The global race for the AI ​​application layer continues

The United States is well ahead of Europe in the race for large AI models, but the picture is different at the application layer, with emerging category leaders similar to Lovable, a vibration encoding startup, and Synthesia, which creates AI-generated video for enterprises.

This is the conclusion reached by the global VC company Accel in its report Global landscape 2025 report focusing on the artificial intelligence and cloud market. Worth noting: Accel is an investor in each Lovable and Synthesia.

According to Accel, European and Israeli cloud and artificial intelligence applications have so far contributed 66 cents of every dollar raised by their U.S. counterparts in 2025. “When we started working on this report 10 years ago, Europe was one-tenth of the United States,” Accel partner Philippe Botteri told TechCrunch.

Botteri says this ratio has grown because the region has developed an ecosystem of founders and investors “who really understand how to build great software companies, and this flywheel has been going on for 10 years.”

It’s also a reminder that Europeans and Israelis can do greater than just staff Big Tech AI labs – an statement also shared by Jonathan Userovici, general partner of Paris-based Headline. “In every industry, from law and healthcare to manufacturing and marketing, we see founders who combine world-class technical talent with deep market knowledge,” Userovici told TechCrunch.

This is in line with the Commission’s findings AI Europe 100 report published by Headline earlier this yr, which curated AI-native application startups across Europe that the company says have the “potential to become Europe’s winners of tomorrow” due to their combination of growth speed, team and technological sophistication.

This speed of growth is one of the key differences Accel sees between this wave of AI and previous ones. A brand new breed of native AI apps achieved $100 million in annual recurring revenue inside a few years, a feat that previously took many years.

“They are growing faster than anything we have seen in the past, and they are doing so with incredible efficiency, which means revenue per employee is the highest we have ever seen for software companies,” Botteri said. “And it’s happening on either side [Atlantic] ocean.”

However, he noted that “existing cloud software companies will not disappear.” Accel’s public cloud index is up 25% year-over-year, and these players are “all adding agent capabilities to their products.” As for private firms, some are integrating AI so deeply that they could be considered AI-native, he argued, citing portfolio company Accel Doctolib as an example.

While Europe had high hopes for domestic modeling firms like Mistral AI, Accel’s prospects for European modeling firms are less sunny. But Botteri didn’t completely dismiss it as a space where future leaders would emerge. This can still occur with smaller models, but “it’s not a target-rich environment,” he added.

In turn, VCs are actively competing for investment opportunities in the AI ​​application layer, despite recurring questions about defense. According to Botteri, there is still a defensible approach to build an offering that is product-centric and has rapid adoption.

Another false dichotomy is the belief that there is no space beyond models and applications. “We see that much of the market today is chasing models, calculations and applications, and we believe that data is currently undervalued,” said Lotan Levkowitz, managing partner at Israeli VC firm Grove Ventures. “We strongly believe that companies that focus on proprietary data and data flywheels are very profitable indeed.”