Exaptation AI Trackers

Startup Growth Bets - Where & How New Firms Search

Exaptation AI · Trackers · Release 1

Startup Growth Bets

Where startups actually place their growth bets, as new entrants, when they expand and when they pivot, sorted by where each search starts.

Growth is a search problem. Our first tracker, Growth Search Signals, shows where leaders say to look. This one shows where startups are actually searching: the bets they make with their money, people and products. Every day, new startups launch, expand or change direction, and most of that information passes by unread. We collect a selected set of recent moves, link each one to its source, and sort it into the same four frames.

The four frames

Demand Pull

Also called Market Pull

The search starts from what customers or markets want.

Supply Push

 

The search starts from what a firm already makes or does well, pushed toward the market in its designed function.

Technology Push

 

The search starts from a capability that is new to the firm.

Resource Expansion

Also called Asset Expansion

Getting more value from what you already own or can access, by expanding what it does or finding functions it was never designed for.

Three kinds of moves

Entry

A young company entering a market with its first big bet.

Expand

The core business stays and the scope grows: new markets, products or business models.

Pivot

The core changes or is dropped.

Some moves are both. Replit expanded what its product does and pivoted its business.

Special remark. Early public information rarely shows whether an existing asset has taken on a function it was never designed for, the way Uber's network, built to move people, started delivering food. When a case is clear, we mark it with a special remark.

How to read it

Each row shows the company and its founders, the move, the bet in one sentence, a dated reason why now, the frame and the closest theory, and an ExAI note that ends with where we would start. Open any row for the source, the company's stage at the time, who is telling the story, what carried over, and what we will check later. Sort by number, company, move, frame or theory.

Release 1 holds 25 companies: 10 US unicorns from the CB Insights unicorn list (valued at $50 billion or less, and unicorns for 10 years or less), plus 15 from recent funding and pivot news, the Forbes Next Billion-Dollar Startups list and the a16z portfolio. Large scale-ups will appear in our upcoming incumbent tracker.

We are not a neutral party. We build tools for the fourth frame, so everything is in the open to disagree with and debate. We expect special remarks to stay rare, and we will report how many we find at 50 bets, whatever it shows.

ExAI notes are written by Shripal Shah with AI assistance and human review (Formed Mindfully with AI, or FMAI).

Release 1 · 25 bets · Updated Nov 17, 2026 · Next update in about two weeks · Coming next: where large incumbents search for growth

Where the search starts · 25 bets

Demand Pull312%
Supply Push1664%
Technology Push28%
Resource Expansion416%

Shares of all 25 bets. Frames are our reading of where each bet starts.

How they move · 25 bets

Entry624%
Expand1352%
Pivot520%
Expand + Pivot14%

Expand + Pivot: one company did both (Replit).

Showing 25 of 25
The BetWhy Now ExAI Notes (FMAI)
1 Abby CareHavi Nguyen (founder, CEO) Entry Trains and pays family members as certified Medicaid caregivers for relatives with complex needs, and is scaling through health-plan partnerships across states. $82M Series C led by Transformation Capital, with CVS Health Ventures, announced Oct 15, 2026, amid a home care workforce shortage. Resource Expansion Resource-based view (an untapped workforce already in the home) The workforce Abby Care needs was already in the home: family members doing unpaid care. Training and pay turn that care into a billable Medicaid service, which the company says cut reported hospitalizations 40% in year one.Where we'd start: Stakeholder-Led Innovation (experiment). Ask family caregivers what they notice at home that clinicians and health plans would pay to know.

SourceFortune, Oct 15, 2026.

Selected fromForbes Next Billion-Dollar Startups

Stage at the time of the moveSeries C ($82M); Forbes valued it at $225M in July 2026, before this round; no post-round valuation disclosed

Who is telling the storyFounder interviews; the 40% hospitalization figure is company-reported, not an independent controlled result

What carried over / what was droppedStarts from: family members already giving unpaid care at home, Medicaid home-care rules, and its own training and documentation app.

What we'll check, and whenEight states (company). Check: new states, health-plan contracts, and state moves to cut family caregiver pay in Colorado and Indiana (revisit Feb 15, 2027).

2 BiltAnkur Jain (founder, CEO) ExpandNew market From rewards on rent and mortgage payments to Neighborhood OS, software that lets restaurants, shops and sports teams run rewards and track the demand they drive. Launched Neighborhood OS on Nov 10, 2026, with Mastercard, Fiserv (Clover), Adyen and OpenTable as technology partners. Supply Push Platform theory (adding merchants as a new side) Bilt's members spend near home, and its ledger already knows where. Neighborhood OS turns that ledger from a points tracker into a way for merchants to see which rewards bring people in.Where we'd start: Data Science Experiments (experiment). Test whether rent-day and neighborhood spending patterns predict which local promotions work.

SourceBilt newsroom: Introducing Bilt Neighborhood OS, Nov 10, 2026.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since Oct 2022 (CB Insights); last priced at $10.75B (Jul 2025)

Who is telling the storyFounder interviews and company release; partner list is company-reported

What carried over / what was droppedCarried over: the rewards ledger, 50,000 merchant partners, the Bilt card, the resident base. Dropped: nothing announced; housing stays the core.

What we'll check, and whenLaunched with early partners including Feels Like Home's three Dallas restaurants (Mamani, Bar Colette, Namo), Bungalow Hospitality in New York, Altamarea Group and Dalan Real Estate; volume not disclosed. Check: merchants live and team deals (revisit Feb 15, 2027).

Timeline2021 (rent rewards) to Oct 2025 (mortgage rewards) to Nov 2026 (local commerce)

3 BitesEran Heffetz (CEO), Tal Nagler Almog, Hagai Horovitz ExpandNew market From a microlearning app for frontline staff to a platform that ties short lessons to the sales, service and operating results they are meant to move. $40M Series A announced Oct 29, 2026; frontline workers are about 80% of the global workforce, the company says. Demand Pull Lean Startup (customer need pivot) Bites already collects learning and engagement data from millions of frontline interactions. Tied to store results, that record could take on a second function: early warning on which teams or stores are slipping.Where we'd start: Data Science Experiments (experiment). Test whether engagement records predict store-level results before selling them as performance data.

SourceBites release (GlobeNewswire), Oct 29, 2026.

Selected fromFunding and pivot news

Stage at the time of the moveSeries A ($40M; $55M total raised)

Who is telling the storyCompany announcement relayed by trade press; growth and retention figures are company-reported

What carried over / what was droppedCarried over: the microlearning engine, delivery inside WhatsApp, SMS, Slack and Teams, and engagement data. Added: links to POS, CRM, ERP and HR-system data; the full training-to-results loop is described largely as next-generation work. Dropped: positioning as a learning tool.

What we'll check, and whenNet revenue retention of 161% (company release, not audited). Check: how many customers use the performance layer end to end, not just lessons (revisit Feb 15, 2027).

TimelineDec 2025 (described as microlearning) to Oct 2026 (performance enablement). Positioning evidence only.

4 ClairNico Simko (CEO), Alex Kostecki, Erich Nussbaumer ExpandNew market From wage advances embedded in payroll platforms like QuickBooks and BambooHR to Clair Connect, which puts the same advances inside banking, fintech and workforce apps. Launched Clair Connect on Oct 14, 2026, after reaching a $100M revenue run rate in August. Supply Push Embedded finance (distribution through partners) Each advance rests on verified, real-time earnings data pulled from payroll. That record is close to proof of income, which lenders, landlords and benefits programs all ask for.Where we'd start: Reusability Experiments (experiment). Test where else a verified pay record could replace paperwork, with the worker's consent.

SourceClair release: Clair Connect, Oct 14, 2026.

Selected fromForbes Next Billion-Dollar Startups

Stage at the time of the moveValued at $145M (Forbes, Jul 2026); $69M raised

Who is telling the storyCompany and partner releases; the $100M revenue run rate, cash-flow-positive status and 500,000 monthly users are company-reported

What carried over / what was droppedCarried over: payroll integrations, Pathward as lender, the advance product. Added: a way to reach users outside payroll apps.

What we'll check, and whenFirst Clair Connect partner expected 'in the coming months' (company); no live integration named yet. Check: named partners and volume (revisit Feb 15, 2027).

Timeline2019 (founded) to Oct 2026 (Clair Connect)

5 FortellMatt de Jonge (CEO), Cole Morris (COO), Andrew Casper, Igor Lovchinsky Entry Prescription AI hearing aids with a custom chip that separates speech from background noise, sold with audiologist care, now in 30+ states through new clinics and remote fittings. Expanded to 30+ states on Oct 15, 2026, through new clinics and Fortell Anywhere, its remote fitting service. Technology Push General Purpose Technology (on-device AI) The hard part Fortell solved is a chip that pulls one voice out of a noisy room in real time. That capability is not limited to hearing loss.Where we'd start: New Application Domains (cause). Test the speech-separation chip with people who hear fine but work in noise, such as call centers or factory floors.

SourceThe Hearing Review, Oct 16, 2026.

Selected fromForbes Next Billion-Dollar Startups

Stage at the time of the moveValued at $740M (Forbes, Jul 2026); $163M raised

Who is telling the storyCompany release; the peer-reviewed preference study (International Journal of Audiology, 2026) tested 20 adults listening to recordings, with company-affiliated authors; the '95% preferred' line is a separate company metric

What carried over / what was droppedStarts from: the founders' AI audio work, its own chip, and a staffed audiology team (the stated growth bottleneck).

What we'll check, and when30+ states; $6,800 a pair (may vary by package), outside insurance. Check: clinic count and share of remote fittings (revisit Feb 15, 2027).

6 GroundedSam Shapiro (founder, CEO) ExpandNew market, then New technology From modular electric camper vans for van-lifers to modular upfits for small-business and municipal fleets on any chassis: electric, hybrid or gas. Opened a 50,000 sq ft Detroit plant on Sept 8, 2026, after GM ended BrightDrop and Ford shelved its next electric Transit. Supply Push Lean Startup (customer segment pivot; platform swap forced by suppliers) The modular interior designed for living on the road now outfits clinics, cafes and service fleets. Same function, new buyers, and the buyers asked first.Where we'd start: Modularity Experiments (experiment). Hand fleet customers the module set and record combinations Grounded never designed.

SourceCity of Detroit and DEGC, Sept 8, 2026.

Selected fromFunding and pivot news

Stage at the time of the movePre-seed ($3.5M, Oct 2024) at the small-business shift; seed ($5M, Aug 2026) at the chassis shift

Who is telling the storyFounder account via independent outlets and a city release; the 90% faster upfit claim is company-reported and unaudited

What carried over / what was droppedCarried over: the modular interior system, Grounded+ vehicle software, the Detroit team, and its camper vans, which are still marketed. Added: commercial fleet builds (a box platform for electric and gas chassis; the C1 hybrid mobile command center). Dropped: EV-only builds.

What we'll check, and whenProduction underway at the new plant, ramping in 2027; reported customers include Colgate, Nokia and Wayne State University Medical. Check: delivered fleet and municipal orders (revisit Feb 15, 2027).

Timeline2023 to 2024 (small businesses ask for custom vans) to Sept 2026 (chassis-agnostic plant opens)

7 HealthleapJosiah Meyer (CEO), Jemima Meyer ExpandNew market From a clinical nutrition tool built for dietitians to a nightly screen of every adult inpatient's chart that flags malnutrition and delirium today, with three more conditions in clinical validation. Grew from 3 to 50+ hospitals in a year; $8M seed plus $30M Series A announced Oct 7, 2026. Supply Push Lean Startup (zoom-out pivot) A screen built to help dietitians now also lifts hospital revenue: Healthleap says $6.3M of its $23.8M annual impact at Penn came from added reimbursement. Whether care or finance drives renewals is the open question.Where we'd start: Stakeholder-Led Innovation (experiment). Put care, coding and finance teams in front of the same nightly scan and see which uses each keeps.

SourceTechCrunch, Oct 7, 2026.

Selected fromFunding and pivot news

Stage at the time of the movePre-seed ($1.1M, 2022) at founding; now an $8M seed (Sequoia, First Round) plus a $30M Series A (Hummingbird); valuation undisclosed

Who is telling the storyFounder account in an exclusive interview; ROI, revenue growth and Penn figures are company estimates; the published study covers malnutrition only

What carried over / what was droppedCarried over: the note-reading engine and the malnutrition model. Dropped: the dietitian-only tool.

What we'll check, and when50+ hospitals; revenue up 10x in a year (company). Check: which conditions beyond malnutrition and delirium clear validation; outpatient launch (revisit Feb 15, 2027).

Timeline2022 (dietitian tool) to 2025 or 2026 (multi-condition platform); exact date unknown

8 Hyphenate (formerly Maximor)Ramnandan Krishnamurthy (CEO), Ajay Krishna Amudan ExpandNew market From AI agents for one part of finance, under the Maximor name, to agents running five areas of the CFO's office on top of a company's existing ERP and bank systems. Renamed from Maximor on Oct 1, 2026, after 86x revenue growth; a $45M Series A led by Nexus followed on Oct 27. Demand Pull Lean Startup (zoom-out pivot) Agents that sit across order-to-cash and procure-to-pay see both sides of a company's cash. No single module was built for that joint view.Where we'd start: Interoperability Experiments (experiment). Connect the modules' records and look for uses none of them was built for.

SourceHyphenate: Series A announcement, Oct 27, 2026.

Selected fromFunding and pivot news

Stage at the time of the moveSeed (about $9M) at the rename; $45M Series A days later; about $54M raised in total

Who is telling the storyCompany announcement; 86x growth (undisclosed base) and client expansion are self-reported; a $270M valuation appears in one finance outlet, not from the company

What carried over / what was droppedCarried over: core agents, customers, team. Dropped: the Maximor name, which the company said had come to stand for one part of finance (which part is not specified).

What we'll check, and when86x revenue growth (company). Check: modules per customer and renewal rates (revisit Feb 15, 2027).

Timeline2024 (founded as Maximor) to Oct 2026 (rename)

9 Inflection AISean White (CEO); founded by Mustafa Suleyman, Karén Simonyan and Reid Hoffman PivotNew market (twice) From Pi, a consumer AI companion, to enterprise AI services after its founders left for Microsoft in 2024, and back to consumers in 2026 with Pi Journeys, a companion built around life stages. Launched Pi Journeys and Inflection AI Labs on Jul 21, 2026; research posts followed through October. Supply Push Lean Startup (customer segment pivot, then reversal) Pi's model is no longer the best on the market, but its users and its design for remembering people and relationships still are its own. Pi Journeys maps a user's relationships and life stages, which the CEO expects enterprises to want next.Where we'd start: Constraint-Driven Reframing Test (experiment). Ask what Pi can do that bigger models cannot, given that it cannot outspend them.

SourceInflection AI release (GlobeNewswire), Jul 21, 2026.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since May 2022 ($4B in the CB Insights table, a 2023 figure); valuation since the 2024 Microsoft deal undisclosed

Who is telling the storyCEO interview and company posts; the website change was observed by a third party

What carried over / what was droppedCarried over: Pi, its users and its companion design, plus agent work from the enterprise years. Dropped: building its own frontier models (2024). Enterprise: no longer promoted on its website; whether contracts continue is unconfirmed.

What we'll check, and whenPi is free with no paid tier. Check: Pi Journeys usage and any revenue model (revisit Feb 15, 2027).

Timeline2023 (Pi) to 2024 (enterprise) to Jul 2026 (back to consumers)

10 KalshiTarek Mansour (CEO), Luana Lopes Lara (COO) ExpandNew market From event contracts on elections and sports to a multi-asset exchange adding stock indexes, single-stock perpetual futures and markets on computing capacity. $1B Series G at a $40B valuation on Oct 29, 2026; Kalshi called event contracts a starting point. Supply Push Dynamic Capabilities (regulated exchange extended to new products) Kalshi's asset is a federally regulated exchange plus 30M+ accounts, not any one contract. Markets on computing capacity would price a resource that has never had a traded price.Where we'd start: Catalytic Reactors (cause). Ask which assets become tradable, and more valuable, once a regulated market exists for them.

SourceKalshi: Series G announcement, Oct 29, 2026.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since Jun 2025 ($2B); last priced at $40B (Oct 2026)

Who is telling the storyCompany announcement; volume and account figures company-reported; $65B is reported talks only

What carried over / what was droppedCarried over: the CFTC-regulated exchange, 30M+ accounts, sports and election markets. Index, perpetual-futures and compute markets are announced; not all are live.

What we'll check, and whenCompany cites more than $1T in annualized volume (an extrapolated run rate). Check: which index, perpetual and compute markets are live, and state court rulings on sports contracts (revisit Feb 15, 2027).

Timeline2018 (founded) to Oct 2026 (multi-asset plan)

11 NoahShah Ramezani (CEO), Thijn Lamers (President) PivotNew business model From a consumer 'bitcoin-backed neobank' app to stablecoin payment rails that other firms, now including Western Union, use to move money across borders. Closed a $38M seed on Oct 7, 2026; Western Union picked Noah for a planned stablecoin payout service on Oct 21. Supply Push Lean Startup (customer segment pivot, consumer to business) The licenses and bank links built for a consumer app became the product: Western Union now runs stablecoin payouts on them. Same function, new customers, and a stronger business.Where we'd start: Resource Identification (experiment). List each license and rail integration, then ask what else it permits beyond payouts.

SourceWestern Union and Noah release (GlobeNewswire), Oct 21, 2026.

Selected fromFunding and pivot news

Stage at the time of the movePre-launch or early seed at the pivot (date unknown); $38M seed now

Who is telling the storyCompany announcements relayed by trade press; market-size figures cite McKinsey

What carried over / what was droppedCarried over: licenses, banking relationships, compliance, and local payment-rail integrations. Dropped: the consumer app as the product (whether it still runs: verify).

What we'll check, and when150+ new customers in 2026 (company). Western Union's quoted reach (200+ countries and territories, 60+ currencies) describes the planned network, not live Noah coverage. Check: service launch and live volume (revisit Feb 15, 2027).

Timeline2021 (consumer app announced) to 2025 (infrastructure seed round); exact date unknown

12 Ossium HealthKevin Caldwell (CEO), Erik Woods (CSO) Entry The first bank of bone marrow recovered from deceased organ donors, now extending the same donated vertebrae to emergency preparedness and non-cancer diseases, alongside its orthopedic graft (OssiGraft, since 2023). Announced on Nov 10, 2026 the first reported organ-donor marrow transplant for a non-cancer disease, severe aplastic anemia. Resource Expansion Resource-based view (recovered donor tissue) Marrow in organ donors' vertebrae went unused until Ossium found a way to bank it. Now one donation can yield marrow for transplants and bone for orthopedic grafts (OssiGraft), and a federal contract adds radiation emergencies as a third buyer.Where we'd start: Resource Identification (experiment). Inventory every tissue in a donated vertebra and ask what else each could treat.

SourceOssium Health and Franciscan Health release, Nov 10, 2026.

Selected fromForbes Next Billion-Dollar Startups

Stage at the time of the moveValued at $300M (Forbes, Jul 2026); $158M raised

Who is telling the storyCompany releases; the 29-transplant count is company-reported (Oct 2026 poster); peer-reviewed studies cover the marrow source mostly in cancer or preclinical settings, not the new aplastic anemia case

What carried over / what was droppedStarts from: organ donation networks, cryopreservation science, and an FDA-registered facility in Indiana.

What we'll check, and when29 transplants (company); BARDA base funding $21M, with options up to $125M. Check: commercial launch and further non-cancer cases (revisit Feb 15, 2027).

13 PartifulShreya Murthy (CEO), Joy Tao ExpandNew business model From a free party-invite app to a paid events platform: Crush, a mutual-match feature (rolled out from late Dec 2025), ticketing (Jun 2026) and the Partiful Plus subscription (Oct 2026). Launched the Partiful Plus subscription on Oct 14, 2026, as Snapchat and WhatsApp added event features aimed at its market. Supply Push Lean Startup (value capture pivot) Guest lists were built to coordinate a party. Peek charges to see them, and Crush reads them as signals of who might click.Where we'd start: Spotting Emergent Patterns (cause). Watch what people already do with guest lists, then price the habits that repeat.

SourcePartiful release (PR Newswire), Oct 14, 2026.

Selected froma16z portfolio

Stage at the time of the moveSeries A ($20M, Nov 2022; $27.4M raised); valuation estimate about $140M (Sacra, 2025)

Who is telling the storyCompany releases; headcount growth from third-party data

What carried over / what was droppedCarried over: the invite and RSVP graph and the free core product. Added: payments, a subscription, mutual matching (Crush).

What we'll check, and whenFirst monetization in six years (ticketing). Check: share of hosts selling tickets, Plus uptake and Crush usage (revisit Feb 15, 2027).

Timeline2020 (free invites) to Dec 2025 (Crush) to Jun 2026 (ticketing) to Oct 2026 (subscription)

14 ReplitAmjad Masad (CEO), Haya Odeh, Faris MasadSpecial remark Expand + Pivot From a browser coding environment for learners and developers to an AI agent that turns plain-English requests into deployed apps, built on outside AI models plus Replit's decade-old run-and-deploy environment. Launched Replit for Business on Oct 26, 2026, to govern AI agent deployments inside companies. Resource Expansion Exaptation Growth Strategy Replit spent a decade building a place where code runs and ships, and collecting the record of how millions of people write and fix code. When AI models got good enough at code, those two assets took on a new function: building software from plain English. People still code by hand on Replit too.Where we'd start: Data Anomaly Analysis (experiment). Mine error and fix records for patterns that no model vendor can see.
Special remarkAn exaptation. Replit's run-and-deploy environment, built so people learning to code could write, run and ship code in one place, took on a new function once outside AI models crossed a capability threshold in 2022 to 2023: turning plain-English requests into working, deployed software. A decade of know-how about how people write and fix code fed that shift. The original function, people coding in the browser, remains.

SourceTechCrunch, Oct 17, 2026.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since Apr 2023 ($1.16B); last priced at $9B (Mar 2026); the CB Insights table still shows $3B

Who is telling the storyFounder statements; the $1B ARR target is company guidance, not achieved ARR; the onstage Anthropic remark was walked back

What carried over / what was droppedCarried over: the cloud coding environment, hosting and deployment, and the record of how people write and debug code. Dropped: education as the core market, and about half the staff before the agent bet (per AI Magazine, citing TechCrunch).

What we'll check, and whenTargeting $1B ARR by end of 2026 (company). Check: reported ARR and enterprise seats (revisit Feb 15, 2027).

Timeline2016 (founded) to 2022 (Ghostwriter) to Sept 2024 (Agent) to Oct 2026 (enterprise)

15 SkimsKim Kardashian, Jens Grede (CEO) ExpandNew market From shapewear and loungewear to a broader lifestyle brand: activewear with Nike (NikeSkims), stores abroad, and Skims Beauty, now set for 2027. On Sept 22, 2026, Kardashian moved the Skims Beauty launch to 2027. Supply Push Resource-based view (brand and customer relationship) The asset carrying each move is the audience relationship, not a product line. Beauty was tried twice before (KKW Beauty, SKKN) and closed both times; Skims now folds it under one brand.Where we'd start: Edge Case Portfolio (experiment). Study how customers already use Skims products in ways the line was not designed for.

SourceGoldman Sachs Asset Management release, Nov 2025.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since Apr 2021 (CB Insights); last priced at about $5B ($225M equity financing, Nov 2025)

Who is telling the storyFounder posts on X and company statements via trade press

What carried over / what was droppedCarried over: the brand, customer base, SKKN by Kim assets (acquired Mar 2025), Ami Colé founder as head of beauty. Dropped: SKKN as a separate brand.

What we'll check, and whenManagement expected over $1B in 2025 sales (a forecast, not a reported result). Check: beauty launch date and NikeSkims results (revisit Feb 15, 2027).

Timeline2019 (shapewear) to 2025 (Nike, beauty plans) to 2027 (beauty launch)

16 SpiritusCharles Cadieu (CEO), Matt Lee (CTO) PivotNew market From capturing CO2 from the air to bury it, to supplying that CO2 to oil producers for enhanced oil recovery. Letters of intent with three US oil producers for 3M+ tonnes a year of CO2, Sept 23, 2026, as removal buyers pulled back. Resource Expansion Lean Startup (customer segment pivot) The capture machine does the same thing; what changes is the captured CO2, from a removal credit to an oil-field input. Some peers call that a red line, so it may cost removal buyers.Where we'd start: New Application Domains (cause). Test other CO2 buyers before oil fields set the price.

SourceSpiritus release, Sept 23, 2026.

Selected fromFunding and pivot news

Stage at the time of the moveSeries A ($30M, Mar 2025) plus an earlier $11M seed

Who is telling the storyCEO interview with Bloomberg and company release; volumes and the 70M-barrel figure are company estimates; letters of intent are non-binding

What carried over / what was droppedCarried over: sorbent capture technology. Changed: its first commercial plant is now planned at an oil producer's site within two years (on-site CO2). Dropped: storage as the only buyer (whether removal sales and the Wyoming Orchard One plan continue is unconfirmed).

What we'll check, and whenThree non-binding letters of intent; no delivered tonnes or contracted revenue. Check: binding offtake contracts, plant siting and reaction from removal buyers (revisit Feb 15, 2027).

Timeline2022 (founded for capture and storage) to Sept 2026 (oil recovery deals)

17 SubstackChris Best (CEO), Hamish McKenzie, Jairaj Sethi ExpandNew business model From a subscription platform where sponsorships were arranged by individual writers to Substack's own native sponsorship formats for newsletters and podcasts, where brands pay to unlock paywalled posts or fund readers' subscriptions. Unveiled native sponsorship formats on Oct 8, 2026, and podcast sponsorships on Oct 21. Supply Push Platform theory (adding advertisers as a third side) The paywall was built to keep non-payers out. Now brands pay to open it, so the paywall itself becomes the ad unit.Where we'd start: Shift of Function (experiment). Test what else a paywall can sell, such as access gifted by employers or schools.

SourceSubstack: A new vision for sponsorships, Oct 8, 2026.

Selected fromUnicorn (company and CB Insights profile)

Stage at the time of the moveUnicorn since Jul 2025 (about $1.1B, $100M Series C); has a CB Insights company profile

Who is telling the storyCompany posts; the $200M-a-year podcast subscription figure is company-reported

What carried over / what was droppedCarried over: the paywall, subscriber relationships, the Notes feed, video and the TV app (Jan 2026). Changed: Substack now sells and runs sponsorship formats itself, for creators who opt in.

What we'll check, and whenLimited pilots with brands such as Uber and T-Mobile; dynamic podcast ads still rolling out. Check: creator opt-in, repeat brand demand and disclosure labels (revisit Feb 15, 2027).

Timeline2017 (subscriptions) to Jun 2026 (sponsorships, phase 1) to Oct 2026 (new formats, podcasts)

18 SunoMikey Shulman (CEO) PivotNew technology From AI song generation trained on unlicensed music to models rebuilt on licensed catalogs (v6), plus a new model that makes speech and music together. Launched v6, built on licensed catalogs from Warner, BMG and Believe, on Sept 9, 2026; Speech beta followed Oct 1. Supply Push Lean Startup (technology pivot forced by litigation) Settling with labels turned Suno's biggest liability, its training data, into a licensed supply chain. Speech points the same engine at podcasts and ads.Where we'd start: Need-Solution Pairs (cause). Pair the voice-and-music model with buyers who now pay separately for narration and background tracks.

SourceTechCrunch, Sept 9, 2026.

Selected fromUnicorn (company-confirmed)

Stage at the time of the moveUnicorn since Nov 2025 ($2.45B); last priced at $5.4B ($400M+ Series D, Jun 2026); not on the CB Insights unicorn list as read

Who is telling the storyCompany releases; user and revenue figures from the CEO

What carried over / what was droppedCarried over: the generation engine, 100M+ users, Songkick (bought from Warner, Nov 2025). Dropping: models trained on the old data (retirement announced).

What we'll check, and whenUniversal and Sony sued again on Sept 18, 2026 (allegations, not findings); a German first-instance court ruled for GEMA (not final). Check: v6 adoption, completion of the older-model retirement and new label deals (revisit Feb 15, 2027).

TimelineNov 2025 (Warner settlement) to Sept 2026 (v6) to Oct 2026 (Speech)

19 Superhuman (formerly Grammarly)Shishir Mehrotra (CEO) ExpandNew market From a writing assistant to a suite of AI work tools (Grammarly, Coda, Superhuman Mail, Superhuman Go), with agents that follow users across every app. One year after renaming itself from Grammarly, Superhuman posted an update on Nov 9, 2026, ahead of its first Assembly. Supply Push Dynamic Capabilities (acquisition-led expansion) Grammarly's real asset was a layer that sits in every app and text box. That layer, built to fix grammar, is now the delivery route for agents.Where we'd start: Interoperability Experiments (experiment). Map which apps the layer already sits in and test agent tasks that span two of them.

SourceGrammarly: Announcing our rebrand to Superhuman, Oct 2025.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since Oct 2019 (as Grammarly); last priced at $13B in 2021, a Grammarly-era figure, not a value for today's combined company

Who is telling the storyCompany and CEO posts; revenue ($700M+) and user figures from one secondary profile, unverified

What carried over / what was droppedCarried over: the cross-app writing layer and its users. Added: Coda (Dec 2024) and Superhuman Mail (Jul 2025). Dropped: the Grammarly company name and the 'Expert Review' feature (Mar 2026, after a lawsuit).

What we'll check, and whenCheck: Superhuman Go adoption and paid seats; the Expert Review class action (revisit Feb 15, 2027).

TimelineDec 2024 (Coda announced; closed Jan 2025) to Oct 2025 (rename) to Nov 2026

20 TrashLabNeil Chaudhary (CEO), John Tan (CTO) Entry One AI-native system that replaces the patchwork of legacy software waste haulers use for CRM, dispatch, routing, billing, payments and calls. $12M Series A led by Ventana on Oct 29, 2026; annual contract value up 7x since 2024, the company says. Demand Pull General demand pull (legacy vertical software replacement) Every route, bin and camera frame now passes through one system: over 5 million tons of material so far. That record could tell recyclers and cities what the trucks carry, not just where they went.Where we'd start: Data Anomaly Analysis (experiment). Flagged loads, like overweight or contaminated bins, may carry information buyers outside hauling want.

SourceTrashLab: About us, Nov 2026.

Selected fromFunding and pivot news

Stage at the time of the moveSeries A ($12M; confirmed on the company's About page, Nov 2026)

Who is telling the storyCompany announcement; the 7x contract value, roughly 80% Haul-E adoption and 5M+ tons are company-reported

What carried over / what was droppedStarts from: founders Neil Chaudhary (CEO, Stanford) and John Tan (CTO, ex-Google), the waste industry's legacy software, in-cab cameras, and hauler customers in 40+ US states and Canada. Competes with Hauler Hero and CurbWaste.

What we'll check, and whenHundreds of haulers; 5M+ tons managed (company). Check: share of inbound orders the agent actually handles (not just customers using it), and customers won from rivals (revisit Feb 15, 2027).

21 UltraJon Miller Schwartz (CEO), Max Friefeld (COO), Oliver Ortlieb (CTO), Chetan Parthiban Entry Stationary AI robots that pack, sort and kit e-commerce orders, rented to warehouses by the month and installed in under a day. $50M Series A led by Framework Ventures on Oct 9, 2026, with a deeper partnership with Physical Intelligence. Technology Push General Purpose Technology (AI manipulation) Each deployment also produces data on how odd-shaped items behave when packed. That record trains Ultra's own fleet today.Where we'd start: Edge Case Portfolio (experiment). Log the items the robots fail on; some point to functions beyond packing, such as grading returns.

SourceFortune, Oct 9, 2026.

Selected fromFunding and pivot news

Stage at the time of the moveSeries A ($50M, led by Framework, with YC participating); earlier $12M seed led by YC and NextView; $62M total

Who is telling the storyFounder interviews; order counts and pricing power are company-reported

What carried over / what was droppedStarts from: the founders' earlier manufacturing business with in-house fulfillment (Voodoo Manufacturing), so they knew packing pain firsthand.

What we'll check, and when500,000+ orders packed; customers in NY, GA, NJ and TX (company). Check: deployed robot count and price trend (revisit Feb 15, 2027).

22 URoboticsByoungho Yoo (CEO), Minho Oh, Donggyu Lee Entry A bolt-on autonomy kit (E-DONG-KIT) that lets legged and wheeled robots cross rough ground, taken from the lab into shipyards, steel plants and defense reconnaissance. ₩40 billion Series A co-led by Kiwoom and K2, with Hyundai's ZER01NE, announced Nov 4, 2026. Supply Push Resource-based view (core technology into adjacent markets) Shipyards, steel mills and reconnaissance all ask the kit to do what it already does: move through rough ground. The maps its robots build as they move are a by-product nobody has priced yet.Where we'd start: Forward Search Experiments (experiment). Run the kit where it wasn't built to go and record what its maps reveal about the site.

SourceEdaily (Korean), Nov 4, 2026.

Selected fromFunding and pivot news

Stage at the time of the moveSeries A (₩40B; US-dollar figures vary with the conversion used)

Who is telling the storyCompany statements relayed by trade press

What carried over / what was droppedStarts from: KAIST Urban Robotics Lab research in Daejeon (spun out Aug 2024), SLAM competition wins and a legged-locomotion patent filing; its public contact address is now in Seoul.

What we'll check, and whenValuation up more than 10x from seed (Dealroom). Named shipyard, steel or defense deployments not yet confirmed. Check: pilots announced (revisit Feb 15, 2027).

23 Utopai Studios (formerly Cybever)Cecilia Shen (CEO), Jie Yang (CTO) PivotNew business model From selling AI 3D-world tools to Hollywood to running its own AI-native film and TV studio, with its tool, PAI, as the production engine. Raised $60M at a $1B valuation, reported Oct 23, 2026; three feature films slated for theaters in 2027. Supply Push Lean Startup (value capture pivot) The tool stayed; the business model changed, from selling software to owning the films it makes. Every 3D world built for a production is an asset that outlives the shoot.Where we'd start: Reusability Experiments (experiment). Reuse each production's worlds across sequels, games and outside licensing.

SourceAI Weekly, citing The Information, Oct 23, 2026.

Selected fromFunding and pivot news

Stage at the time of the moveAt the pivot (Aug 2025): revenue about $7.5M for H1 2025 (Forbes). Now: $60M round at $1B (Oct 2026; round label varies by source). PAI licensing reached $11M ARR within 60 days of its Mar 2026 launch (Forbes).

Who is telling the storyFounder interviews; the $110M figure is the value of the project slate, not company revenue; the $3B offers are reported, not confirmed

What carried over / what was droppedCarried over: 3D world generation and video tech, rebuilt as PAI, which is still licensed (e.g., DeNA, Huashi Media). Changed: the company now leads with its own films and series; tool licensing continues alongside.

What we'll check, and when20+ projects in development; no release yet; first theatrical features planned for 2027. Check: release dates and PAI licensing revenue (revisit Feb 15, 2027).

Timeline2022 (Cybever) to Aug 26, 2025 (rename) to 2027 (first releases)

24 WhoopWill Ahmed (founder, CEO) ExpandNew market From a fitness band to a preventive-health platform: blood testing sold even to non-members, medical-grade features, and a Gulf expansion backed by Mubadala and a government clinical research program. Research collaboration with Abu Dhabi's Department of Health signed Oct 21, 2026, starting with a six-month clinical study. Supply Push Business model innovation (device to health service) Years of continuous heart, sleep and strain data from 2.5M+ members is a record few clinics hold. Abu Dhabi's health department now plans to use that data in clinical research, a function the band was never sold for.Where we'd start: Data Science Experiments (experiment). Test which member signals predict blood-test results before anyone draws blood.

SourceWAM (Emirates News Agency), Oct 21, 2026.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since Oct 2020 ($1.2B); last priced at $10.1B (Mar 2026)

Who is telling the storyCEO interviews and company releases; member and bookings figures company-reported

What carried over / what was droppedCarried over: the band, member data, the subscription model. Added: Advanced Labs (with Quest) open to non-members, Grail's cancer test, a UAE office and Arabic localization (May 2026). Dropped: nothing announced.

What we'll check, and whenGulf membership up 85% in a year (company, via Khaleej Times). No study results yet. Check: study start, consent and data terms, Labs revenue share and IPO timing (revisit Feb 15, 2027).

Timeline2015 (athlete band) to Sept 2025 (Labs) to May 2026 (Mubadala, UAE) to Oct 2026 (Abu Dhabi health department research)

25 ZiplineKeller Rinaudo Cliffton (CEO), Keenan Wyrobek; co-founders also include Ryan Oksenhorn and Will Hetzler ExpandNew market From delivering blood and medicine by drone in Rwanda to a US home-delivery network for Walmart and restaurants, adding 300 Walmart sites. $1.5B raised at $21.5B on Nov 5, 2026, with a Walmart deal for 300 more drone sites. Supply Push Lean Startup (customer segment pivot) The network built to move blood across rough terrain now moves dinners across suburbs: same function, a far bigger market. For a group of long-running restaurant partners, Zipline says its share of their delivery revenue rose from 10.8% to 34.9% in seven months.Where we'd start: Simulations with Resource Properties (experiment). Model what else a fleet that already flies over every customer could sense or carry.

SourceZipline newsroom, Nov 5, 2026.

Selected fromCB Insights unicorn list

Stage at the time of the moveUnicorn since 2019 ($1.2B); last priced at $21.5B (Nov 2026)

Who is telling the storyCompany release; restaurant figures cover a selected cohort and are company-reported

What carried over / what was droppedCarried over: aircraft, launch and landing systems, logistics software, safety record. The original company (Romotive) made consumer robots before drones.

What we'll check, and when21 Walmart sites live at announcement; 300 more planned over several years. Check: sites live and community pushback (revisit Feb 15, 2027).

Timeline2011 (Romotive) to 2016 (Rwanda) to 2023 onward (US consumer delivery)

Frames, move labels and theory labels are our reading and may change as the tracker grows. Figures are as reported by the company or in the linked coverage unless noted. Every entry links to its primary source.

We are building the AI search engine for every asset's undiscovered functions.