The AI Funding Paradox: Why "AI-Native" Startups Are Actually Raising Less Money
 
Why are companies calling themselves "AI startups" actually raising smaller rounds at lower valuations than traditional software businesses?
 
Everyone knows AI is transforming venture capital. OpenAI commands a $480 billion valuation. Anthropic just raised $13 billion. The AI gold rush is real - but only for a select few building foundation models and core infrastructure.
 
For the thousands of European startups positioning themselves as "AI-first" companies, the reality is starkly different.
 
It’s hard to separate out from all the noise, but the AI market has split in two. At the top, a handful of model builders and infrastructure players command astronomical valuations. But beneath them, the vast majority of "AI-native" startups - those building applications and wrapping APIs - are discovering a harsh reality. They're actually raising less money at lower valuations than traditional SaaS companies at the same stage.
 
An analysis of Pitchbook European data from 1H 2025 reveals the paradox in stark numbers: In aggregate across all stages, AI median valuations have trumped SaaS since 2018. Now we have reached parity at €8.7 million. But an even more interesting story emerges when you look by stage.
 
At Pre-seed and Seed, AI still commands a 23% premium (€5.9 million versus €4.8 million). But by Series A and B, SaaS companies achieve median valuations of €45.3 million - a stunning 42% premium over AI's €31.9 million.
 
The valuation flip happens at precisely the moment investors demand real business fundamentals over promises. And the gap is widening.
 
 
The Numbers Tell an Uncomfortable Truth
 
The European venture market has delivered its verdict on AI startups, and it's not what founders expected. Finrofca's Q4 2025 analysis of 565 AI companies reveals a market structure that explains why most AI startups face this funding paradox.
 
The taxonomy matters. ‘Foundational’ level vendors and ‘Infrastructure’ companies are valued for their position in the stack and their ability to shape the direction of the field, while ‘Applied’ categories are priced more like SaaS, with emphasis on sales cycles and customer stickiness. This distinction is crucial - the vast majority of European AI startups sit in the applied layer, not the infrastructure or foundational layers that command premium valuations.
 
Applied categories including PropTech, Legal Tech, HR Tech, and Productivity Tools now trade closer to enterprise SaaS benchmarks than frontier AI layers. This isn't a temporary correction - it's the market recognising that most "AI startups" are simply traditional businesses with AI features, not revolutionary platforms.
 
The PitchBook data confirms what investors already know: when it comes to Series A and beyond, they're not paying for AI promises anymore. They're paying for business fundamentals. And on that metric, traditional SaaS companies are winning decisively.
 
 
The "Feature Not a Company" Problem

The collapse of AI wrapper startups has become a cautionary tale that every founder should study. When Jasper AI raised $150 million at a $1.5 billion valuation in October 2022, it seemed to validate the AI-first approach. By 2024, its revenue had plummeted from $120 million to just $55 million - a 54% decline - after OpenAI improved ChatGPT's native capabilities.
 
This wasn't an isolated incident. It was the canary in the coal mine.
 
The fundamental problem is defensibility. Most AI startups are building on top of foundation models they don't control, using APIs they can't differentiate, solving problems that the model providers themselves will eventually address. They're features masquerading as companies, and investors have noticed.
 
As we explored in May, the shift from knowledge to intelligence means that simply having access to AI models isn't enough. You need to build ‘systems of intelligence’ - complete solutions that combine AI with proprietary data, workflows, and deep vertical expertise. Without these elements, you're just another wrapper waiting to be disintermediated.
 
 
The Circular Revenue Illusion
 
Bill Gurley's warning about circular revenue in AI investments has proven prophetic. Major tech companies are "using their balance sheet to drive their income statement" - investing billions in AI startups that turn around and spend those billions on cloud services from the same investors.
 
Microsoft invests in OpenAI. OpenAI spends on Azure. Amazon invests in Anthropic. Anthropic spends on AWS. The money goes round and round, creating an illusion of revenue growth whilst masking the true economics of these businesses.
 
This shell game works for the giants at the top of the AI stack. But for everyone else, it's created impossible expectations. Investors see OpenAI's revenue "growth" and expect similar trajectories from applied AI startups. When these companies reveal their actual unit economics - high compute costs, low margins, commoditised offerings - the valuations collapse.
 
Up until last year, it was common for us to encounter valuations in some AI sectors as high as 50x multiple of revenue due to investor enthusiasm outpacing financial performance. Those days are over. The market has learned to distinguish between real revenue and circular flows, between genuine product-market fit and temporary AI hype.
 
 
Why Traditional SaaS Is Winning
 
The PitchBook data reveals something profound: by Series A and B, traditional SaaS companies in Europe command 42% higher valuations than AI startups. This isn't because investors don't believe in AI - it's because they've learned what actually creates value.
 
Traditional SaaS companies have predictable revenue models. They have established customer acquisition costs and lifetime values. They have defensible moats built on network effects, switching costs, and deep workflow integration. Most importantly, they use AI as a tool to improve their economics, not as their entire identity.
 
Consider the numbers: whilst European AI startups show median valuations of €31.9 million at Series A/B, SaaS companies achieve €45.3 million. This isn't a small difference - it's a fundamental repricing of what the market values. Investors are voting with their chequebooks, and they're choosing ‘boring’, predictable SaaS businesses over exciting but unproven AI ventures.
 
As we discussed in February, only 15.4% of startups now graduate from Seed to Series A. In this brutal environment, investors want companies with proven economics, not promises of AI transformation. The premium has shifted from potential to performance.
 
 
The Path Forward for Founders
 
The implications for European founders are clear and actionable. Stop positioning your company as "AI-first" - it's become a red flag, not a green light. The data proves it: you'll raise less money at a lower valuation than if you positioned as a traditional software business that happens to use AI brilliantly.
 
Seed and Series A multiples remain stretched on potential, while later stages compress as scale, efficiency, and profitability increasingly matter. This compression is particularly testing for AI startups because their unit economics often worsen with scale due to compute costs, whilst traditional SaaS economics improve.
 
The winning formula is increasingly clear: deep vertical expertise, genuine customer problems, sustainable unit economics, and AI as an accelerant rather than the entire proposition. Build a business that would succeed even if AI disappeared tomorrow, then use AI to make it unbeatable.
 
Remember the takeaway from our recent piece on the anti-pitch: investors fund founders who don't need them. The same principle applies here. Investors aren't funding AI experiments - they're funding businesses that will thrive regardless of which model wins the AI race.
 
 
The Takeaways
 
The AI funding paradox resolves itself once you understand what's really happening. It's not that AI isn't valuable - it's that AI alone isn't a business. The PitchBook data tells the complete story: overall, European AI and SaaS startups are valued identically at €8.7 million median valuations. But dig deeper, and you find that early promise evaporates into harsh reality.
 
At Seed stage, AI commands a 23% premium - investors are still betting on potential. By Series A, that premium has transformed into a 42% discount. This isn't market inefficiency - it's market intelligence. The valuation flip happens at precisely the moment investors stop funding promises and start demanding proof.
 
For UK and European founders, this creates both challenge and opportunity. Unlike Silicon Valley, we have fewer foundational model companies and less infrastructure investment. Most European AI startups operate in the applied layer, building vertical solutions for specific industries. This isn't a weakness - it's reality. And the data suggests embracing this reality is more lucrative than fighting it.
 
The startups raising premium rounds aren't the ones shouting loudest about AI. They're the ones quietly building monopolies in specific verticals, using AI as a tool rather than an identity. SaaS companies raising at premiums to AI companies aren't apologising for not being "AI-native." They're too busy building sustainable businesses with defensible moats, predictable revenue models, and improving unit economics.
 
The smart money has already figured this out. The global analysis of 565 AI companies confirms that applied AI categories trade closer to enterprise SaaS benchmarks than frontier AI valuations.
 
For founders navigating this landscape, the message is clear: the age of raising money on AI promises alone is over. The age of building real businesses that happen to use AI brilliantly has just begun.
 
Position accordingly, and you'll raise better rounds with less dilution than the "AI-native" companies still searching for a business model. The PitchBook data proves it. Now you have the evidence to act on it.

Let's talk.