AI vertical companies need to own the entire customer ecosystem, not just replace a point solution

Harry Stebbings puts the stakes in concrete terms. To justify a valuation of $15 billion to $20 billion, a company like Sierra cannot simply build a better next-generation replacement for existing service software. It has to own the entire customer ecosystem for a business. That is a materially higher bar than the one most AI vertical startups are currently pitching against.

The distinction matters because a large share of the current AI vertical market is being built on a displacement thesis: find an incumbent SaaS product, build an AI-native version, and capture the category. Stebbings is arguing that displacement, even when successful, does not produce the kind of value that justifies the largest current AI valuations. Owning a point in the workflow is not the same as owning the workflow.

The valuation logic is worth unpacking. If a company replaces a service cloud product and captures that product’s revenue multiple, the ceiling is set by the incumbent’s market cap, not by the full scope of the customer relationship. Stebbings is saying the companies being priced at $15 billion to $20 billion today need to be priced against a different ceiling entirely: the value of the whole customer ecosystem that passes through them. That is an argument about what kind of business a vertical AI company has to become, not merely about how good its core product is.

Separately, Eddie Lazar describes building a tool to assess what he calls the AI resiliency of public companies, using a vibe-coded methodology. That kind of exercise suggests investors are already trying to distinguish companies structurally positioned to hold wide scope from those occupying narrow roles. A company that can be categorized as resilient by that framework is almost certainly one that controls more of the customer relationship than a single point solution allows. The framing is early and the methodology informal by Lazar’s own description, but the impulse behind it is the same one Stebbings is naming: investors want a way to identify which AI businesses are building toward ecosystem ownership and which are building toward feature parity with an existing tool.

The only way I can justify 15 or 20 billion in market cap for Sierra is not if I build a slightly better next generation service cloud. If I am the entire customer ecosystem for your entire business. Harry Stebbings

For founders, the implication is that the wedge strategy, entering through a narrow integration and expanding from there, may not produce a defensible position if a well-capitalized competitor is willing to own the full stack from the start. A wedge that does not have a credible path to ecosystem control may be a wedge that never becomes anything larger. The question founders need to answer is not whether their product is better than the incumbent’s, but whether their product becomes the surface through which a business manages its relationship with its own customers.

The counterargument to ecosystem ownership as a strategic frame is that most B2B workflows do not have a single control point available to be captured. Integration dependencies, data siloes, and procurement inertia all limit how much of a customer relationship any one vendor can realistically hold. A company that sets out to own the entire ecosystem may find that the ecosystem is not actually available for ownership, and that the narrower point-solution position it passed over was the durable one.

That friction between the expansionist and skeptical readings is useful context for why Stebbings frames this as a valuation problem rather than a product one. The debate is not about whether AI vertical companies can displace incumbents. The debate is about what the ceiling looks like once displacement is achieved, and whether that ceiling justifies the capital being deployed today.

None of this is settled. The full-stack ownership thesis is a valuation argument, not a proven operating model, and the companies that would validate it at scale have not yet done so. But Stebbings’ framing, specific in its figure and its named example, is the kind of claim that tends to recalibrate how investors ask questions in diligence. If the threshold for a large AI vertical valuation is ecosystem ownership rather than category replacement, a meaningful portion of current pitches are answering the wrong question.