Teardown

Daily digest · 2026-09-04

Scan #046: a16z bets $30M on procurement's agent-orchestration wedge, Sun King's 40k-agent PAYGO book keeps compounding into securitisation — meet four US incumbents where the operating math has quietly turned: Cornerstone's creditors just organized under a cooperation agreement, CVS's DOJ prior-auth probe collides with 900 store closures, the Hartford is pruning Personal Lines rather than defending it, and Costco just added a $1.33B/quarter membership-fee annuity that no rival can copy

Four emerging companies buying category permission with 2026 rounds — Sun King on a Series D-anchored $260M+ pushing off-grid solar to 200M people by 2030 and now securitising the receivables with Citi, Lio's $30M/a16z Series A that pitches procurement as replacing BPOs with agents (not tools), Vecna Robotics' $100M Series C plus Karl Iagnemma reset trying to make the standalone RaaS pallet-jack math work before Locus/Fetch/OEM autonomous forklifts converge, and Stand Insurance's $35M/Eclipse Series B for a physics-based California wildfire underwriter that has to prove one loss-ratio season on Concert Specialty paper — meet four US incumbents where the operating math has turned quietly against them: Cornerstone Building Brands with 90%+ of creditors under a Moelis/Paul Weiss cooperation agreement ahead of CD&R talks by February 20, 2026, CVS Health where a July 2026 FTC insulin settlement, 1,170 store closures, Blue Shield California PBM loss to Amazon+Cost Plus and DOJ Aetna prior-auth probes are all attacking one leg each of the vertical, The Hartford quietly pruning to Business Insurance small-commercial (88 combined ratio, 5.5% renewal price) rather than defending Personal Lines direct (-7% Q2 2026), and Costco just adding a first-since-2017 $1.33B/quarter membership-fee annuity on top of 92.7% renewal — the moat is compounding, not eroding.

Today’s eight surface one thread: whether the incumbent’s cost structure is the thing that will kill it, or the thing that will save it. Four emerging companies are trying to price the incumbent’s structural cost against them — Sun King against grid extension, Lio against BPO seat pricing, Vecna against manual labor plus manual jacks, Stand against admitted-market ISO rates. Four incumbents are on the other side of the same math: three (Cornerstone, CVS, and — arguably — the Hartford Personal Lines run-off) are watching their moat get monetized against them; one (Costco) just proved that with 79.6M paid households at 92.7% renewal, the moat is the moat.

Sun King — Energy · Emerging. The Nairobi off-grid PAYGO solar operator has reached ~82M customers on ~$1.5B of cumulative financing, closed a $260M BeyondNetZero-led Series D in 2022 that now anchors a $5.6B / 50M-kit / 200M-people plan through 2030, and just added a $150M Ethiopia commitment in March 2026. The 40,000-Energy-Officer field network collecting daily mobile-money PAYGO payments is now the securitisable receivables book Citi has underwritten twice — and the reason both grid extension in peri-urban Kenya and M-KOPA’s smartphone-finance pivot are the two things that decide whether this compounds or plateaus.

Lio — Supply Chain · Emerging. The Munich-founded YC S23 procurement AI just banked $30M Series A from a16z (Vlad Keil, Lukas Heinzmann, Till Wagner, not the names our funding-note trackers had). Positioning is deliberate: replace the BPO seat, not the ERP tab. The bet is that a standalone agent-orchestration layer above SAP Ariba and Coupa retains enough of the workflow to defend a $30M+ round once those incumbents ship their own agents “good enough” inside the systems that already own PO, contract and supplier-master data. That is a live question, not a settled one.

Vecna Robotics — Logistics · Emerging. $100M Series C in 2024 with a Karl Iagnemma-led reset in late 2024 after Blackstone-era stumbles, deployed autonomous pallet jacks at DHL, FedEx, GEODIS, Milton CAT and Lineage. The uncomfortable question buried in trade coverage: does a $3-5k/month RaaS pallet-jack actually beat a human plus a manual jack across a two-year deployment, or does the standalone economics story never get to 24-month payback the way Berkshire Grey’s never did — forcing a strategic exit. Locus, Fetch (Zebra), ForwardX, Geek+ and Toyota/Jungheinrich/Crown OEM autonomous forklifts are all closing on the same shipper wallet.

Stand Insurance — Insurance · Emerging. Vishaal Melwani’s AI-native California wildfire homeowners MGA raised a $35M Series B led by Eclipse in October 2025 (not 2026 as the funding tape suggested) on top of prior seed capital — total ~$50M+, sitting on Concert Specialty surplus-lines paper. The physics-based structure-level “World Model” underwriting is either about to be validated in one full California fire season against Kin, Delos and Openly’s LiDAR-plus-ML approach — or the up-to-60% Frontline mitigation credits will subsidise the risk faster than the model prices it. That’s a falsifiable question that answers itself in twelve months.

The Hartford — Insurance · Incumbent, well positioned. Not because it is unassailable — because it is doing the harder thing. Business Insurance small-commercial’s 1.66M policies in force, 5.5% renewal price and 88 combined ratio in Q4 2025 are compounding. Personal Insurance direct-and-AARP written premium is down 7% Q2 2026 as intense shopping crushes the direct-response P&C economics. Management is not defending the losing side; the 2018 Talcott divestiture and the Wellington-Hartford Funds sale precedent both say Hartford prunes to the profitable core rather than holding revenue at any margin. That is the discipline that separates well-positioned incumbents from at-risk ones — and it is precisely what Chubb tried and failed to buy in March 2021 for $70/share, when HIG closed above $130 by early 2026.

Costco — Retail · Incumbent, well positioned. This is the base case: ~900 warehouses, 79.6M paid households, 92.7% US renewal rate, ~4,000 SKUs at capped 14% markup running against Walmart’s ~120,000 at ~25% gross. The first membership-fee hike since 2017 (September 2024, $5 to $65 Gold Star / $10 to $130 Executive) is now flowing through — Q1 FY2026 fee income $1.33B, +14% YoY, materially ahead of the 8-10% analysts had modeled. Fee income is near-100% margin. The moat is not the SKU count; it is the promise that Costco cannot mark up beyond 14% without breaking the format’s own gravity — which is the one thing Sam’s Club’s redesign, Amazon Prime and Instacart cannot copy without ruining their own economics.

CVS Health — Retail · Incumbent, at risk. Four attackers, one leg each of the vertical. Caremark PBM lost the Blue Shield California book to Amazon Pharmacy + Cost Plus + Navitus in 2024 and settled with the FTC on insulin dispensing in July 2026. Aetna’s star ratings, prior-authorization denials and DOJ probes are producing $1B+ EPS swings and now sit inside a broader federal investigation. Front-of-store retail is in structural decline — 1,170 stores gone by end of 2025 (900-store plan announced October 2023 completed early). And David Joyner (CEO since October 2024, ex-Caremark) is now the target of Glenview Capital’s break-up campaign. The counterargument — vertical integration is the moat — has to explain why every leg is losing share simultaneously. It doesn’t.

Cornerstone Building Brands — Construction · Incumbent, at risk. The CD&R 2022 $5.8B take-private on top of the 2018 NCI+Ply Gem merger left CNR carrying ~$1.5B+ of 8.75%-9.50% secured notes into a single-family housing print that has not cooperated. On February 20, 2026 Bloomberg reported >90% of creditors under a cooperation agreement with Moelis and Paul Weiss retained — the textbook pre-LME setup. Ply Gem, Silverline, MI, Simonton and Great Lakes windows all sit at recurring homeowner complaints on installation, warranty and glass failure that a direct-to-homeowner attacker with a modern service ops layer (think what Renewal by Andersen did for the premium tier) could turn into share loss even as CD&R sponsor-owns through the next holding period.

Full digest: teardown.ai/digest/2026-09-04.

Full deep dives