Europe’s Thriving Generative AI Market: A Story of Competition and Innovation

New research and expert consensus indicate Europe's generative AI market is thriving and competitive, driven by diverse business models, genuine user choice, and widespread multi-sourcing of AI models by enterprises to optimize solutions and avoid vendor lock-in. Contrary to concerns of market consolidation, low barriers to entry and flexible switching are fostering competition, prompting calls for EU regulators to maintain a 'hands-off but observant' stance to encourage innovation and scale. This approach aims to capitalize on the sector's significant growth potential, notwithstanding Europe's current deficit in AI infrastructure investment compared to the US.

Analysis

Right before the summer, a panel of leading voices convened to explore the rapidly evolving generative AI market in Europe. And the verdict was clear: innovation is thriving without the need for heavy-handed intervention. Anchored by new research from RBB Economics, the ‘From Lab to Market: Navigating the Competitive Dynamics of Generative AI in Europe’ debate brought together policymakers, economists, and industry to assess the state of play. Their conclusion? The market is working. Diverse business models, genuine user choice, and flourishing partnerships are fueling progress. The smartest role for EU regulators, they argued, is a ‘hands-off but observant’ stance — one that fosters scale and leaves room for global ambition. 1. Competition is working The study’s authors, Katie Curry and Meryem Haraj Touzani, introduced their findings on competitive dynamics in the European gen AI market. They emphasised that “we’re starting to see, for the first time, real activity at the deployment layer of the value chain,” adding that the “transformative properties of this amazing technology actually start to take real-life form in terms of industrial applications.” In terms of competition, the authors identified “a lot of market-led solutions helping to keep barriers to entry down, helping to facilitate uptake, and crucially, to help facilitate switching between models flexibly and at low cost.” While both national and EU competition authorities previously expressed concerns that the gen AI market could “tip towards a situation where there’s just one or a few dominant suppliers,” the RBB study shows “no signs that those concerns are materialising.” Instead, the authors found that firms deploying AI-powered solutions are opting for a multisourcing approach. “They’re not relying on a single provider; they’re using multiple models for different purposes, and they’re able to switch flexibly between those.” Stefano Roncoroni, Case Handler at the European Commission’s DG Competition, expressed his general agreement with the study’s findings. From his enforcement perspective, much of the current focus has been on partnerships between companies operating at different levels of the AI ecosystem. According to Roncoroni, these “partnerships bring a lot of potential pro-competitive benefits.” He explained that they “open up new distribution channels for small developers” and allow them “to reach end users at scale and quite quickly.” Roncoroni emphasised that, from an antitrust perspective, it is important to carefully assess how these partnerships are structured. However, he agreed “that for the moment, we see a lot of promising signs: users are experimenting, they are trying different apps, they are willing to engage with this new technology” At the same time, the Commission will remain vigilant, noting that “we don’t want to break things, but we do want to be able to move fast” if any serious issues emerge in the generative AI sector. 2. The power of multi-sourcing and flexible tools Indeed, the panel provided strong evidence of a healthy, competitive ecosystem. Curry and Touzani noted that firms are actively embracing “multisourcing for models,” a practice that “seems to be fairly common.” Marta Przywala of SAP, a leading German company that develops business applications, confirmed this multi-vendor reality. She explained: “we are one of these companies that pursue a multi-vendor strategy, and we use various models from Europe and from the United States.” This approach is driven by SAP’s strategy “to be relevant, reliable, and responsible” and to ensure that AI applications “really solve business problems.” Przywala stressed that using multiple models “helps to find the sweet spot between accuracy, robustness, and a lack of biases.” “One model can be used for market analytics; another model may be better for customer service or risk mitigation – as well to avoid vendor lock-ins.” Suzy Wild of Anthropic agreed, noting that “we’re seeing a lot of new entrants successfully challenging the established digital firms in this space.” Adding that the flexibility of using different models “is absolutely crucial in preparing the European economy for what’s going to be a huge pace in how technology is applied.” Wild argued that multisourcing not only helps to boost the uptake of AI across the entire EU economy, but is also “equipping a workforce with the right skills.” She also highlighted a crucial issue, pointing out the “huge deficit in [Europe’s] investment compared to the US in all the infrastructure,” suggesting that a supportive environment, rather than premature regulation, is what Europe needs to thrive in generative AI. The RBB authors also shared the example of a major bank, which “actively and dynamically routes tasks between diverse models” to leverage each model’s strengths. This approach is not just about “hedging bets; it’s also about optimisation, because no single model is best at doing everything.” To facilitate this, they highlighted how a number of market-led solutions are emerging “to help manage this flexibility at a lower cost and with less technical difficulty.” Conclusion The panelists had a strong message for EU policymakers: while there is a “clear role here for active monitoring of the sector,” the “balance of risk guards against any early intervention and points more towards a ‘watch-and-see’ approach,” Curry concluded. Indeed, from an industry point of view, I’d like to echo this sentiment, it is absolutely crucial for the EU to resist the old instincts of immediately rushing to regulation when there’s something new or disruptive. The panel discussion made one thing very clear: Europe’s generative AI market is vibrant and competitive at every stage of the value chain, with its diverse business models and low barriers to switching. The deployment layer offers enormous untapped potential, drawing increasing investment thanks to the high value of real-world AI applications. Europe’s thriving gen AI sector presents a clear case for why a ‘do not regulate before you fully understand’ philosophy is the most effective way to foster the EU’s digital future. The continent’s digital ecosystem will only be able to compete, thrive, and lead if the EU supports innovation and finally manages to resist the temptation to rush into regulation. New research from RBB Economics, supported by a panel of industry experts and policymakers, indicates that Europe's generative AI market is demonstrating robust competitive dynamics, contrary to earlier fears of market tipping towards a few dominant suppliers. The core driver of this healthy ecosystem is the enterprise adoption of a "multisourcing" strategy, where firms like SAP actively use multiple AI models from different providers to optimize for specific business problems, enhance accuracy, and avoid vendor lock-in. This trend is fostering low barriers to entry and enabling new entrants like Anthropic to challenge established firms. Consequently, market-led solutions are emerging to facilitate low-cost switching between models. While the European Commission's DG Competition acknowledges the pro-competitive benefits of current partnerships, it remains vigilant. The prevailing consensus advocates for a 'hands-off but observant' regulatory stance to avoid stifling innovation. A key headwind identified, however, is Europe's significant deficit in AI infrastructure investment compared to the US, which could impact the region's long-term scalability and global competitiveness.

Key Decisions for Investors

• Investors should assess companies that enable multi-sourcing strategies, as platforms managing multiple AI models are strategically positioned to benefit from the dominant enterprise trend of avoiding vendor lock-in.
• The current 'hands-off' regulatory sentiment in the EU lowers short-term risk for the sector, but investors should monitor statements from competition authorities like DG Competition for any change in stance, which could rapidly alter the investment landscape.
• Given that enterprises are using a portfolio of AI models, investors in foundational model providers should recognize that market share may be fragmented; differentiation for specific use cases could be more critical for a model's success than all-purpose dominance.
• Consider overweighting exposure to firms at the AI application and deployment layer, which is showing vibrant growth, while being cautious about the European AI infrastructure space until there are clearer signals of investment to close the gap with the US.