"What types of businesses outside of the labs will get stronger as raw intelligence explodes?"
The businesses whose margin goes up:
- Owners of physical bottlenecks. Power generation, transmission, grid interconnects, data center land and permits, cooling, and the long tail of the chip supply chain. Intelligence scales with energy and atoms, and those don't follow Moore's law. Permits and interconnect queues are the moat.
- Liability-bearers and licensed entities. Insurers, banks, auditors, licensed clinicians and pharmacies, anyone who signs and can be sued. A model can produce the answer, but someone still has to be accountable for it. That accountability gets more valuable as the volume of answers explodes.
- Proprietary feedback loops. Businesses that generate outcome data that isn't on the internet: claims histories, manufacturing yields, logistics telemetry, clinical outcomes. Frontier models commoditize public knowledge. Closed, compounding ground truth is the part they can't train their way to.
- Distribution and trust. Brands, two-sided marketplaces, and creators with real audiences. When content and products become effectively infinite, attention and credibility are the scarce inputs. Your Morning Brew lesson gets more true, not less.
- Owners of the customer in fragmented physical services. Roll-ups of HVAC, accounting, property management, home health. AI collapses their back-office cost, and the operator who owns the relationship keeps the margin. The software vendor selling to them doesn't.
- Industries with elastic demand and an experimental bottleneck. Biotech, materials, chip design. Ideas get cheap, so the constraint shifts to wet labs, test benches, and trial throughput. Lab automation, CROs, and anyone with fast physical iteration cycles win.
- Verification, security, and provenance. When generation is free, checking becomes the job: testing, evals, cybersecurity, audit trails, authenticity. Every 10x in AI output creates demand for verification.
- Deployment and integration, with a big caveat. Enterprises absorb change far slower than models improve, so the gap between what's possible and what's deployed widens. That gap is a business. But firms that sell hours of cognition get squeezed on price every model release. The durable version moves toward outcome-based pricing, owned workflows and IP, and embedded accountability. That's the strategic fork for a firm like Tenex: be a liability-bearing operator inside the client, not a labor arbitrage play that gets cheaper every quarter.
- Human-premium goods. Live events, in-person experiences, craftsmanship, status goods. Scarcity of human time becomes the product.
Nov 1, 2026.