Teardown

Daily digest · 2026-08-21

Scan #037: Four incumbents caught between category maturity and rebuilt balance sheets — four emerging companies whose whole thesis is a subsidy or a shortage that expires

Willis Towers Watson trailing MMC and Aon on organic growth 100+ bps while betting $625M of AI capex on a 30% margin story; Halliburton running #2 in a US frac market that has been deflating since 2023 and shipping ZEUS electric spreads international for the first time; US Foods watching Sysco pull further ahead on case growth while the September 2025 PFG merger collapses under antitrust; Fluor finally clearing the fixed-price tail with an $26.9B 85%-reimbursable backlog aimed at fabs, LNG and DoE nuclear — paired with Kodiak Robotics driverless in the Permian at a $2.5B SPAC mark and $185M of runway, Kettle betting a 140M-parameter wildfire model against a state-sanctioned public model coming down the pike, Aurora Solar on the wrong side of the ITC cliff with three layoffs and a founder-CEO change already priced in, and Attentive holding a peak $7B mark while Klaviyo bundles email plus SMS beneath it.

The thread today is the difference between a franchise defending itself and a franchise being redefined around it. WTW, Halliburton and US Foods are all in categories that never really shrank — insurance broking, oilfield services, foodservice distribution — and each is nonetheless losing relative position while the #1 pulls further away. Fluor is the counter-example: an incumbent that was worse-off five years ago than any of the other three, spent a full cycle taking the writedowns, and now sits with the cleanest backlog in a generation. All four emerging companies are wedges into categories where the terms of trade are changing in 2025-2026 for reasons the incumbents did not choose — the ITC ending, the wildfire market repricing, TCPA class actions compounding, and driverless economics finally clearing on a specific 21-mile Permian route.

Willis Towers Watson — Insurance · Incumbent, at-risk. The 46,900-person, 140-country #3 global broker that Yale-trained actuary Carl Hess has run since Jan 2022, born of the 1828 London commodities-brokerage Henry Willis inherited and rebuilt as a shipping specialist. The market read the DOJ’s 2021 blocking of the $30B Aon merger and the sale of Willis Re to Gallagher for $3.25B as a permanent seat-loss; five years later, MMC’s $7.75B McGriff deal and Aon’s $13B NFP deal have rolled up the middle market WTW was supposed to defend, WTW’s organic growth trailed Aon by more than 100 bps in 2025, and Hess is now betting a $625M “AI-powered efficiency” program to hit ~30% adjusted operating margin by 2028 while the stock trades down 22% YTD and the buyback dwarfs strategic reinvestment. The moat here is not the client book, it is the perception among Fortune 500 risk managers that #3 is still Big Three; that perception has a shelf life.

Halliburton — Energy · Incumbent, at-risk. The 107-year-old Duncan, Oklahoma cementer Erle P. Halliburton built into a completions franchise, still the #2 US pressure pumper by fleet count under Jeff Miller. Q2 2026 revenue was $5.71B at a 14% consolidated operating margin — respectable, and worse than it reads. The North American frac market has been oversupplied and deflating in real terms since 2023, ZEUS all-electric spreads are shipping to YPF / Vaca Muerta in Q4 2026 for the first time but each fleet cannibalizes a diesel spread already booked at higher margin, SLB is pulling ahead on integrated international projects and digital ARR, and Baker Hughes is pivoting to the LNG and industrial-energy tailwind Halliburton does not have a natural place in. The Permian rig count is late-cycle; the international backlog is growing but concentrated in short-fuse contracts. The management response is buybacks, which reveal the belief that reinvestment does not clear the cost of capital.

US Foods — Retail · Incumbent, at-risk. The #2 US broadline foodservice distributor Dave Flitman has run since Jan 2023 — ~$39.4B FY2025 revenue across 30,000 employees, 70+ DCs, and 90+ CHEF’STORE cash-and-carry stores — a 2016 IPO out of a nine-year KKR / CD&R LBO whose case-growth deficit to Sysco has now compounded for a decade. The 4.9% adjusted EBITDA margin still lags Sysco’s ~6% and Performance Food Group’s rising specialty mix. Then September 2025 happened: Sachem Head pushed a merger with Performance Food Group that fell over antitrust concerns before a definitive agreement, leaving a standalone company that has to out-execute two better-positioned peers while ordering apps quietly disintermediate the DSR — the district sales representative — the entire cost model is built around. Truck drivers, warehouse selectors, and DSRs all show up on Glassdoor as the operating pressure point; the private-label Exclusive Brands story is real but not yet enough to close the margin gap.

Fluor — Construction · Incumbent, well-positioned. The 114-year-old Swiss-immigrant California carpenter’s oil-and-gas EPC that walked itself into a 2019-2022 fixed-price near-death — Mission Solutions margin resets, Radford Army ammunition plant writedowns, the Chapter 15 in Australia — and then, under David Constable, spent an entire cycle taking the pain. The result as of Q2 2026 is the cleanest backlog in a generation: $26.9B total at ~85% reimbursable, only $119M of legacy fixed-price still on the pad, $3B of cash, a $1.4B buyback authorization, and NuScale monetized for roughly $2B across 2025-2026. The pipeline is aimed at semiconductor fabs, LNG trains, gas-fired power for data centers, and DoE nuclear cleanup — exactly the megaprojects the US is now willing to pay reimbursable-basis rates for. The one live grenade is the August 2025 A$1.07B Santos judgment on appeal in mid-2026; the base case is that Fluor is finally the business the 1912 rebar-and-carpentry origin implied it could be.

Kodiak Robotics — Logistics · Emerging. Don Burnette’s Mountain View driverless-truck company — Otto and Uber ATG DNA — that hauls proppant driverless for Atlas Energy on a 21-mile off-road Permian route and is chasing public-highway long-haul next. Went public September 2025 via merger with Ares Acquisition Corp II at a $2.5B enterprise value; ~$185M of pro forma cash against $160-170M of 2026 free-cash burn. The open question is whether a retrofit-first, OEM-agnostic stack proves out at under $2/mile Driver-as-a-Service on public roads before the money runs out — and does so before Aurora’s $1.2B war chest and end-2026 ~200-truck commercial fleet locks in the hub-to-hub interstate standard, Waabi’s simulation-first stack captures the Volvo factory OEM lane, and Torc / Daimler’s 2027 Freightliner Cascadia launch corners the captive-OEM channel. Non-obvious detail: the Atlas Energy contract is proppant, not freight — the unit economics that work at 21 miles of caliche do not automatically translate to 700 miles of I-10.

Kettle — Insurance · Emerging. The 2020 San Francisco insurtech Andrew Engler (Google) and Nathaniel Manning (Ushahidi, Obama admin) built around a 140-million-parameter deep-learning wildfire model — parametric reinsurance capacity from PartnerRe and RLI, an MGA product on California brush, and a September 2024 CEO handoff to Root reinsurance veteran Isaac Espinoza that reads as an “operator now, please” adjustment. The Palisades and Eaton fires of January 2025 were the first live test; the book is small enough that Kettle mostly survived it by not having written much. The open question is whether ~23 people convert a genuine modeling edge into durable, priced capacity before Munich Re, Swiss Re and Everest Re build the same models in-house, before Cal Poly Humboldt’s forthcoming public wildfire model becomes the state-sanctioned default under the California Sustainable Insurance Strategy, and before a single systemic fire season blows through the tiny book Kettle actually retains. The pattern to watch is not premium volume — it is whether the reinsurance panel keeps renewing.

Aurora Solar — Energy · Emerging. Christopher Hopper and Samuel Adeyemo’s residential-solar design and sales SaaS — ~$523.5M raised across four rounds through a Feb 2022 Series D at $4B led by Coatue and Iconiq, embedded inside roughly 80% of the US top-75 residential installers. Three layoffs and a founder-CEO change in two years have already been priced in; the newer thing is that the 30% Section 25D residential ITC dies December 31, 2025 under the One Big Beautiful Bill Act signed July 4, 2025, and Wood Mackenzie / SEIA forecast the US residential installer channel contracts 21% in 2026 alone with the overall market down 27% YoY in Q1 2026. Sunnova filed Chapter 11 in June 2025; SunPower did in August 2024. The open question is whether $135-$220 per user per month holds against OpenSolar’s genuinely free tier and Solargraf’s inverter-bundled zero-marginal-cost distribution through Enphase, and whether HelioScope C&I plus the just-launched Aug 5, 2026 Aurora Solar Marketplace grows fast enough to offset the residential seat churn. The likelier exit is a strategic acquisition (Autodesk, Trimble, Enphase, Bentley) at a fraction of $4B.

Attentive Mobile — Ecommerce · Emerging. Brian Long and Andrew Jones’s SMS-marketing platform for DTC brands — the two founders sold TapCommerce to Twitter, then built the highest-ROI channel in ecommerce and, in a $470M March 2021 Series E led by Coatue and Sequoia, priced themselves at $7B as if that channel would compound forever. Public Klaviyo (~$8B market cap) added SMS and now bundles email + SMS at the enterprise seat; Shopify’s own promotional messaging is beneath both. TCPA class-action exposure is compounding as a category-wide liability — a specific named suit against Attentive was hard to source in a day of searching, but the underlying dynamic is real. The open question is whether Attentive defends enterprise ecommerce SMS spend against Klaviyo’s bundle and Shopify’s incumbency without another priced round to reset the mark, or whether the market clears at the ~$2B secondary print and forces a strategic sale. The non-obvious detail is not the $7B mark itself — it is that the founders are the ones who repeatedly cite “profitability” now, three years after growth-at-all-costs was the whole thesis.

Full deep dives