For decades, venture capital has run on instinct as a relationship-driven business built on gut calls and insider access to Silicon Valley’s hottest startups. Now, in the middle of the AI boom, one firm is trying to turn startup investing into a data science.
San Francisco-based TRAC has used a proprietary AI model to generate a fresh list of early-stage startups it predicts are most likely to become unicorns, or companies valued at more than a billion dollars. And two companies from its last list have already exploded into $11 billion juggernauts.
In 2023, TRAC first revealed 30 of the startups its model identified exclusively with Business Insider and also pulled back the curtain on its methodology.
The list included two relatively unknown companies at the time that are now among the hottest startups in venture: Harvey, a legal tech startup, and Kalshi, a prediction market, both valued at $11 billion.
TRAC’s model, which the firm calls “Moneyball for venture capital,” is based on over 30 sources of both public and private data. The model starts by identifying the many startups unlikely to succeed rather than picking winners.
“We don’t look for needles in the haystack,” explained Joe Aaron, TRAC’s cofounder and managing partner. “We remove the haystack.”
TRAC has found that a company’s founders do not predict success. Instead, its algorithm prioritizes 286 top investors.
“These extraordinary investors make a profit on two-thirds of their positions and one in five of their investments returns over 10X,” Aaron explained.
Less than 2% of all startups attract these elite investors, so that eliminates over 98% of all startups from TRAC’s formula.
How accurate is the algorithm? Venture investing is typically judged after a decade or more, so it is too soon to know for sure.
The firm says the companies it identifies have a one-in-five probability of becoming a unicorn. Advances in AI make this year’s list more accurate than ever, but that is not necessarily good news for investors.
“It is easier today for our AI to identify future unicorns than ever before,” said Fred Campbell, TRAC’s managing partner. “Conversely, it is harder than ever to secure allocation in funding rounds of the fastest-growing future unicorns.”
That is because so many VCs are chasing the hottest companies. The hard part is not identifying the companies, but getting founders to accept a check.
“Competition to invest in the ‘hot’ companies is as fierce as we’ve seen in decades,” Campbell said. “TRAC is seeing investor demand exceed round sizes by 10X, meaning for every $1 the company wants to raise, investors are offering $10.”
Companies are raising so quickly that, in the weeks Business Insider was working on the list, five companies fell off because they raised new funding at valuations above $1 billion.
Before any more startups graduate, here is the updated list of 30 companies in 2026, identified by TRAC’s model, in alphabetical order.
Allara
Allara founder and CEO Rachel Blank Allara
What it does: A Virtual care platform for women’s health, covered by insurance
When it was founded: 2020
Last post-money valuation: $145 million, according to Pitchbook
Total raised: $46.99 million, according to Pitchbook
Founder and CEO: Rachel Blank
Select investors: Index Ventures, GV, Vanterra, Maggie Sellers
Anara
Naveed Janmohamed is the founder of Anara Anara
What it does: Helps researchers find, understand, and produce scientific content faster by deeply integrating AI into every part of the research workflow.
When it was founded: 2023
Last post-money valuation: $70 million
Total raised: $13 million
Founder: Naveed Janmohamed
Select investors: Pioneer Fund, Rebel Fund, Y Combinator, Preston-Werner Ventures
Blueprint
Danny Freed is the founder and CEO of Blueprint Blueprint
What it does: AI Operating System for Mental Health Clinicians
Browser Use cofounders Magnus Muller (left) and Gregor Žunič Browser Use
What it does: Enables AI agents to interact with websites. Leading open source library and infrastructure for intelligent browser automation.
When it was founded: 2024
Total raised: $17 million
Last post money valuation: Undisclosed
Cofounder and CEO: Magnus Muller
Select investors: Y Combinator, Felicis Ventures, A Capital, Nexus Venture Partners
Created By Humans
The cofounders of Created by Humans, left to right: Jen Singerman, Trip Adler and Edward Igushev Created by Humans
What it does: Develops an artificial intelligence-based rights licensing platform that connects human creators with AI companies.
When it was founded: 2024
Last post money valuation: Undisclosed
Total raised: $12 million
Cofounder and CEO: Trip Adler
Select investors: Floodgate, Giant Ventures, Daft Capital, Garry Tan
Crosby
John Sarihan and Ryan Daniels, Crosby Brandon Sampson/BranCrosbydon Sampson
What it does: A hybrid law firm that uses AI and human lawyers to review and negotiate business contracts.
When it was founded: 2024
Last post-money valuation: not disclosed
Total raised: $25.8 million, according to the company
Founders: Ryan Daniels and John Sarihan
Select investors: Sequoia Capital, Bain Capital Ventures, and Index Ventures
Draftwise
What it does: Develops a knowledge management and intelligence platform that assists clients with technology and compliance needs.
When it was founded: 2020
Last post money valuation: Undisclosed
Total raised: $25 million, according to Pitchbook
Cofounder and CEO: James Ding
Select investors: Index Ventures, Y Combinator, Soma Capital
Extropic
What it does: Develops thermodynamic computing hardware that it says is radically more energy efficient than GPUs.
When it was founded: 2022
Last post-money valuation: $50 million, according to PitchBook
Total raised: $14 million, according to PitchBook
Cofounder and CEO: Guillaume Verdon-Akzam
Select investors: E1 Ventures, Kindred Ventures, Valor Equity Partners, Weekend Fund
Exabits AI
What it does: Develops decentralized infrastructure for artificial intelligence facilities.
When it was founded: 2021
Total raised: $15 million, according to Pitchbook
Last post-money valuation: $150 million, according to Pitchbook
Cofounder and CEO: Hoansoo Lee
Notable investors: Google Accelerator, Harvard Innovation Launch Lab, Hack VC
Eyebot
Matthias Hofmann, Eyebot Eyebot
What it does: Makes self-serve kiosks that run a vision test and generate a doctor-verified glasses prescription in about 90 seconds
When it was founded: 2021
Last post-money valuation: $100 million, according to the company
Total raised: $28 million, according to the company
Cofounder and CEO: Matthias Hofmann
Select investors: AlleyCorp, Baukunst, General Catalyst, National Science Foundation, Village Global
Fathom
What it does: Builds AI that turns clinicians’ notes into the standardized billing and reimbursement codes insurers require, cutting down the paperwork that slows down payments
When it was founded: 2015
Last post-money valuation: $45 million, according to PitchBook
Total raised: $61 million, according to the company
What it does: Sends a trained professional to your home or office to draw your blood and bring it to a lab for testing, so you don’t have to go to a clinic
When it was founded: 2018
Last post-money valuation: $68 million, according to PitchBook
What it does: Uses AI to make camera images sharper and cleaner by correcting flaws from the camera’s lens and sensor, so devices like phones and drones can produce much higher-quality images
When it was founded: 2019
Last post-money valuation: $100 million, according to PitchBook
What it does: Uses AI to automatically keep a small business’s books up to date, tracking income and expenses and producing tax-ready financial reports
When it was founded: 2021
Last post-money valuation: $90 million in 2024; Kick declined to share its latest valuation
Total raised: $20 million, according to the company
Founder and CEO: Conrad Wadowski
Select investors: Felicis, General Catalyst, GV OpenAI Startup Fund
Knowunity
What it does: Develops interactive learning software to help students with their everyday school life.
What it does: Send autonomous drones to 911 calls in order to assist first responders.
When it was founded: 2018
Last post-money valuation: not disclosed
Total raised: Roughly $15 million, according to the company.
Founder: Divyaditya Shrivastava
Select investors: Y Combinator, Khosla Ventures, and Toyota Ventures
Patlytics
Paul Lee and Arthur Jen, Patlytics Patlytics
What it does: Patlytics builds AI software that helps companies and law firms draft, analyze, and manage patents much faster by automating complex legal work.
What it does: Polymath Robotics builds software, so heavy-duty vehicles in industries like mining and farming can drive themselves safely.
When it was founded: 2021
Last post-money valuation: Undisclosed
Total raised: Undisclosed
Founders: Stefan Seltz-Axmacher (CEO), Ilia Baranov (CTO)
Select investors: TRAC.vc, Thursday Ventures, Y Combinator, SOMA Ventures
Rocketlane
Deepak Bala, Srikrishnan Ganesan, and Vignesh Girishankar, Rocketlane Rocketlane
What it does: A tool that helps companies run client projects smoothly by keeping all the tasks, updates, and paperwork in one place and automating the busywork.
When it was founded: 2020
Last post-money valuation: not disclosed
Total raised: $45 million
Founders: Deepak Bala, Srikrishnan Ganesan, Vignesh Girishankar
