Teardown

Daily digest · 2026-08-11

Scan #031: Eight AI wedges into eight legacy books

Four public/entrenched incumbents whose next chapter depends on defending against a new AI-plus-hardware entrant, paired with the specific challenger cutting into their category — insurance, HVAC, retail apparel, and MRO distribution.

Today’s eight are paired on purpose. Four are public or privately-owned incumbents holding fat, hard-earned share in insurance, HVAC, apparel retail, and industrial distribution. Four are the specific challengers most credibly attacking one of the incumbent’s soft points with a lot of new capital and a defensible technical wedge — aerial-imagery underwriting, human-in-the-loop takeoff, overhead-sensor inventory accuracy, and drone-based warehouse reconciliation. The story of the day is what the recent print of each incumbent’s quarter actually said about how much cushion is left.

Progressive Corporation — Insurance · Incumbent, at risk. The largest auto insurer in the US just reported June 2026 net written premiums up ~3% year-over-year — a dramatic deceleration from mid-teens in 2024-25 — and shares fell 9% on July 15 as Wells Fargo and Piper flagged the turn of the pricing cycle. Commercial P&C premiums declined 1.2% in Q1 2026, the first drop since Q3 2017. The Snapshot telematics moat is being replicated by Root, OEM-embedded UBI (Tesla, Ford), and the state DOI-driven forced rate rollbacks that made 2024-25 so profitable are now the ceiling.

Lennox International — Construction/HVAC · Incumbent, at risk. Record margins on the A2L refrigerant transition, but the pull-forward leaves 2026-27 comps ugly against US HVAC shipments already down ~20% in 2025 (AHRI). Carrier’s €12B Viessmann close has widened the premium-dealer network gap; Daikin/Goodman opened a 4.1M-sq-ft plant in Waller, TX — 40 miles from Lennox’s Richardson HQ — and Midea’s Dallas R&D center opened Q3 2025. The direct-dealer moat that carried Lennox to peak margins is being flanked by better-capitalized rivals on both premium and value tiers, with the stock priced near a peak-cycle multiple.

Gap Inc — Retail/Apparel · Incumbent, at risk. Old Navy — ~55% of the portfolio — missed the spring 2026 fashion cycle, Athleta bled share to Alo and Vuori, and Wells Fargo downgraded in August 2026. FY26 sales guidance was cut to 1-2% from 2-3% after Q1. Richard Dickson (ex-Mattel, Barbie halo) has stabilized operations but the merchandising engine hasn’t restarted. The four-brand portfolio is being squeezed from below by Shein/Temu real-time trend replication and above by Uniqlo essentials; Banana Republic sits with no clear position between J.Crew’s elevated basics and Todd Snyder/Buck Mason menswear.

W.W. Grainger — Supply chain / MRO distribution · Incumbent, well positioned. The counter-example to Amazon-eats-distribution. $17B revenue, ~15% EBIT margin, mid-40s ROIC — the enterprise High-Touch franchise (KeepStock vending, Ariba/Coupa punchout integrations, technical sales) is protected by inventory availability and procurement lock-in that Amazon Business has not cracked; Endless Assortment (Zoro + MonotaRO) plays the long-tail game on Amazon’s own economics. The 2011-14 pricing reset — cutting list to close the online arbitrage gap and trading margin for share — is now the case study every distribution CEO points to.

Honeycomb Insurance — Insurance · Emerging. AI-underwritten condo/HOA/small commercial multifamily MGA. No inspectors — aerial imagery plus AI to underwrite the buildings incumbent carriers can’t afford to touch at $2-5K premiums. $275M GWP by year-end 2025, live in 20+ states, extension round June 2026 at higher valuation than the 2024 Series B (Fortune). SiriusPoint is lead reinsurance. Picked the one property segment where the incumbent process is genuinely mispriced — small commercial multifamily is under-served precisely because inspection cost overwhelms premium. If loss ratios hold through a real cat cycle, it’s structural.

Beam AI (Attentive.AI) — Construction · Emerging. 1,100+ paying contractors, 500K+ takeoffs completed, $30.5M Series B led by Insight Partners in Nov 2025 ($48M total raised). The construction takeoff category is being flooded with AI weekly — Beam’s wedge is a QA-reviewed human-in-the-loop layer that delivers auditable accuracy contractors trust. Trunk Tools’ own stated strategy explicitly cedes takeoff to Beam and focuses on adjacent workflows. The risk is sandwich: pure-AI competitors race to zero, and Autodesk/Procore/Trimble bundle “good enough” takeoff into what a GC already pays for.

RADAR — Retail/Commerce · Emerging. $170M Series B May 2026 at $1B valuation (Gideon Strategic Partners + Nimble Partners) — overhead-sensor + AI platform live in 1,400+ storefronts, anchor customer American Eagle Outfitters, claimed 99% item-level inventory accuracy. Physical retail is ~80% of global commerce and has never had the data layer ecommerce takes for granted. The category has a long history of overpromise (Amazon Just Walk Out sold off, Standard AI pivoted to services, Simbe still niche); RADAR’s traction with a marquee chain and $1B mark without leaking pilot-failure war stories is unusual. Real questions live in the hardware CapEx per store, services drag, and pilot-to-portfolio pricing concessions.

Gather AI — Supply chain / Warehouse robotics · Emerging. $40M Series B Feb 2026 led by Smith Point Capital ($74M total). CMU-spinout autonomous drones fly warehouse aisles capturing pallet-position imagery, reconcile against WMS, flag discrepancies. Live at NFI, Barrett Distribution, DHL. More named enterprise customers than better-funded rivals Verity, Dexory, Corvus. Sits on top of Manhattan/Blue Yonder/SAP EWM — which is exactly the risk: if those incumbent WMS suites bundle equivalent CV reconciliation natively, Gather’s per-warehouse economics compress. The head-to-head with Verity in shipper bake-offs is the critical test.

Full deep dives