Teardown

Daily digest · 2026-07-26

Scan #015: The window trade

Eight companies priced on temporary dislocations — Pallet, Torus, PermitFlow, Syrup Tech, Slide, Infios, Williams-Sonoma, and Enphase.

The thread running through today’s eight is the window trade: businesses whose economics were underwritten on a dislocation that has a closing date. Slide harvested Florida’s insurer collapse and the Citizens well is now nearly dry; Enphase compounded inside the 25D tax-credit era and Congress repealed it; Torus is selling into the data-center power panic while it lasts; PermitFlow monetizes municipal chaos that statutes are starting to legislate away; Pallet is racing TMS vendors to the agent layer before it becomes a checkbox. Syrup Tech shows one way the trade ends — a quiet sub-scale sale to Thoma Bravo’s Anaplan. The deliberate pairing is in energy: Enphase, the distributed-energy incumbent losing its subsidy window, and Torus, the challenger whose whole pitch is someone else’s scarcity window. Williams-Sonoma is today’s control group — the rare operator whose margins were built on discipline rather than a dislocation, which is why they survive one.

Pallet — Logistics · Emerging. The San Francisco “AI workforce” vendor sells fine-tuned agents that do order entry, quoting and track-and-trace inside brokers’ existing TMSs — $27M Series B led by General Catalyst (May 2025), $50M total. What the funding coverage skips: the last disclosed revenue figure is $3M ARR (Oct 2024), there is effectively no G2 footprint, Glassdoor sits at 2.2 with a June 2026 review alleging a toxic CEO, and its better-funded rivals — Augment ($110M) and HappyRobot ($500M valuation) — are pushing the same automation into the same mid-market brokers. Having abandoned its own TMS wedge, Pallet now visits systems of record that McLeod, Revenova and Turvo can bundle agents into natively.

Torus — Energy · Emerging. The Lehi, Utah storage company from ex-Pluralsight CPO Nate Walkingshaw bolts its Nova Spin steel flywheel in front of commodity LFP batteries and sells modular power plants into the data-center scarcity panic — ~$304M raised, including $200M from Magnetar (Sept 2025). The research turned up a pointed skeptic case the press releases don’t engage: a April 2026 CleanTechnica teardown naming Torus directly, DOE’s 2024 storage assessments excluding flywheels from grid-scale duty, ~12-minute energy duration on the flywheel itself, and a small-sample Glassdoor with “bully culture” complaints. The open question is whether the flywheel layer measurably lowers lifetime $/MWh versus Tesla and Fluence simply oversizing inverters — or is an expensive UPS welded onto a subscale BESS.

PermitFlow — Construction · Emerging. The YC W22 “TurboTax for construction permits” — Kleiner-led $31M Series A (Feb 2024), a reported ~$500M-valuation Series B, and a claimed ~10x revenue ramp since the A — sells contractors a hand-curated database of 7,000+ permitting jurisdictions plus human expediters wrapped in AI workflow. The uncomfortable findings: the software-versus-services revenue split is undisclosed, 2024-25 Glassdoor reviews describe hire-and-fire operations churn, and the complexity it monetizes is being attacked from both ends — municipal-side standardization by Accela and Cox-owned OpenGov, and instant-permitting statutes like California’s SB 379. It may be a tech-enabled expediting agency whose unit economics never escape the payroll.

Syrup Tech — Retail / Ecommerce · Emerging. The NYC AI inventory-forecasting startup — Accel-led $17.5M Series A, ~$24M total, mid-market fashion brands like Reformation and Faherty — is today’s cautionary datapoint: Anaplan, itself Thoma Bravo-owned, quietly acquired it on September 9, 2025 for undisclosed terms, and no Series B was ever raised. The page treats the exit as the analytical spine: what worked was a small team hand-tuning SKU-level demand models against each brand’s messy ERP feed, and the open question is whether that generalizes across Anaplan’s 300-plus-retailer base under PE margin discipline — or dissolves into the checkbox “AI forecasting” feature o9, Blue Yonder and Impact Analytics already claim.

Slide Insurance — Insurance · Incumbent, at risk. Bruce Lucas’s second act after Heritage: a Tampa homeowners insurer founded 2021 that built ~$1B of premium by assuming policies from Citizens depopulation and the renewal rights of dead carriers (St. Johns, UPC), then IPO’d on Nasdaq in June 2025. It is a superbly executed harvest of a dislocation that is ending — the stock broke to $12.53 by September 2025, Morgan Stanley cut it on valuation in June 2026, 2026 growth is guided to single digits, and 99%+ of its risk still sits in one hurricane state. The book has never absorbed a major landfall; the multiple assumes it will.

Infios — Supply chain · Incumbent, at risk. The Körber/KKR supply-chain-software JV — the old HighJump WMS roll-up plus MercuryGate TMS (2024), rebranded Infios in March 2025, 5,000+ customers, Gartner-estimated $500M–1B revenue — just took its eighth straight Gartner Leader placement. The at-risk call rests on what the awards don’t show: it is a mid-integration patchwork of acquired codebases squeezed between cloud-native Manhattan above and ERP-bundled SAP below, with 2025 layoffs surfacing only through Glassdoor and reviews describing the tenured engineers who hold the roll-up together walking out. Tier-2 WMS stickiness is real; it is also exactly the position agentic re-platforming erodes first.

Williams-Sonoma — Retail · Incumbent, well positioned. The San Francisco home-goods house (Pottery Barn ~$3.0B, West Elm, B2B growing double digits toward a $2B target) runs industry-best 18%+ operating margins on a debt-free balance sheet with ecommerce built in-house since the dial-up era. The moat survived the research: the complaints are real — West Elm sits at 1.8/5 on PissedConsumer on delivery and damaged-furniture themes — and furniture tariffs plus a frozen housing cycle are genuine drags, but they are timing risks, not structural ones. The no-promotion discipline and owned-brand sourcing that produced those margins were built across cycles, not inside one.

Enphase Energy — Energy · Incumbent, at risk. The Fremont microinverter maker that T.J. Rodgers and Badri Kothandaraman rescued from near-death in 2017 rode the residential-solar boom to dominance — then Congress’s 2025 repeal of the 25D homeowner credit shrank its market ~21% and tilted it toward cost-ruled third-party ownership, where Tesla’s cheap integrated Powerwall 3 stack wins. Q1 2026 revenue was $282.9M — roughly a third of the 2022 peak run-rate — after three layoff rounds since December 2023, and Q2 guidance ($280–310M) lands July 28, two days after this page. The premium per-panel architecture is a structural mismatch with the market the subsidy repeal is creating; survivable, but no longer dominant.

Full deep dives