Over the past two decades, few technologies have had such a transformative and cross-sector impact as artificial intelligence (AI). While its origins go back much further, the true inflection point arrived with the rise of deep learning, the exponential growth of computing power, and, more recently, the emergence of generative AI. For venture capital funds, this represents a real-time revolution where opportunities flourish but demand sharp analytical skills and precise market timing.
The numbers speak for themselves: according to data from The Economist, global AI investment reached USD 200 billion in 2024—twice as much as in 2023. This includes both venture capital and investment in infrastructure, data centers, and the development of specialized chips for model inference and training. This investment frenzy has been led by players such as Nvidia, Amazon, Microsoft and Alphabet, which are not only betting on their own models but also building the hardware and infrastructure needed to scale AI.
In Europe, the landscape is not far behind. According to an IDC study cited by Europa Press, corporate spending on artificial intelligence in the EMEA region grew by 61% in 2024 compared to the previous year. This growth has been driven by public incentives, adoption in traditional sectors such as banking and industry, and a new generation of startups focused on vertical AI solutions.
During the first wave of enthusiasm around models like ChatGPT, Bard or Claude, many funds explored investing directly in startups building large language models (LLMs). However, the high technical barriers to entry and the dominance of a handful of players (OpenAI, Google, Anthropic, Mistral, Cohere) have shifted attention toward practical applications and smaller models that can be adapted to local contexts or specific business needs.
Many VCs have taken note that the winners won’t necessarily be those who build the most powerful models, but those who manage to integrate them effectively into real-world workflows.
The biggest promise lies in vertical solutions. Startups applying generative or predictive AI to sectors such as legaltech, insurtech, healthcare, logistics, energy or agrifood are increasingly attractive to investors. These companies aren’t competing to build the next ChatGPT—they’re solving specific problems with real competitive advantages: proprietary data, sectoral know-how, integration with legacy systems, etc.
Examples include:
Much like in the gold rush, the winners are often those selling the shovels rather than those digging for gold. Many funds have understood this and are investing in the infrastructure needed to scale AI: chips, storage, distribution networks, and middleware platforms for orchestrating models in production.
Companies such as Modular, Weights & Biases, Hugging Face and Lambda Labs are capitalizing on this need for infrastructure that does not depend on a specific use case but on the broader evolution of the ecosystem.
One of today’s main bottlenecks is effective AI implementation in traditional companies. New startups are emerging that provide low-code platforms for building AI-powered workflows, data governance tools, bias evaluation, explainability, and regulatory compliance (AI Governance). This will be a critical segment in Europe, especially with the entry into force of the EU AI Act.
While classic metrics such as TAM, traction and team remain essential, VCs have added new AI-specific criteria:
Contrary to appearances, the future of AI is not only “bigger.” A growing trend toward small models (like Microsoft’s Phi-3 or Google’s Gemma) is emerging—models that can run on edge devices or be finetuned with fewer data. This will unlock applications in sectors where privacy, latency or connectivity are critical factors.
The combination of text, image, video, audio and structured data in the same system (such as GPT-4V or Gemini 1.5) will transform sectors like customer service, product design, medicine and entertainment. Investors are watching for startups capable of converting this multimodality into tangible competitive advantages.
We are no longer talking about AI that simply responds, but AI that acts: systems able to make decisions, execute actions on external platforms, interact with APIs, and continuously optimize processes. Although still experimental, AI agents will be key in areas such as operations, automated marketing or logistics.
With increasing regulatory pressure, startups that integrate strong safety, ethics and explainability practices from the outset will be better positioned to scale. Responsible AI won’t be optional—it will be a commercial and legal requirement, especially for gaining corporate or institutional clients.
Artificial intelligence is setting the pace of a new industrial revolution. For venture capital funds, the key is not only to anticipate which technology will dominate, but to understand how and where it will be implemented efficiently and ethically. In this new era, differentiation will not come only from having the most powerful model but from the ability to integrate it with purpose, solve real problems and scale sustainably.
In short, the next generation of AI unicorns will not necessarily be the flashiest—just the most useful.
Laurent Arens
Director
BStartup10
cloud technology axon
Xcalibur Smart Mapping, recognised as the global leader in airborne...
