Every few weeks, another vertical AI startup announces a new mega-round: “$100M to transform [industry] with AI”. The speed at which these startups race through funding rounds, and the sheer scale, is dizzying. Beneath it all sits a question few seem to ask: what is this money actually buying?
The answer, upon closer inspection, is startlingly simple. It's buying go-to-market. Not technology. Not data moats. Not network effects. GTM. Field engineers, sales teams, customer success reps, and the entire apparatus of making enterprise buyers comfortable enough to sign.
This should terrify investors. In a world of abundant code and mature procurement, funding sales forces to push undifferentiated software through low friction channels is the entirely wrong thing to fund. The model is broken. The real question is what replaces it.
The answer requires rethinking how vertical software captures value from the ground up. When code is abundant, keeping software closed source is an unnecessary constraint on the customer. The promise of open source SaaS, long stalled by the cost of customization and maintenance, can finally be fulfilled. Imagine instead, thousands of domain experts around the world, customizing an open code base for specific industries using nothing more than a coding agent and their own expertise.
Winner-take-all requires a winner
The software industry spent three decades producing winner-take-all outcomes. The entire VC model depends on it. What if AI applications simply don't generate them?
Every previous technological cycle came with its own set of startup challenges, and venture was the right tool to combat them: winning new surfaces like the browser, rewiring procurement, proving reliability long before the market bought in. SaaS was disruptive to the business model of software. Cloud was foreign, requiring enterprises to abandon the comfort of their own data centers. Mobile demanded entirely new interaction paradigms. These cycles were contrarian, misunderstood, or structurally disruptive. Venture capital thrived because incumbents were slow to respond and startups had the window to build compounding advantages before anyone else showed up.
AI is none of these things. It works. Every CEO wants to adopt it now. For the first time in decades, startups and incumbents are chasing the same territory with the same urgency. This isn't an adoption problem. It's an execution race. And if all that is being funded is a go-to-market apparatus, the advantage evaporates the moment a well-run competitor watches the first mover navigate the product maze, and then walks the cleared path at half the cost.
Risk capital, to be justified, has to offer asymmetric returns. To paraphrase Howard Marks: for great returns it's not enough to be right, everyone else has to be wrong. In AI, everyone is right. So even in the best case, the results of these investments will be average. Average is not how venture funds work.
The product isn't a product
The cost of building software is collapsing toward zero. This single fact restructures everything about product competition in AI. When anyone can build anything, product advantages don't compound. Without network effects or deep operational complexity, all you are doing as a first mover is charting the path for everyone behind you.
AI application companies positioned themselves as middleware: sitting between foundation models and enterprise customers, filling gaps that raw models couldn't handle. Limited context windows, hallucinations, no access to private data. Early movers like Harvey, Hebbia, and Legora deserve credit for identifying and solving them. But these gaps are closing with every model generation and labs are now delivering comparable experiences with far less boilerplate.
Speed of execution is not a moat. Being first to market, as Thiel would put it, means you are paid to make the expensive mistakes that help everyone else. Unless you hold the last technological advantage in a chain, you are perpetually exposed to whoever builds next. And in AI, "next" arrives every few months.
Scale doesn't save you either. In a world where generating code costs nearly nothing, it becomes viable to build narrowly scoped, domain-expert agents that outperform general-purpose tools on every metric that matters. A specialized regulatory compliance agent will always beat a do-everything legal AI for the customers who care most.
Even when you do capture a market, there is no pricing power. The knowledge that your software can be replicated shifts leverage to the buyer. They don't need to actually switch.
It's the distribution, stupid
So the product isn't durable. What about distribution?
Distribution has always been the decisive advantage in vertical software: the company that pushes through procurement barriers, in a category where innovation is exhausted and the channel carries friction, occupies a position that is nearly impossible to dislodge. This is last mover's advantage.
Vertical AI apps are running the same playbook, but neither condition holds. AI is iterating so fast that no one has reached the last innovation in any category. At the same time, the procurement channel has lost its friction entirely. Software ate the world and Covid compressed another decade of digitization into two years. Today, nearly every organization knows how to buy software now. The path is clean. Friction was not just an obstacle for the seller. It was the barrier that kept competitors at bay.
Yet the industry hasn’t adapted. The Field Deployment Engineer model, popularized by Palantir, is now the default sales motion across AI startups. But Palantir operates where friction is abundant: government clearances, compliance, and relationship depth that takes years to accumulate. Eight-figure contracts justify that cost. Most AI startups sell $200K ARR deals into enterprises where the procurement path is wide open.
In a desperate attempt to manufacture winner-take-all effects where none exist organically, the industry has embraced kingmaking: flood a chosen player with capital, install it as the category leader, and let the self-fulfilling prophecy do its work. Once a market is "won" with costly GTM, margins are so thin the company can never become self-sustaining. It expands into adjacent markets, where the war of attrition continues against other well-funded players doing the same thing.
The valuation expectations attached to venture capital assume distribution friction creates lasting advantages. The capital being injected is not funding durable assets. It is funding temporary demand. The business may be real, but the returns won't be. And the need to justify those valuations is what makes these companies fragile: it forces growth at all costs, which forces unprofitable sales motions, which forces the next round, which raises the bar even higher.
The real opportunity is not to outspend these companies on go-to-market, but rather, to disrupt the status quo by building a fundamentally different distribution architecture.
The counterarguments don't hold
Some will say go-to-market and product are inseparable and that FDEs are really de facto product managers that embed with customers, extract tacit knowledge, and feed those insights back into the software over time. Yet, any feature that proves critical for a particular customer is trivial to copy. The real question is not whether FDEs contribute to product development, but whether that contribution translates into pricing power. It doesn't. Customers that know what they want can increasingly get those modifications built directly with AI coding agents.
Others point to switching costs. Once an enterprise has spent months implementing one vendor's system, trained workflows around its quirks, and built institutional knowledge about its failure modes, the appetite for starting over dwindles. But the calculus is shifting. Many buyers already view AI products as simultaneously expensive and easy to replace. As costs continue falling, the ROI equation for building and customizing internally only becomes more attractive. Switching costs matter less when the alternative is not switching to a competitor but building your own.
The outcome-based pricing thesis is the strongest case: land with a customer, iterate on their specific workflows, then transition to pricing tied to measurable results. If executed well,, this is genuinely disruptive. But the implication is non-trivial. Outcome-based pricing means assuming responsibility for a function that was previously strategic to the buyer. At that point, one is no longer just a vendor, but a direct competitor to the customer.Any provided solution must be costly enough to justify outsourcing while remaining sufficiently non-strategic that the buyer is comfortable handing it over. Navigating this balance at scale will prove challenging.
A new distribution engine
Competing on GTM without matching FDE spend is like trying to ride clean during the 2000s Tour de France. The question is not how to outspend them. It's how to build a distribution architecture that makes their spending irrelevant.
The answer is to rethink how vertical AI software is sold entirely. Release the core product as open source and let customers adopt it on their own terms. Instead of sending FDEs on-site, give customers the ability to open private issues on a repository, have their own coding agents assemble the changes, and rely on an automated process to merge and maintain those contributions upstream. In this model, the product itself is free. The value captured is the ecosystem that is formed around it: agent-first documentation, a contribution model that makes upstream merges seamless, and domain-specific system integrators who are part prompt engineer, part consultant. The software is abundant. The ecosystem is scarce.
This is genuinely disruptive, and it should feel uncomfortable. The code is published openly for the world to see. Anyone, including competitors, can fork it. This deliberately burns the bridges that proprietary software companies rely on for defensibility. That is how the last mover wins: make the product free, make switching costs zero, and compete on the only thing that matters, building the best ecosystem for modification, extensions, and domain-specific customization.
A new breed of system integrators emerges in this model. Not the Accentures and Deloittes of the old world, but prompt engineers and AI-native consultants who customize an open source system for specific industries, workflows, and regulatory environments. They make money for themselves by building on top of an existing platform. Think force.com. Salesforce didn't just sell software, it built an ecosystem where thousands of independent builders made a living extending the core product. The same dynamic applies here, except the builders are armed with coding agents, and the platform is open.
Open source SaaS is not a new idea. The conventional wisdom was sound: enterprises wanted customization but lacked the engineering capacity to maintain forked software, and open source vendors couldn't capture enough value to sustain themselves. But that was a world without abundant AI. Coding agents are the enabling technology that changes the equation. They collapse the customization barrier that keeps enterprises locked into proprietary SaaS.
What makes this model structurally durable is that it is aligned with what customers actually want. Enterprises don’t want to be locked into a vendor’s roadmap or pay for features they’ll never use. They want software that solves a specific problem, can be modified as needs change, and never holds them hostage. The open source model gives them exactly this and aligns with the customer’s interests. Every SaaS company that charges per seat for a product the customer could plausibly build internally is swimming against the current. Take the world as it is: software is abundant, customization is cheap, and buyers have leverage. Build a business that thrives under those conditions rather than one that depends on pretending they don't exist.
Meanwhile, the incumbents are funding your customer education. Every dollar spent on FDEs, onboarding, and enterprise handholding is a dollar spent teaching the market what AI can do, how to evaluate it, and how to integrate it. Being first to market means you are paid to educate customers for everyone who follows.
The entire VC thesis depends on one player capturing a market and extracting monopoly rents. That model assumes proprietary software, friction-laden channels, and winner-take-all dynamics. Remove those assumptions and the funding logic collapses. And the world is removing them whether the industry likes it or not. Software costs are falling toward zero. Procurement friction has evaporated. AI coding agents are making customization trivial. The founders who recognize this early and build for it will define the next decade of vertical software-not because they raised the most money but because they built something that couldn't be killed.
Let everyone else fund the field engineers, the brand campaigns, and the kingmaking rounds. Then walk the cleared path with a business that was built to last from the start.