Teardown

Daily digest · 2026-08-08

Scan #028: The inputs to the physical economy, and who keeps the margin

Eight companies about the guts of the physical economy — power, warehouses, jobsite layout, grocery shelves, mispriced drivers and ecommerce plumbing — split into incumbents monetizing today's version and challengers rebuilding the inputs, including two head-to-head sector pairs in energy and construction.

Today’s eight are about the unglamorous inputs to the physical economy — the electrons, the square footage, the chalk lines on a slab, the shelf, the auto policy, the product feed — and, in each case, who actually keeps the margin on them. Four are entrenched incumbents monetizing the current version of that plumbing; four are emerging companies betting a narrower mechanism reprices it. Two of the pairs sit in the same sector on purpose: in energy, a silicon-anode materials maker (Group14) against the gas-and-geothermal fleet (Calpine) that just produced one of the most profitable PE deals ever; in construction, a layout robot (Dusty) against the mechanical contractor (EMCOR) whose scarce craft labor is the actual bottleneck. The recurring lesson is that the incumbents’ moats are real but bruised — a blocked merger, a PE roll-up on the defensive, a carbon-heavy fleet, a nosebleed multiple — and the challengers’ edges are real but conditional, each resting on a single falsifiable bet that a hard market, a factory ramp or a giant’s copycat product could erase.

Albertsons — Retail / Grocery · Incumbent, at risk. The No. 2 traditional U.S. supermarket operator — ~2,243 stores across 22 banners (Safeway, Vons, Jewel-Osco, Acme) doing ~$83.2B in FY2025 sales, still ~30% owned and board-controlled by Cerberus. The number the “scale player” framing buries: roughly $15B of net debt and a company stranded in strategic limbo after courts blocked its $24.6B sale to Kroger in December 2024 — then Albertsons turned around and sued Kroger for the breakup. Squeezed between Walmart and Aldi on price and Amazon on digital, with a retail-media arm too small to plug the margin gap, it is the clearest public test of whether a levered, mid-share grocer survives the squeeze.

Rithum — Ecommerce · Incumbent, at risk. The channel-management plumbing stitched together from CommerceHub (1997) and ChannelAdvisor (1999), merged in 2022 and rebranded in December 2023, Insight Partners-controlled since a 2020 recap that valued it near $1.9B — syndicating 40,000+ sellers’ product data, dropship orders and listings across 400+ channels and moving $50B+ of GMV a year. The tell other coverage misses: this twice-flipped PE roll-up just cut roughly a third of its staff, and its position is exactly the layer Amazon, Walmart, Shopify and BigCommerce are absorbing natively — an integration toll booth on roads its customers increasingly own.

Marshmallow — Insurance / Insurtech · Emerging. A London digital motor insurer built on a single arbitrage — the UK industry misprices drivers without a local credit and claims history, and newly arrived migrants are the largest such pool — using alternative data to quote them 15-40% below the mainstream. Founded 2017 by the Kent-Braham twins and David Goaté (all ex-Yoti), it hit a $1.25B mark in 2021, roughly doubled to just over $2B in an April 2025 Series C led by Portage, and — unusually for an insurtech — crossed into profit in 2024 (£20.3M net on ~£289M revenue, per Sacra) with 1M+ insured drivers. Its open question is whether the thin-file pricing edge is structural or a wasting asset that decays as incumbents’ models ingest the same alternative data.

Flexe — Logistics / Supply chain · Emerging. The original “Airbnb for warehouses” — a Seattle marketplace and shared-WMS layer, founded 2013, that rents brands short-term space and fulfillment across 3,000+ independently owned warehouses without a lease, serving Walmart, Staples and BJ’s. It raised ~$250M+ and minted a $1B valuation on a $119M Series D in July 2022 — right as the category cracked. The part the unicorn tag hides: a freight recession forced two brutal layoff rounds (33% in 2023, another 38% in early 2024) and there’s been no fresh mark since. Its open question is whether an asset-light model can hold a margin as Amazon commoditizes flexible fulfillment and asset-heavy 3PLs copy the network.

Group14 — Energy / Battery materials · Emerging. A silicon-carbon anode material (SCC55) engineered as a drop-in graphite replacement that can lift lithium-ion energy density up to 50% — sold not as a new battery but as a powder for the lines cell makers already run. Founded 2015 in Washington by two EnerG2 veterans, it has raised $1.1B+ (a Porsche-led $614M Series C, an SK-led $463M Series D in August 2025) plus a ~$100M DOE award. The sober frame beneath the “95% of battery production” customer claim: its flagship Moses Lake plant has slipped more than a year and furloughed workers, the center of gravity has shifted to Korea, and revenue stays undisclosed against $750M+ of signed-but-unconverted offtake. Its open question is whether demand and the ramp arrive before the cash runs out.

Calpine — Energy · Incumbent, well positioned. The largest independent power producer in the U.S. — ~79 plants and 27+ GW, overwhelmingly gas combined-cycle, plus The Geysers geothermal complex and a retail arm serving 60 TWh a year. The detail that reframes it: this is a company that filed one of the largest industrial bankruptcies in history in 2005, and whose 2018 take-private by Energy Capital Partners ($17B incl. debt) was a contrarian bet that flexible gas would stay valuable — a bet that, as AI data centers reset power demand, roughly doubled EBITDA and produced a ~4x return when Constellation closed its ~$26.6B acquisition in January 2026, by some accounts the most profitable PE deal ever by dollars. The bruise: a carbon-heavy fleet whose value now depends on load growth outrunning decarbonization.

Dusty Robotics — Construction · Emerging. The FieldPrinter — a knee-high autonomous robot that prints the coordinated BIM model straight onto a poured slab to 1/16-inch accuracy, up to 10x faster than the two-person chalk-line crew that has laid out buildings by hand for a century. Founded 2018 in Mountain View by two Savioke veterans (Tessa Lau, Philipp Herget), it has raised ~$70M — capped by a $45M Series B led by Scale Venture Partners in May 2022 — and had printed 100M+ square feet for GCs like DPR, JE Dunn and Skanska by August 2024. The uncomfortable read: no disclosed round since 2022, a narrow single-product wedge, and HP now attacking the exact same job with SitePrint. Its open question is whether layout printing becomes standard GC-owned equipment before a hardware giant and manual crews cap its price and TAM.

EMCOR Group — Construction · Incumbent, well positioned. The largest independent mechanical and electrical contractor in the U.S. — ~$17B FY2025 revenue, ~40,000 employees installing and servicing the HVAC, power, plumbing and fire-protection guts of buildings — assembled in 1994 from the wreckage of JWP Inc.’s bankruptcy. Why it matters now: its skilled-trades workforce is precisely the bottleneck the data-center build-out, semiconductor reshoring and grid electrification are starving for, driving backlog to a record $13.25B at year-end 2025 and the stock up ~135% in the year to spring 2026. The catch the momentum narrative skates past: at that valuation EME prices in near-flawless execution on ever-larger fixed-price work, and the same craft-labor scarcity that is the moat is also the constraint on how fast it can grow.

Full deep dives