Teardown

Daily digest · 2026-07-30

Scan #019: Receipts and asking prices

Four emerging companies whose numbers are still mostly self-reported and four incumbents being repriced against proof — Freehand, Fleek, Ominimo, Gritt, Carvana, Ryder, USI, Core & Main.

The thread running through today’s eight: the gap between the receipts a company can produce and the price being asked of it. Carvana printed record everything and lost 15-20% after hours because the guidance didn’t cover the multiple; Core & Main derated by a quarter while its moat didn’t move; Ryder spent eight years converting cyclical revenue into contracts and just got its proof at the bottom of the cycle. On the emerging side the receipts are thinner than the headlines — Freehand’s autonomous-payment claims are entirely company-sourced, Gritt’s 4-5x install multiplier is unaudited, Fleek has never disclosed GMV, and Ominimo’s €1.4B mark rests on loss ratios proven in exactly one underpriced market. Two deliberate pairings today: insurance from both ends (Ominimo the carrier-challenger, USI the broker-incumbent) and construction from both ends (Gritt’s robots, Core & Main’s pipe).

Freehand — Supply chain · Emerging. AI agents that audit, dispute and pay Fortune 500 freight and supplier invoices inside existing ERPs — founded February 2024 by Pando’s two founders, surfaced from stealth in February 2026, $75M Series B co-led by Battery and NewRoad announced July 29. Claims ~50 customers (Meta, Unilever, Pfizer, J&J) and billions paid autonomously — but after two years in stealth there is not a single independent review, no disclosed ARR, and the 5-10% spend-recovery figure is the company’s own. The open question: does that recovery persist after year one’s accumulated leakage is cleaned out, and does no-human-in-the-loop payment survive its first material error before SAP, Oracle and Coupa bundle agentic audit into the systems of record Freehand merely writes into?

Fleek — Ecommerce · Emerging. London B2B marketplace for the used-clothing bale trade — ~2,000 graders in Karachi, Delhi and Dubai selling to 50,000+ retailers in 100+ countries; $25M Series B in July 2026 led by Burda (Vinted’s early backer), with eBay joining, $45M total. What the funding coverage skips: Fleek has never disclosed GMV, take rate or valuation, and every completed shipment teaches wholesaler and buyer who each other are — the WhatsApp disintermediation problem is structural. The bet has quietly shifted to Fleek Sort, a vision model that grades a garment from one photo; the open question is whether it becomes the pricing standard inside sorting hubs before cheap off-the-shelf vision models let both sides route around the commission.

Ominimo — Insurance · Emerging. Hungarian motor insurer launched 2024, ~7% of its home market in year one, annualised GWP from €26.3M (2024) to ~€306.8M (July 2026), profitably — Serbia’s first unicorn at €1.4B on barely €30M raised, with Zurich fronting its expansion paper. The caveats under the headline: the 92-100% combined ratio is company-reported and unaudited, per-market loss ratios outside Hungary aren’t broken out, and the edge was proven against CEE incumbents that rate on a handful of variables. Now it wants its own licence, Western Europe and a 2027 US entry — swapping capital-light MGA economics for balance-sheet risk in markets where incumbents already rate on dense data.

Gritt — Energy · Emerging. San Francisco robotics startup, out of stealth July 21 with a $26M Series A led by Obvious Ventures: it retrofits rented skid steers and off-the-shelf Kawasaki arms with AI to place utility-scale solar panels, claiming crews go from ~800 to 3,000-4,000 panels a day and 2.8 GW contracted over 18 months. Every one of those numbers is company-reported, its customers are unnamed, and even its total funding is inconsistently reported across outlets ($32M-$34M). The wedge is placement only — humans still fasten, rack and pile — so the question is whether retrofit AI absorbs the rest of the install stack before vertically integrated rivals (AES’s Maximo, Terabase’s Terafab) and post-2025 tax-credit phase-outs shrink the market it’s automating.

Carvana — Retail · Incumbent, well positioned. The online used-car retailer completed the decade’s best comeback — $3.72 a share and near-bankruptcy in December 2022 to Q2 2026’s record 197,325 retail units, $7.38B revenue and $513M net income (July 29 release) — and the stock fell 15-20% after hours anyway, because $2.7-3.0B of 2026 EBITDA guidance and a margin down from 12.4% to 10.4% don’t cover a ~$73B market cap. The parts other coverage soft-pedals: Ernest Garcia II’s securities-fraud history and the family’s control stake, the Hindenburg paper trail, and persistent title-and-registration complaints. Verdict: well positioned — 1.6% of a 40-million-unit market with the only national reconditioning-and-logistics machine; the risks threaten the stock more than the machine.

Ryder — Logistics · Incumbent, well positioned. The company that invented US truck leasing spent eight years shrinking away from it: 56% of 2018 revenue was leasing and rental, while by 2024, 61% came from contractual dedicated and supply-chain work (Whiplash, Dotcom, Cardinal acquisitions). The receipt arrived at the bottom of the cycle — a seventh straight quarter of comparable EPS growth ($3.73, +12%, July 23) through the longest freight recession on record, with 17% adjusted ROE. The complaints are real (BBB maintenance gripes, a 3.4 Glassdoor, the 2019-20 residual-value writedown that showed what happens when used-truck math breaks), but the transformation has now been stress-tested in the worst tape imaginable.

USI Insurance Services — Insurance · Incumbent, at risk. Valhalla, NY middle-market brokerage — $2.78B of 2024 revenue, 200+ offices, KKR/CDPQ-owned since the $4.3B 2017 buyout. While Hub files confidentially and Acrisure flags a $32B listing, USI has gone quiet: it fell out of the Business Insurance top 10 in 2024 for the first time in over a decade, closed exactly one acquisition that year, carries roughly 7x leverage with ~2x coverage, and defends its producer bench with non-compete litigation while rivals recruit with public-currency equity. Verdict: at risk — profitable and sticky, but losing the scale race that defines the industry, and the longer the IPO window stays open without USI in it, the more the silence reads as the story.

Core & Main — Construction · Incumbent, well positioned. The largest US pure-play waterworks distributor — pipe, valves, hydrants, meters from 370+ branches; $7.65B revenue and $931M adjusted EBITDA in fiscal 2025 — has derated from above $60 to ~$45 since late 2024 on a guidance cut and the fear that its growth was M&A plus PVC inflation. The findings that cut both ways: EBITDA has been roughly flat for four years ex-acquisitions, and an August 2024 PVC price-fixing suit names it as an unindicted co-conspirator (it denies wrongdoing). Verdict: well positioned — ~17% of a $39B fragmented market where the product is too heavy to ship, half of demand is non-discretionary municipal repair-and-replace funded by IIJA’s $55B water tranche, and there is no disruption vector for 24-inch ductile iron.

Full pages linked above. Yesterday’s scan: Scan #018.

Full deep dives