How to Survive Using AI — and Dominion Capital Group as Your Vendor Partner, 2026–2030

The Vendor Squeeze

Through 2026–2030, the operators who survive won't be the ones who spend the most on AI. They'll be the ones who refuse to let AI capex walk out of their operating margin. Every hyperscaler capex cycle through the last decade flowed the same direction: into the hands of a smaller set of vendors, on shorter contracts, with more pricing power. The sub-$500M operator who treats AI like 2014 cloud — buy a little, run a pilot, scale when the unit economics clear — will absorb the squeeze instead of routing around it.

The tell is already in the 2024 and 2025 capex prints. Big-three hyperscaler spend is up double digits year-over-year, the long tail of model providers is being picked off in primary financings, and the cost of compute is being amortized into longer, lock-in-heavy vendor commitments. Margin compression shows up first in the operators who don't have a durable counter-party on the other side of those contracts. That's the cohort this report is built for.

The rest of this brief lays out where the squeeze lands first, which categories consolidate before the cycle peaks, and what a defensible vendor posture looks like for an operator who can't afford to be locked in and can't afford to be left out.

The AI Productivity Curve, 2024–2030

Set aside the hype-cycle headlines for a moment and look at the underlying productivity index. The compounding curve from 2024 through 2030 is the single most important number an operator can plug into a five-year plan — it tells you how much output the same headcount can produce, which in turn tells you how much labor cost you can defer, and which vendor categories have a defensible margin left after the cycle.

The takeaway: by 2030, the same team produces more than double the 2024 baseline. Operators who price their services assuming flat productivity through the cycle will price themselves out of the market before 2028. The compounding is real, and the laggards will feel it in gross margin before they see it in revenue.

Texas Capital Flows Through 2030

Texas is the cleanest proxy for the 2026–2030 capital reallocation cycle because the deals are disclosed, the verticals are visible, and the operators are sub-$500M. Across energy, real estate, AI-SaaS, data center, and manufactured housing, projected deployment tracks the same thesis: capital follows the verticals where the vendor squeeze is most navigable and the operator margin holds up longest.

Read the bars left to right and the order tells the story: the verticals with the deepest existing operator base (Energy, Real Estate) absorb the bulk of the deployment, while the verticals built specifically for the AI cycle (Data Center, Manufactured Housing) start smaller because the operator tooling is younger. The operators who position between these bars — owning the bridge from Energy capital into Data Center deployment — capture the spread.

Categories That Will Consolidate

Three categories consolidate before 2030, and the operator who picks the wrong counter-party in any of them pays twice — once in margin, once in migration cost. The pattern is consistent across all three: fragmented vendor landscape, falling switching costs, and a buyer with enough leverage to set price.

Software: the long tail of vertical SaaS gets folded into platform plays. The independent operator who built a defensible niche in 2020–2024 will get a friendly call from a strategic in 2027 or 2028; the one who built a me-too product in a crowded category will be acqui-hired or quietly shut down. Pick your software vendors the way you pick your acquisition targets — by the balance sheet that will still be there in five years.

Hardware: the GPU-and-rack layer consolidates faster than the consensus expects because the capex is too large for the operator balance sheet and the supply is too concentrated for the market to absorb disruption. Hardware vendors who look diversified in 2026 will look narrow by 2029. Lock in service-level commitments, not unit prices.

Services: the consulting, implementation, and managed-services layer is the most exposed category of the three. The same AI productivity curve that doubles operator output by 2030 also halves the billable hour for the firms that bill by the hour. The services firms that survive will be the ones that price on outcome, not time.

Building Durable Vendor Relationships

The playbook for sub-$500M operators is the same one Dominion uses for its own portfolio: pick the counter-party before you pick the product, structure the contract around the macro cycle, and keep an explicit exit path so neither side has an incentive to renegotiate under pressure. The operators who come out of this cycle with intact margins are the ones who treated vendor selection as a capital allocation decision, not a procurement decision.

This report is the operator's brief on how to do that in 2026 — before the next capex print, before the next vendor lock-in cycle, before the next consolidation wave turns a defensible niche into an unfundable one. The full 22-page version walks through each vertical with deal-level detail; this page is the strategic frame.