Teardown

Daily digest · 2026-08-03

Scan #023: The wedge is not the moat

Four incumbents that bought their positions cheap — Dollar Tree, Majesco, Expeditors, NRG — and four emerging companies (Zip, FERMÀT, Forto, Rebar) racing to convert a clever wedge into something defensible before it commoditizes.

A thread runs through today’s eight: the entry wedge and the durable position are different things. All four incumbents hold positions that were bought — a divested drag, an LBO roll-up, forty years of compounding branch economics, 13 GW of gas below replacement cost — while all four emerging companies are discovering that the wedge that got them in (an intake form, a post-click funnel, a digital quote, a blueprint reader) is exactly the part their platform dependencies and the foundation-model wave commoditize first. The forwarding pair makes the point directly: Expeditors never built the app, and it outlasted the people who did.

Dollar TreeRetail · Incumbent, well positioned. Shed Family Dollar for roughly $1B — a decade after paying $8.5B for it — and emerged a cleaner, single-banner business. Multi-price rollout is driving traffic and the comp engine is running, though heavy China sourcing keeps tariff math permanently on the risk ledger. Store-condition and understaffing complaints haven’t gone away; the divestiture just stopped them from being subsidized.

MajescoInsurance software · Incumbent, well positioned. Thoma Bravo’s 2020 take-private (~$729M) has been followed by a quiet acquisition spree — ClaimVantage, Utilant, Global IQX, Vitech — building a ~$500M-revenue core-systems consolidator in Guidewire’s shadow. The tell to watch: how much of that growth is organic; core-systems buyers barely leave public reviews, and the audited numbers stay private.

ExpeditorsLogistics · Incumbent, well positioned. The anti-Flexport: no acquisitions, no venture spend, branch P&Ls with operators paid on district profit — and it outlasted a decade of digital forwarders that promised to disrupt it. The deferred technology bill and DSV-scale consolidation are the real threats, not the startups. Q2 numbers land August 4.

NRG EnergyEnergy · Incumbent, well positioned. Bought ~13 GW of dispatchable gas from LS Power for ~$12B — arithmetic works out near $920/kW, well below replacement cost — while holding the largest Texas retail electricity book in a supply-short, data-center-hungry ERCOT. The retail complaint file (Reliant, Direct Energy) and post-deal leverage are the ballast the data-center narrative glosses over.

ZipSupply chain · Emerging. Procurement intake-and-orchestration layer (~$371M raised, $2.2B mark, Oct 2024) sitting on top of the ERPs it routes around. Revenue estimates diverge wildly — $193M vs numbers implied by an ~$82K average contract — and the open question is whether the intake control point converts to suite displacement before Coupa (Thoma Bravo) and SAP bundle agentic intake into what customers already pay for.

FERMÀTEcommerce · Emerging. AI-native post-click funnels for DTC brands; $74M raised, $45M Series B (VMG, 2025), claimed 5x ARR growth — off an undisclosed base. Detectable store installs are down ~7.6% year over year (Storeleads, Aug 2026), pricing reportedly runs $36K–$84K/yr against $49/mo self-serve rivals, and Glassdoor is mixed. The question is whether the cross-brand “Commerce Brain” proves lift before Shopify’s native AI collapses the price floor.

FortoLogistics · Emerging. Berlin digital forwarder, $610M raised, $2.1B mark (Mar 2022) — then a failed Moelis-run sale process (Feb 2025), a founder-CEO replaced, and roughly a third of staff cut in April 2025. The Maersk acquisition that made headlines never completed; Maersk’s existing stake now hangs over any exit price. The FortoLabs SaaS pivot (Oct 2025) is an attempt to reprice the company as software before the shrinking forwarding book prices it as distressed freight.

RebarConstruction · Emerging. Founded October 2024; $14M Series A (Prudence, March 2026) for computer-vision takeoffs and quoting for commercial HVAC distributors — seven of its customers invested. All growth numbers are company-supplied multiples off undisclosed bases. Open question: does the HVAC-specialist data layer become the quoting system of record before better-funded generalists (Bobyard, Beam AI) and multimodal models make blueprint reading free?

Full digest: Scan #023


Full deep dives