But here’s the catch: unicorn status doesn’t guarantee a strong business. In fact, many of these AI “success stories” are at risk of becoming what VCs now call zombiecorns – startups with sky-high valuations, shallow tech credentials, and weak revenue models. For first-time founders especially, the danger is clear: chasing the wrong milestones can put your company on shaky ground from day one.

Despite the record investment flowing into tech right now, it would be a mistake for the industry to approach it all with rose tinted spectacles. Some industry experts are even starting to warn of an “AI chill”, with the market exhibiting bubble characteristics. Start-ups need to be ready and focus on capitalising on AI in the right way, rather than chasing capital without a plan.

The better question to ask is not: “How fast can I become a unicorn?” It’s: “How do I build something that will last?”

From unicorns to thoroughbreds

AI, undoubtedly, is a generational technological force in the economy. Its impact will be felt across businesses, from wholesale operational optimisation and automation through to the decisions made by individuals every day. The best AI firms, however, will approach it in a comprehensive and meaningful way.

A better framing for early-stage entrepreneurs comes from VC Saul Klein: instead of unicorns, think thoroughbreds. Thoroughbreds are companies that hit $100 million in annual revenue – not just valuations – through repeatable, resilient business models.

Why does this matter for today’s sprawling AI market? Because revenue is proof that customers are willing to pay for what you’re building, and that your company can scale sustainably. Valuation is only a reflection of short-term investor confidence, while sustainable revenue reflects actual market traction. Many of today’s AI leaders may not be around in a decade’s time – the opportunity for scalable, focused and resilient business is real.

And just like in horse racing, there’s a natural progression. Colts (ventures with ~$25 million revenue) are the ones showing real growth signals. For startups, aiming to become a colt first is a far healthier milestone than chasing unicorn badges.

Some of today’s most durable AI ventures followed this path. Take UiPath, the Romanian-born automation company: it grew steadily by targeting enterprise workflows before going public and now has $1.66 billion ARR. Similarly, DeepL, the German AI translation company, avoided the hype cycle, quietly scaling a product people paid for before raising at multi-billion valuations. These companies weren’t chasing headlines – they were building thoroughbreds.

Wrapper-ware is not the answer

The easiest mistake founders make today is building wrapper-ware – startups that simply package ChatGPT or another foundation model into a new interface. Yes, it’s fast to launch, and yes, it looks good on a pitch deck. But these ideas are shallow, replicable, and quickly commoditised.

The founders who will stand out are those who iterate at “AI native speeds” and:

• Build around infrastructure: robust data pipelines, vertical-specific models, and automation frameworks that underpin mission-critical operations. Hugging Face, for example, became a thoroughbred by building a trusted platform around open-source model sharing
• Solve deep problems: industries facing labour shortages or inefficiencies, such as healthcare, logistics, and advanced manufacturing, are crying out for automation. Covariant, an AI robotics company, has seen strong adoption by tackling automation in warehouses – an unglamorous but enormous market
• Think long-term: a product demo may win you a round, but only sticky customer bases, like those built by Synthesia in AI video generation, prove staying power

As a rule of thumb: if your product can be replicated by a weekend hackathon or two, it’s likely not defensible enough to carry you to colt status, let alone thoroughbred scale.

Focus on the big picture

The most promising AI startups aren’t consumer apps or productivity widgets. They’re companies quietly transforming sectors where traditional software fell short. Robotics in elder care, AI-enabled automation in logistics warehouses, or predictive maintenance in heavy industry – these aren’t flashy, but they address fundamental economic needs.

McKinsey research estimates that AI could add $4.4 trillion annually to global GDP, with the largest gains in sectors like manufacturing, healthcare, and logistics. Yet these are precisely the industries where software adoption has historically lagged. The opportunity lies not in chasing consumer buzz but in tackling structural inefficiencies.

For European founders, there’s an underappreciated advantage. Unlike in Silicon Valley, where abundant capital fuels unsustainable burn rates, European investors have learned to be disciplined. Smaller fund sizes and stricter disclosure rules mean they focus on fundamentals: revenue growth, unit economics, and sustainable business models.

It may feel like a disadvantage when you’re raising – but in the long run, it pushes fast moving founders to build stronger companies.

In the end, hype fades. But a business with real customers, strong fundamentals, and scalable economics will keep running long after the unicorns stumble. For founders, that’s the real prize – not the mythical horn, but the staying power of a thoroughbred.

For more startup news, check out the other articles on the website, and subscribe to the magazine for free. Listen to The Cereal Entrepreneur podcast for more interviews with entrepreneurs and big-hitters in the startup ecosystem.