Teardown

Daily digest · 2026-07-16

Scan #007: Where the margin goes

Eight companies across energy storage, insurance software, logistics, construction and retail — four emerging, four entrenched — read through one question: when a shortage or a platform shift moves value around, which layer actually captures the margin?

Today’s eight are a study in where the money settles when a market is being reshaped. Three of them are challenger-and-incumbent pairs pointed at the same shift: sodium-ion upstart Peak Energy against grid-storage integrator Fluence; underwriting-AI layer Federato against core-systems owner Guidewire; digital-native fulfiller Stord against contract-logistics giant GXO. In each pair the question is the same — does the durable margin sit with the physical asset, or with the intelligence layer bolted on top? The incumbents show what that fight does to economics: Fluence neither makes the cell nor owns the customer and runs a ~13% gross margin; GXO turns double-digit revenue growth into near-zero net income; Guidewire owns the deepest moat in P&C software yet just watched its stock halve; and Staples, the outlier, is a useful distribution franchise whose cash flow is being siphoned to creditors. The four challengers are all betting the margin is migrating to them. Whether it actually does is the open question on each page.

Peak Energy — Energy / Grid storage · Emerging. A sodium-ion grid-storage startup founded in 2023 by Tesla, Fluence and Enovix veterans, raising ~$80M Series B at a ~$475M pre-money (July 2026) and building America’s first sodium-ion BESS factory in Sacramento. The pitch is passive cooling, safety and a China-free supply chain — but the number the deck skips is that its sodium cells run roughly $70/kWh against Chinese LFP near $40-45, and sodium’s lower energy density means more space per usable kWh. The whole case rests on tariffs, safety and domestic-supply value offsetting that gap before the subsidy window closes.

Federato — Insurance / Insurtech · Emerging. A “RiskOps” platform (founded 2020) that sits on top of insurers’ core systems and tells underwriters which risks to chase and how a book is drifting; it has now raised past $180M after a $100M Series D led by Goldman Sachs Alternatives (Nov 2025). The uncomfortable backdrop: rival Cytora was just absorbed by Applied Systems (a core-platform incumbent), which is exactly the commoditization risk Federato faces — the workflow layer getting swallowed by the systems beneath it or by foundation-model agents above it.

Trunk Tools — Construction · Emerging. An AI-agent platform (founded 2021 by Dr. Sarah Buchner, a carpenter-turned-PhD) that ingests a project’s drawings, specs and RFIs so field crews can query them and automate document work; ~$70M raised, capped by a $40M Series B led by Insight (July 2025). The most revealing datum is its own: a Gilbane case study logged 87% validated-correct answers over 246 questions — impressive, and also the crux, because on safety-critical drawings the gap between 87% and near-perfect is where trust and liability live, and Procore/Autodesk can bundle “good enough” on data they already host.

Stord — Logistics / Supply chain · Emerging. An ecommerce fulfillment network plus software (founded 2015, Georgia Tech) pitched as an Amazon-fulfillment alternative for independent brands; ~$775M raised, last marked at $3B (Series F, May 2026). The tension the valuation hides: Stord spent years pivoting from asset-light software to owning warehouses through a string of acquisitions (Fulfillment Works, ProPack, Pitney Bowes fulfillment, Ware2Go, Shipwire) — its 2024 profitability claim is company-stated and hard to verify, and owning the boxes is what could cap the software multiple it’s really being priced on.

Fluence Energy — Energy / Grid storage · Incumbent, at risk. The largest Western grid-scale battery-storage integrator (a 2018 Siemens/AES JV, IPO’d 2021), sitting on a record ~$5.3B backlog. But it turned its first small profit only in FY2024 and revenue fell in FY2025 at a ~13% gross margin, because it neither makes the cell nor owns the end customer — squeezed between Chinese hardware deflation and US-China tariff whiplash (Section 301 duties stepping to 25% in Jan 2026), with Siemens selling down $420M of stock in May 2026. The equity now depends on a software-and-services story still far too small to carry it.

Guidewire Software — Insurance · Incumbent, well positioned. The dominant P&C insurance core-systems platform (founded 2001; 500-plus carriers on PolicyCenter/BillingCenter/ClaimCenter), with cloud ARR now past $1B and compounding. The twist since we’d have called it overvalued: the stock round-tripped from ~$262 (Sept 2025) to ~$117 (June 2026), halving the market cap to ~$10B after ARR landed at the low end of guidance — which arguably removes the one live bear case (valuation) while leaving the moat intact. The remaining question is whether AI-workflow layers like Federato stay partners or eventually erode the core’s primacy.

GXO Logistics — Logistics / Supply chain · Incumbent, well positioned. The world’s largest pure-play contract-logistics operator (spun from XPO in 2021; ~$13.2B revenue in 2025, automation-led warehousing for blue-chip brands). It is the scaled leader in a fragmented ~$375B market with a real outsourcing tailwind — and yet the stock has gone nowhere since the spin, because the model converts 12%+ revenue growth into ~7-8% EBITDA margins and roughly $36M of net income, and a late-2024 sale process ended with the board rejecting offers and the CEO departing. Well positioned operationally; cheap for a reason.

Staples — Retail · Incumbent, at risk. The office-supply retailer Sycamore Partners took private for ~$6.9B in 2017, now mostly a B2B contract-distribution business (Staples Business Advantage) wrapped around a shrinking 900-store US fleet. The real story isn’t the stores, it’s the balance sheet: a genuinely useful delivery franchise ($700-900M EBITDA by credit-market estimates) is trapped in a leveraged structure whose 2024 refinancing carries double-digit coupons, so every year more of the franchise’s cash flow goes to creditors instead of reinvestment, in a category still in secular decline.

Full deep dives