raised funds are unlikely to return to their previous employers, such as DeepMind and Meta, as they have chosen to leave these big companies to achieve their goals in a different environment with more freedom to develop and ship products (15m27s).
• These startups are on a mission to succeed, and while some may go back to their previous employers if things don't work out, this is not the current pattern being observed (16m32s).

M&A Activity and Exit Strategies for European Startups

• Some European AI startups will likely be successful and remain standalone, independent companies, while others will be acquired due to regulatory constraints (16m49s).
• The state of the exit market for European startups is expected to see increased M&A activity, with global M&A deals expected to reach around $75.4 billion in 2024, driven by large exits (17m42s).
• The $75 billion M&A figure is a global number, not specific to Europe, and represents deals worth over half a billion dollars (17m59s).
• A significant portion of the expected M&A activity in 2024 is attributed to a single deal, the acquisition of Unisys by Synopsis, worth $35 billion (18m22s).
• Europe has seen fewer large M&A deals recently, but there is a large pool of late-stage software companies with revenues in the $200-400 million range that are potential targets for acquisition (18m51s).
• Many of these European companies missed the IPO wave of 2012 and are now looking for alternative exit options (19m4s).
• European AI startups are expected to see exits through IPOs or strategic M&A in the next two to three years, once the IPO market reopens (19m8s).
• Big tech companies are missing in action in the M&A market due to regulatory scrutiny, making it difficult for them to make large acquisitions without huge uncertainty (19m25s).
• This regulatory scrutiny creates a high-risk environment for both the acquiring and acquired companies, impacting the tech M&A market (20m2s).

Alternative Investment Strategies and Late-Stage Funding

• Instead of traditional M&A, some big tech companies are opting for "reverse acqui-hires," where they hire key talent from AI startups, as seen in deals involving Character, Inflection, and Adep (20m13s).
• Late-stage funding in Europe is challenging, but the region is attracting capital, with new funds emerging, such as 20 VC's $400 million fund and Node VC's 71 million euro fund (21m30s).
• The influx of capital into European funds is due to the recognition of Europe's strength in the global venture ecosystem, and the potential for great companies to emerge from the region (21m52s).
• The notion that a specific country is missing late-stage capital is outdated, as capital has no boundaries, and funds invest across regions; instead, the issue is that late-stage capital is selective and only invests in companies with strong growth rates and global ambitions (22m20s).
• Europe faces similar challenges as far-away islands like New Zealand in terms of accessing capital, but the key factor in securing funding is the potential return on investment, not the location of the company (23m12s).
• If a company has the potential to create billions of dollars in value by addressing a global market, it can secure funding regardless of its location (23m36s).
• The capacity of the venture market to invest $100 million in a company is not the primary concern, but rather the company's potential, scope, and ability to go global or address a large market (23m50s).

The Agentic Revolution and Future of Software Development

• Next year is expected to be the year of the "thetic revolution," with Europe playing a significant role, particularly with top research teams like DeepMind working on this problem (24m10s).
• Agent technology is expected to advance to the point where it can interact with web applications and execute digital tasks for people and companies (24m30s).
• A new way of developing code is expected to emerge, with models becoming more capable of generating code, leading to a shift from code completion to application creation (24m50s).
• Developers will focus on defining software specifications, and AI will write and maintain the applications, as envisioned by companies like Tessel in London (25m10s).

Conclusion and Acknowledgements

• Listeners can connect with the guest online through LinkedIn and X, and with the host, Rebecca Balon, on X and Threads (25m44s).
• The team and Henry Pickrill, who manages TechCrunch audio products, are acknowledged for their efforts.
• The audience is thanked for listening to the content.
• The conversation will be continued in a future episode.