Electricity replaced compute as the binding constraint on AI buildout, and the one credible independent tool was acquired out of the market in March 2025.

The category leader has priced itself out of the bottom two-thirds of its own market — but >95% retention means the play is selling beside it, never instead of it.

A 50%-share incumbent took 15,000 dealers offline for three weeks at a measured cost of $1.02B, then got sued for refusing to hand back customer data. The trust rupture is documented, priced, and still open.

Toast is a payments company that sells a POS. It monetises the top line and has no structural reason to tell an operator where food cost is bleeding — which is where a 3–5% net margin actually goes.

Canada has committed unprecedented capital to non-market housing and handed the application process to a sector that cannot use it: roughly half of all community housing units sit with several thousand providers holding fewer than 100 units each, and almost none of them employ a developer. The enterable business is the missing capability, not the building — but the federal window that paid providers to buy that capability closed on 2 February 2026, and until a successor opens, the buyer has the need and not the budget.

The most attractive market structure in the vertical cohort at the application layer — fragmented, 12.7% CAGR, no dominant incumbent — and the two leaders both shipped the obvious AI wedge in 2025, before a new entrant could reach it. Sourced update 2026-09-18, and it argues against the record's own 'fragmented' framing: the ~45% top-five figure describes the agency software layer, but the EVV rail underneath it is far more concentrated. HHAeXchange acquired Sandata on 3 October 2024, combining Sandata's 23 state Medicaid EVV contracts and 20,000+ agencies with HHAeXchange's own aggregator contracts [B]. The switching cost this study already identified now sits mostly inside one company.

The buyer changed. There are ~2,000 dental service organisations where there were ~100 in 2010, and they acquire practices running incompatible systems — creating the same migration wedge as insurance brokerage, but in a market that is growing rather than shrinking.

The pain is real and documented; the customer base is shrinking and the transaction volume driving the wedge is declining 5.1% year over year. A hated incumbent is not the same as a vulnerable one.

The largest and fastest-growing market in the portfolio, and the one where enthusiasm and executable opportunity are most badly mismatched. Budget headlines are not revenue.

The most attractive-looking and least enterable market in the portfolio. Growth you cannot capture is not an asset.

Salesforce earns more CRM revenue than Microsoft, Oracle, Adobe and SAP combined, and the credible free alternative has 45,000 GitHub stars. A new entrant is squeezed between an incumbent it cannot outspend and a substitute it cannot underprice.

Clio just raised $500M at a $5B valuation and spent $1B buying vLex. A market whose leader is actively bundling research, drafting and payments into one suite is the worst possible moment to arrive with a point solution.

The open-source alternative in this category is not a hobby project. General Atlantic valued Odoo at €7B in January 2026. Competing here means facing SAP at 22% share and an $8B open-source company at the same time.

Walk — but for a different reason than the other three. Payroll compliance is a genuine moat, which is exactly why the incumbents are worth $10B and $16.8B. The barrier that would protect an entrant has already been paid for by someone else.

The cheapest category in this research to build and the hardest to get paid for. Ten funded vendors, a $3.2B open-source tier, and a buyer who churns the moment a free tool is good enough.

Two companies hold 79% of a $718M market, and the larger one sells the software to pull revenue from its own diagnostic analysers. It can price the software at zero forever. A software-only entrant cannot answer that.

The most fragmented generic category — the top five vendors hold only 20–25% — and still a walk, because the fragmentation is a symptom of low switching costs rather than an unserved segment.