After the generative AI gold rush, founders are chasing predictable revenue—healthcare, legal and finance are emerging as the safest bets for long-term growth.
Illustration by IMF Alpha editorial · Reviewed by Pedro Marini
The shifting playbook
After two years of hyperactive seed rounds and theatrical demos, reality is starting to quiet down. Investors want predictable revenue and unit economics that hold up under scrutiny. That demand has nudged many AI founders away from grand horizontal plays and toward vertical SaaS — focused products built for a single industry.
This is not defeat. Think of it as a tactical pivot. Horizontal models sell scale as a promise; verticals sell immediate relief. Fewer integration headaches, faster sales with prequalified buyers, and industry-specific data that actually compounds into a defensible moat.
Why the pivot makes sense now
What’s interesting is how these reasons stack: the same industry friction that slows horizontal adoption is what creates the vertical advantage.
Where founders are placing their chips
Veeva’s growth after cloud consolidation is a useful analogy. Life-sciences customers wanted domain workflows, not a generic CRM with an AI sticker. The tech differs now, but buyer psychology looks familiar.
Trade-offs and counterpoints
This shift is not universal. Horizontal infrastructure plays still matter — vector stores, model orchestration, low-latency inference. Big cloud providers and a few platform vendors will capture large-scale margins, while verticals capture value inside industries.
Some investors still believe a strong horizontal layer makes a startup a tidy acquisition for hyperscalers. Others argue the opposite: being industry-specific can make you indispensable to customers and therefore harder to displace.
In practice, then, the market will support both — but they play different games and need different go-to-market muscles.
What this means for M&A, hiring and fundraising
Quick playbook for founders
Why investors should care
Putting AI models into a vertical context is not low ambition; it’s a way to turn bleeding-edge tech into repeatable, billable outcomes. For an investor hunting a multi-bagger, the better bet might be on founders who can convince one industry to pay, then quietly scale that category across similar customers.
Founders who treat domain-specific data as the product and models as the ingredient are the ones likely to survive this funding cycle and emerge with businesses that look steadier and much more profitable.
Enterprises are swapping risky, expensive real-world datasets for generated alternatives. The shift has investment, regulatory, and technical consequences.
Phones are becoming their own AI servers. That matters for privacy, latency, and who wins in silicon and services—cloud is not dead, but its role is shifting.
From NPUs to 4-bit quantized models, on-device generative AI is reshaping apps, monetization and investor bets — and not always in the ways you expect.