Scan #062: three incumbents whose entrenched model has become the thing punishing them, versus four emerging companies racing a specific bundling event

Today's thread: three public incumbents each still own a real asset — MetLife has 90%+ employer-benefits persistency on top of a ~$620B asset manager, Five Below runs ~1,800 extreme-value boxes targeting a loyal tween-teen demographic, PACCAR owns the Kenworth/Peterbilt brands and a ~$5B Parts business with ~27% operating margins — and each of those assets is now the specific thing holding them back. MetLife's employer channel is what Nayya is paid by employers to disintermediate while Athene/Apollo quietly out-reinsures MetLife on pension risk transfer; Five Below's physical-box, China-sourced $1-5 model is exactly what Temu/Shein/Amazon Haul is attacking at the price-point and tariff cost structure; PACCAR's diesel-engine moat is cannibalised by a BEV/autonomy transition where it is explicitly a follower to Daimler-Volvo and Aurora-Waabi. The four emerging companies all shipped product and found customers; each is now racing a specific bundling or disintermediation event — Lily AI vs Shopify Catalog + Google Vertex AI, FourKites vs project44's integrated TMS play plus SAP agents, Jerry vs the carrier-commission squeeze and Google/Carvana embeds, Boston Metal vs Electra's lower-capex iron-ore electrolysis and a critical-metals pivot that reframes the entire company. And one refresh: D.R. Horton with FY25 numbers and 60-plus Midlands SC lawsuits that still make it the single clearest at-risk homebuilder in the group.

Lily AI — Ecommerce / Retail — Product-attribute AI · Emerging. 2015-founded Mountain View startup that turned retailer catalogs into search-and-ads-friendly attribute data — the vision+language pipeline standardizes "A-line petite midi dress, floral, cottagecore, below-the-knee" against a retailer-specific taxonomy and feeds it into on-site search, PLP ranking, Google Shopping and Meta ads. Co-founders Purva Gupta (ex-UNICEF, ClearTax) and Sowmiya Chocka Narayanan (ex-Box, Walmart Labs) have raised ~$57.5M across Series A ($12.5M Canaan, January 2020), Series B ($25M Canaan, August 2022) and a $20M Series B-1 extension led by Fernbrook Capital in March 2024 — notably not a priced Series C, which the extension structure quietly concedes. Tier-1 proof points are real: ThredUp sell-through case study, Tapestry's Google Search ads case study, M&S launch in July 2026. The uncomfortable finding is review-volume absence: G2 and Capterra show near-zero reviews for an 11-year-old enterprise SaaS, consistent with a company whose contracts are large but customer count is small. The open question: does a point enrichment vendor survive once Shopify Catalog, Google Vertex AI Search and Amazon's PXT team ship LLM catalog enrichment natively against a retailer's own product database?

FourKites — Logistics / Supply chain visibility · Emerging. Chicago-based real-time transportation visibility platform founded 2014 by Mathew Elenjickal (ex-i2/JDA/Oracle), Priyesh Ranjan, and Vivek Vaid. Raised ~$200-245M total across seed through a $100M Series D led by Thomas H. Lee Partners in March 2021 at a reported ~$1B+ valuation (SoftBank participation named in some trade press but not confirmed in primary coverage), plus a $30M growth extension in August 2022 immediately after a layoff wave and a $10M Mitsui strategic in September 2022. No priced round since; Caplight tagged FourKites in its November 2024 "zombiecorns" roundup. Four straight years as a Gartner Magic Quadrant Leader (2021-2024) and the 2025 Tracy / Sam / AI Control Tower agent launches are real product. But project44 raised at a ~$2.7B mark that FourKites never matched and now sells an integrated TMS (Movement) that bundles visibility into execution; SAP and Oracle are shipping native agents on data they already own; and Trimble's $2B Transporeon absorption plus Panasonic's $8.5B Blue Yonder reshape the competitive map. February 2024 saw a second round of layoffs and a senior-exec dismissal. The open question is whether a dedicated visibility layer survives once the TMS systems ship their own agents on their own data.

Jerry — Insurance / Consumer auto broker (MGA + aggregator) · Emerging. Palo Alto-founded (2017) consumer auto insurance comparison + vehicle super-app from Art Agrawal, Lina Zhang, and Musawir Shah. ~$240M raised across seed/A/B/C ($75M August 2021 at ~$450M led by Goodwater) and a $110M August 2023 round led by Park West Asset Management at a reported ~$1B valuation. 55+ carrier panel, 15M+ cumulative app downloads reported through 2023, and a public rebuild around a GenAI agent that shops quotes, binds policies, and manages renewals; management claimed ~$4M in annual customer-care savings from the agent rebuild. The two uncomfortable facts: Jerry has not disclosed a priced round in over two years in a category where EverQuote's revenue cliffed from ~$405M (FY22) to ~$287M (FY23) when GEICO cut broker spend 38% — EverQuote recovered to >$500M by FY24, but the whiplash is exactly the systemic carrier-commission risk Jerry's business model carries. App-store, Trustpilot and BBB reviews show a recurring complaint pattern of quote-vs-bind discrepancies and surprise billing. The open question: can an aggregator-broker win when carriers (GEICO, Progressive, State Farm), platforms (Google, Credit Karma) and point-of-sale embedders (Carvana, Rocket Auto) can each strip out the middle?

Boston Metal — Energy / Industrial decarbonisation (green primary steel) · Emerging. 2013 MIT spinout commercialising molten oxide electrolysis (MOE) to make primary steel from iron ore without coke or blast furnaces. Donald Sadoway's lab research, James Yurko's engineering, Guillaume Lambotte as CTO, Tadeu Carneiro as CEO since 2017. Capital raised exceeds ~$400M across Series A, B (BEV-led $50M January 2021 expanded by BHP), C ($120M lead January 2023 by ArcelorMittal with Microsoft Climate Innovation Fund, expanded to $282M total through 2024), a $51M July 2025 convertible, and a $75M May 2026 round. But the critical news is that the May 2026 round came AFTER a February 2026 Woburn-pilot equipment failure, 71 WARN-noticed layoffs, and a Brazil tin-slag halt — MIT Technology Review framed it as a repositioning not a growth round. Boston Metal is now explicitly leading with critical metals (niobium, tantalum, tin recovery) as the near-term revenue line, with primary green steel pushed right. Electra raised $186M in April 2025 and partnered with Nucor; H2 Green Steel/Stegra needed a €1.4bn April 2026 rescue and is seeking another ~$1.1bn; ArcelorMittal has frozen several of its own hydrogen-DRI projects. The 45V/IRA rollback hits hydrogen-DRI rivals harder than MOE, which doesn't need H2 — a weak silver lining. The open question is whether MOE economics work at commodity-steel margins without a carbon price, and whether critical metals can carry the company long enough to find out.

MetLife, Inc. — Insurance / US-based global life & benefits · Incumbent, at risk. ~$56B public (NYSE: MET), the US-based global life and employee benefits giant founded 1868, demutualized March 2000. CEO Michel Khalaf has run the company since May 2019; FY2025 revenue ran ~$73B, employees ~45,000, and MetLife Investment Management AUM is ~$620B. The company passes three at-risk rubric conditions and plausibly a fourth: Group Benefits organic growth tracks the category (4-6% in a 3-5% category) with margins protected by 90-95% employer-side persistency rather than product edge; the enrollment, dental-network administration, and claims adjudication delivery model is recognizable from 2015; Nayya raised $100M+ selling benefits-decision AI directly to employers (MetLife chose to invest-and-partner rather than challenge); and the structural ROE edge in pension risk transfer has been reversed by Athene/Apollo using private-credit-funded reinsurance. The December 2024 New Frontier Investor Day set 2024-2029 targets, but the margin pressure is from asset-side competitors (Apollo, KKR, Ares, Blackstone-owned life insurers) MetLife cannot match on cost of capital. CFO John McCallion since 2019 keeps capital returning; the strategic question is whether the next decade is the asset manager growing or the insurance channels eroding.

Five Below, Inc. — Retail / Extreme-value specialty (tweens & teens) · Incumbent, at risk. ~$10B public (NASDAQ: FIVE) — down from a ~$23B 2021 peak — the ~1,800-store, $1-5 extreme-value chain founded 2002 in Philadelphia by David Schlessinger and Tom Vellios. Winnie Park took over as CEO in November 2024 after Starboard Value pushed Joel Anderson out over shrink and comp-sale decay; Vellios is Executive Chair, Kristy Chipman CFO. FY2024 ran ~$3.9B revenue and -2.7% comparable sales — the first real negative comp in Five Below's listed life. The rubric trips on four conditions: (1) a 22-year delivery model structurally unchanged; (2) Temu/Shein/Amazon Haul as named funded challengers attacking the exact sub-$5 impulse basket from DTC-China directly; (3) the Five Beyond expansion up to $25 as a public admission that $1-5 can't carry growth any more; (4) a technology shift (cross-border DTC) the ~1,800-box physical-store economics discourage. Trump 2.0 tariffs on China sourcing are a direct hit to COGS, and locked-up high-shrink SKUs have become a documented customer-experience complaint. The strategic challenge is whether treasure-hunt-with-a-door remains differentiated once Temu and Amazon Haul match the price with next-day delivery.

PACCAR Inc. — Logistics / Commercial-truck OEM (Kenworth, Peterbilt, DAF) · Incumbent, at risk. ~$55B public (NASDAQ: PCAR), the 1905-founded Seattle-origin global commercial truck maker behind Kenworth, Peterbilt, DAF (Europe), Leyland Trucks (UK), plus PACCAR Parts and PACCAR Financial Services. CEO Preston Feight since July 2019 (previously ran DAF); Harrie Schippers CFO; the Pigott family holds voting control through 90+ years of stewardship — PACCAR has not had an unprofitable year since 1939. FY2024 revenue ~$33B, with Parts' ~27% operating margin masking a Truck-segment cycle compression from ~16-18% peak toward ~10-12%. The at-risk case: five of seven rubric triggers hit — Truck-segment growth down 2+ years off peak into a cyclically-recovering freight TAM; margins increasingly carried by the Parts aftermarket (lock-in on 1.5M+ trucks on the road, not product edge in new sales); diesel-truck delivery model essentially unchanged for decades; named challengers in both powertrain (Tesla Semi, Daimler eCascadia, Volvo VNR Electric, Chinese imports via Mexico) and autonomy (Aurora, Waabi, Kodiak, Plus); and the structural technology shift (BEV + autonomy) that cannibalises PACCAR's exact high-margin pools. The hedge is real — PACCAR partners with Aurora on autonomous Peterbilt/Kenworth, joined the Amplify Cell JV with Daimler Truck and Cummins Accelera for US battery manufacturing — but on every lane of the transition PACCAR is a follower not a leader.

Refreshed

D.R. Horton, Inc. (refreshed with FY25 numbers and Oct 2026 data) — Construction / Entry-level homebuilding · Incumbent, at risk. ~$36B public (NYSE: DHI), the largest US homebuilder by volume (~84K FY2025 closings, ~$34B revenue, ~$369K average sales price). Founded by Donald R. Horton in Fort Worth in 1978 (Horton died May 17, 2024, aged 74). Paul Romanowski took over as CEO in October 2023; David Auld is Executive Chair. DHI hits five of seven at-risk rubric triggers: two years of unit-order declines into a growing new-build category; margins propped by DHI Mortgage rate buydowns rather than product edge (buydowns have cost ~500bps of gross margin through 2024-25); a 48-year unchanged stick-build delivery model; a diffuse named-challenger field (ICON Technology's Lennar partnership at Wolf Ranch, Smith Douglas Homes' January 2024 IPO, LGI Homes and Dream Finders in entry-level); and a cottage industry of specific, documented construction-defect complaints — the Gray Media July 2024 "House from Hell" investigation and 60-plus Midlands South Carolina lawsuits are not opinion but public record. Capital allocation is still disciplined (Forestar 63%-owned, >$4B/yr in buybacks, 632,900-lot pipeline), but the moat is land, not the house.