MarkosWebWeb intelligence for the agentic economy

New startup ideas · Fintech · Fintech

startup idea

Bookform

A foundation model of the limit order book, licensed to funds as a simulator.

Bookform trains a sequence model on full-depth order book messages across US equities, futures and options venues, so a fund can ask what happens to the book if it works a 400,000 share order over two hours in a given regime, and get a distribution instead of a backtest.

3/5

venture judge

27

similar startups, last 2 years (49 all-time)

96%

of 4 nearest real companies still alive

yes

8 matching federal grants and programs

Direction supported by government programs and grants

Test it before you build it

$900 · 4 weeks · 20 prospects

For $900 and four weeks, prove that mid-size systematic equity desks will pay $9,500 for an outside market-impact benchmark against their own internal curve, and that exchange data terms leave room for software margins.

Riskiest assumption · A head of execution who already runs an internal impact curve and already pays for the underlying market data will pay software prices to an outside vendor to benchmark and beat that curve, before any model exists.

1Focus group: who and where

The head of execution or senior execution quant at a systematic equity manager running $500M to $10B, who just spent a quarterly review defending a pre-trade impact curve that the realized slippage numbers keep contradicting

where to find 20 · QWAFAFEW chapter meetings in New York, Boston and Chicago; the SEC's Investment Adviser Public Disclosure database, filtering Form ADV for quantitative equity advisers in the $500M to $10B band; The Microstructure Exchange online seminar series, where execution researchers at exactly these funds present and attend

2Sell first, build later

A design-partner benchmark pilot sold before any production model exists: in six weeks from kickoff, Bookform tests its book model against the fund's internal pre-trade impact curve on a hold-out sample of the fund's own historical child orders, and delivers a written error analysis either way

the ask · $9,500 per pilot, with a $2,500 deposit invoiced up front and the balance due on delivery of the benchmark report

a real yes · A real yes is a paid $2,500 deposit against a signed scope with a dated kickoff and a named data contact; a desk saying the approach is interesting, an unpaid trial request, or an offer to look at results if you send them counts for nothing

3Small experiments

The first one attacks the riskiest assumption; each ends with a number that says whether to run the next.

  1. 1. Sell the benchmark pilot

    $600 · 15 days

    Book 20 calls with heads of execution sourced from Form ADV filtering and QWAFAFEW chapters, opening with a two-page memo showing a distribution-versus-point-estimate cost output built on IEX's free historical depth files. Pitch the paid benchmark pilot on every call: $9,500, six weeks, Bookform's model versus their internal impact curve on a hold-out of their own historical child orders.

    keep going if · 8 of 20 calls end with a booked 45-minute scoping session that their execution data owner attends

  2. 2. Price the data licenses on paper

    $100 · 10 days

    Request written non-display, derived-data and redistribution pricing from Nasdaq (TotalView-ITCH), NYSE (Integrated Feed) and Cboe, and build a unit-economics sheet for a 10-customer book at a $60,000 annual license. This attacks the card's fatal risk with email and reading, not code.

    keep going if · Written venue terms put data cost per customer at or below 30% of a $60,000 annual contract

  3. 3. Close three paid pilots

    $200 · 10 days

    Convert scoping sessions into signed pilot scopes with a dated kickoff, and invoice a $2,500 deposit against each $9,500 pilot. The founder sends the scope document within 24 hours of each scoping call and follows up until the invoice is paid or refused.

    keep going if · 3 signed pilot scopes with dated kickoffs, of which 2 pay the $2,500 deposit within 10 days

4Collect a deposit up front

Tesla took $1,000 refundable reservations for the Model 3 and $100 for the Cybertruck before building either: the deposit is the measurement, not the revenue.

$2,500

per prospect, refundable

how · An invoice for $2,500 paid by wire against a signed one-page pilot scope, because fund operations teams pay invoices routinely but will never put a card into a checkout; the head of execution signs the scope and accounts payable moves the money set up: Stripe Invoicing ↗

what it reserves · One of three design-partner slots, the pilot price locked at $9,500 against a $25,000 list price later, first choice of which venue's depth data the model covers first, and results held private to the fund for six months

refund · Refundable in full any time before the kickoff data call; after delivery the full $9,500 is credited against a first-year license if the fund converts

target · 2 paid deposits from 20 conversations within 28 days

before taking money · Do not ingest, redistribute or train on any exchange depth feed beyond the written terms of each venue's derived-data and non-display agreements, and have the Nasdaq and NYSE agreements reviewed before any customer order data enters the pipeline.

Go: build it if

8 of 20 desks book scoping sessions, 3 sign pilot scopes with dated kickoffs, at least $5,000 in deposits is collected, and written venue terms keep data cost at or below 30% of ACV

Kill: stop if

Fewer than 4 of 20 desks book a scoping session, or 0 deposits are paid after 20 conversations, or exchange derived-data terms put data cost above 50% of ACV at 10 customers

5 Scripts to run itoutreach message, landing copy, deposit terms · click to open

outreach message

Your pre-trade model tells the PM a 400,000 share order costs 32 basis points, and every quarter the realized slippage argues back. I am building Bookform, a sequence model of full-depth order flow that returns a cost distribution instead of a point estimate. Before writing production code I am running three paid benchmark pilots: $9,500, six weeks, tested against a hold-out of your own child orders. If we do not beat your curve, you keep the full error analysis. 20 minutes this week?

landing page

Your impact curve is a point estimate. The book is not. $9,500 buys a six-week benchmark: our model versus your internal curve on your own hold-out orders Reserve one of three design-partner slots with a $2,500 deposit

deposit terms

Your $2,500 deposit, invoiced against a signed scope, reserves one of three design-partner slots at the $9,500 pilot price and sets your kickoff date within 30 days. It is refundable in full any time before the kickoff data call. If you convert to an annual license after delivery, the entire $9,500 is credited against year one.

Would you run this test?

One tap. The yes-share feeds the Demand pillar of this idea's score; nobody sees who answered.

Budgets are out-of-pocket estimates for a team of one to three, US market. Size the deposit to the deal, and check the terms before taking money in a regulated line.

Scorecard

One score that balances how trendy the idea is, the demand for it and its potential for 100x, with competition measured relative to every other idea in the catalog. Recent startup trends first, government priorities second.

68

Idea Score, 0-100 · raw 42.0 x 1.61

Crowded

competition: more crowded than 85% of ideas · headwind x0.57

+5.0

government priorities, secondary (609 matching grants)

Trend

66

Is the wave forming now? 2025-26 entrants vs 2023-24, rounds since 2025, the sector's live-batch direction, the 2026 trend analyst.

  • Entrants 2025-26 vs 2023-24 (similar companies)95
  • Rounds announced 2025+ in the sector87
  • Rounds announced 2025+ matching the idea100
  • Sector direction (live batch)50
  • 2026 trend analyst0

Demand

60

Does anyone want it? YC's current RFS, companies already paid for something similar, the operator judge, founders' yes-rate in decks, readers who would run the test.

  • YC asks for it (current RFS: idea / sector)30
  • Someone already pays (similar companies, recent / all-time)100
  • Operator judge: real pain50

100x potential

67

Can it return a fund? The venture judge (double weight), market-size and moat axes, neighbours still alive, the technologist judge.

  • Venture judge50
  • Market size axis100
  • Moat axis80
  • Neighbours still alive23
  • Technologist judge100

Score = 100 x cbrt(Trend x Demand x 100x) x (1 - 0.5 x crowding) + government bonus (max 5), calibrated so the 95th-percentile idea scores 90 (order never changes). A geometric mean: a weak pillar cannot be papered over. Percentiles are among the 272 ideas in the catalog; the terms matched were foundation, model, limit, order, book, licensed, funds, simulator.

The idea in full

What
Bookform trains a sequence model on full-depth order book messages across US equities, futures and options venues, so a fund can ask what happens to the book if it works a 400,000 share order over two hours in a given regime, and get a distribution instead of a backtest. It ships as a self-serve API and research console with usage-based pricing, run inside the customer's boundary so their orders and alphas never enter the training set. The first three years are data licensing, message-level reconstruction and training; revenue follows once execution desks can show the model beats their internal impact curves.
Why now
The National Science Foundation is funding the underlying academic work directly, including a roughly $200k award for Market Microstructure and Regularized Optimal Transport, and Prodigy Research (yc S26) is already training a foundation model for quantitative finance, which shows both the science and the investor appetite arrived in the same window.
Wedge: first customer and entry point
Sell one thing first: a transaction cost analysis and impact simulator for mid-size systematic equity funds, self-serve on a sandbox of one venue's historical depth data.
Path to 100x
Global buy-side spend on market data, execution analytics and quant research tooling is well past $100B, and today it is split between raw data feeds and consultants; a model that prices market impact better than any single fund's internal curve becomes the default reference and creates the category of licensed microstructure inference. The moat is the stack of venue redistribution licenses and the reconstructed multi-venue message archive, which takes years and legal spend rather than clever code to assemble.
Ceiling
Alpha decay: if the model's impact predictions become common knowledge they stop paying for themselves, and pricing collapses to a data vendor subscription.
Closest real companies, as the generator saw them
Prodigy Research is building a general foundation model for quantitative finance and Standard Signal and KelAI run the money themselves. Bookform deliberately does not trade: it licenses execution and microstructure inference to funds that want to keep their own alpha, which is why funds will feed it order flow context they would never give a competitor.
Main risk
Exchange data redistribution terms make the licensed product economically impossible to sell at software prices, and every fund rebuilds the model on data it already pays for.

Five judges

Each judge scores every idea in the catalog with a named rubric; the venture judge decides whether a card is shown at all (4-5 is venture-grade).

  • Venture investor

    3/5

    Venue redistribution licenses are a real moat, but three years to revenue plus alpha decay collapsing pricing to a data subscription caps the outcome.

  • Bootstrapper

    1/5

    Three years of venue data licensing and message reconstruction before revenue is exactly the capital hole this lens avoids.

  • Operator

    3/5

    Execution desks pay heavily for TCA, but three years of venue redistribution licensing before revenue and internal impact curves already in place.

  • Technologist

    5/5

    Multi-venue message-level book reconstruction plus a sequence model on full depth is years of data licensing and training no fund replicates casually.

  • Risk

    1/5

    The entire product rests on venue data redistribution terms the exchanges control and can reprice, which the card itself names as fatal.

  • trends

    1/5

    Why-now rests on a $200k NSF award and one S26 comparable; a microstructure model was equally pitchable in 2023 and revenue waits three years.

Similar startups in the directory

Companies whose pitch matches most of the idea's terms (foundation, model, limit, order, book, licensed, funds, simulator): 49 all-time, 27 from the last two years. Same matching as Idea Check.

  • Invertixyc F26 · 2026 · Vertical AI agentsalive

    Energy Superintelligence, AI that operates critical infrastructure

  • QodeX Quantumalchemist Alchemist Class 41 · 2026 · AI infra and computeunchecked

    QodeX Quantum builds a platform to train and deploy quantum-native AI models for research and engineering companies - faster, cheaper, and beyond classical limits.

  • Voreliosyc F26 · 2026 · AI infra and computealive

    A foundation model for engineering physics

  • Shepherd Roboticsyc F26 · 2026 · Robotics and physical worldalive

    Robots for high skilled labor powering AI infrastructure

  • NEXI Biotech Incplugandplay PnP 2026 · 2026 · Healthcare and biounchecked

    NEXI Biotech develops in vitro neurotechnology models for improved neurological drug discovery and translation.

  • Eluspeedrun SR007 · 2026 · Developer toolsalive

    Your software, improving itself: An autonomous product, run entirely by agents

  • Contxtspeedrun SR007 · 2026 · Horizontal AI assistantsalive

    Instinct for enterprise

  • Orbitsspeedrun SR007 · 2026 · Vertical AI agentsalive

    Orbits is an AI chief of staff that runs your household end-to-end, saving parents and busy families hours every week on everything from planning the meals to repairing the roof.

  • Voundspeedrun SR007 · 2026 · Data for AIalive

    Data & Deployment Infrastructure for Physical AI.

  • Miraispeedrun SR007 · 2026 · Consumeralive

    Making Any Real Place Explorable With World Models, Starting With Real Estate.

  • Primitive Labsspeedrun SR007 · 2026 · Developer toolsalive

    Primitive Labs is building the behavioral intelligence layer for AI, modeling human behavior so companies can predict how people will respond to software and agents before they launch.

  • hardware intelligenceyc S26 · 2026 · Agent infrastructurealive

    AI tools for the chips your AI runs on

Run this as an Idea Check →

The generator's reference companies

Real companies the model named as closest when it wrote the card, with their fate. A check mark is a company the radar could verify in its directory.

Public money in this direction

US federal grants, SBIR/STTR awards and open opportunities from the radar's public-money feed, matched to the idea's terms; the sector totals give the context.

8

grants and programs matching the idea

2

startup-relevant grants in Fintech

$809K

awarded in the sector, tracked

All public money by sector →

Market signal

What the radar sees in Fintech: new companies by cohort year, the forming YC batch, and outcomes since the February snapshot.

Fintech · 256 → 148 → 186 → 125 → 99 new companies 2022 → 2026 · 89% aliveYC F26 live: 7 in this cluster, 6% of the batch (was 5% in S26)Since February, of 592 YC companies here: 11 acquired, 8 shut down, 56 rewrote their pitch

Fintech: companies, trend and grants →

Design attributes

The card is one cell of a designed set: every axis below was chosen before the text was written, and the text had to realize it.

Buyer
Enterprise
Business model
Software subscription
Path to 100x
Creates a new category
Market size
$100B+ market
Capital intensity
Capital-medium (ops, field teams)
Speed to revenue
R&D first, revenue after 3 years
Technical depth
Deep tech: ML, hardware, bio
Go-to-market
Self-serve
Moat
License or regulatory moat
Geography
US first
Regulation
Some regulation
Vibe
Hot space

Listed under

An idea sits in its own sector and in any sector its text clearly touches.

More ideas like this

Fintech · Fintech

Bookzero

Software that closes a small company's books for a tenth of a bookkeeper's price.

Bookzero is self-serve subscription software that connects to a small business's bank feeds, invoices and payroll and produces a finished monthly close - categorized ledger, reconciliations, financial statements - for roughly $100 a month against the $800-1,500 SMBs pay outsourced bookkeepers.

Score 83Warm competitionVC 5/5Software subscriptionSmall businesstest: $1.2k · 4w79% of 3 neighbours alive

Fintech · Fintech

Parbook

The self-serve secondary market where institutions trade private credit loan participations.

Parbook is a marketplace where private credit funds, insurers and regional banks list, price and settle loan participations and portfolio slices, using a standardized machine-readable data tape that Parbook generates from each seller's loan documents.

Score 72Active competitionVC 5/5MarketplaceEnterprisetest: $2k · 6w79% of 3 neighbours alive

Fintech · Fintech

Perilex

An exchange where machine-priced corporate risk is listed and capital bids.

Perilex is a listing venue for idiosyncratic corporate risk that traditional carriers price badly: fleet and telematics-driven liability, compute and outage exposure, model-failure liability.

Score 72Active competitionVC 5/5MarketplaceEnterprisetest: $1.5k · 5w96% of 4 neighbours alive

Fintech · Fintech

Countersign

Spend controls and audit trail for every dollar an enterprise's AI agents move.

Countersign sits between a company's agents and whatever rails they spend on - issued cards, wallets, stablecoin accounts - enforcing per-agent budgets, merchant allowlists and human approval thresholds before authorization, then reconciling each settled charge back to the agent run and prompt that caused it.

Score 68Crowded competitionVC 5/5Software subscriptionEnterprisetest: $700 · 3w96% of 4 neighbours alive

Fintech · Fintech

Underwatt

Underwriting software that prices AI data center risk from live power and thermal telemetry.

Underwatt sells a subscription platform to carriers, MGAs and reinsurers that prices property, business-interruption and outage risk on GPU data centers using a telemetry fleet Underwatt installs itself: power quality recorders, thermal and coolant sensors, and grid interconnect feeds at the insured sites.

Score 62Crowded competitionVC 4/5Software subscriptionEnterprisetest: $1.9k · 6w96% of 4 neighbours alive

Fintech · Fintech

Subrix

AI agents that settle claims between insurance carriers on one shared network.

Subrix runs autonomous agents that handle subrogation - the process where one insurer recovers claim costs from another at-fault party's insurer.

Score 60Crowded competitionVC 5/5AI agent as a serviceEnterprisetest: $1k · 6w82% of 4 neighbours alive
Swipe ideas like this in the deckTalk to the radar about it

Fictional company written 2026-08-22 from MarkosWeb data; the companies, grants and numbers around it are real and tracked. Treat the idea as a research prompt, not a plan.