Meet the 2026 Thiel Fellows: 12 Gen Z Founders Rewriting the Playbook
From a Polish countryside coder building universal asset markets to a 15-year-old who vibe-coded his way into the Wall Street Journal: this year's class is the most globally diverse and technically radical yet.
| 12 Fellows Named | $250K Each, over 2 years | 7 Countries represented | 300+ Fellows total, since 2011 |
The Thiel Fellowship has never been subtle about its thesis: the conventional career track is broken, and the people who will build the future are the ones brave enough, or contrarian enough, to ignore it. With the 2026 cohort, Peter Thiel's foundation doubled the grant to $250,000 per fellow and unveiled a class that is smaller, sharper, and more globally distributed than any before it.
Twelve founders. Seven countries. A single conviction: that the most important companies of the next decade won't be founded in lecture halls.
"The Boomer career tracks no longer lead anywhere worth going. These Gen Z iconoclasts have chosen their own paths."
— Peter Thiel, Chairman, The Thiel FoundationThe 2026 class is notable not just for the diversity of ideas but for the depth of pre-fellowship traction already baked in. Juicebox just closed a Series B. The Antifraud Company raised $5M and its founder testified before Congress. EveryTicker's teenage creator was profiled by the Wall Street Journal before he could drive. These aren't ideas. They're companies.
- Harry O'Connor (Sentient Machines): foundation models for robotics generalisation
- Victor Boyd (Cavalla): autonomous forklifts toward hypersonic logistics
- Milan Lustig (Opt32): edge AI compilers to run inference on physical hardware
- Samuel Carvalho (Praso): wholesale commerce infrastructure for Brazil's underserved SMBs
- Aubrey Niederhoffer (Swoop): Africa's first super-app, starting with Nigeria
- Galen Mead (Standard Intelligence): models that actively learn from the internet
- Claire Wang: biologically faithful nervous system simulations for BCI
- Kyler Wang (Action): stealth AI with one of the class's deepest CVs
- Nick Dobroshinsky (EveryTicker): AI research across the full US stock universe
- Antoni Kiszka (Derpetual): leveraged markets for any asset class via DeFi
- Ishan Gupta (Juicebox): AI recruiter searching 800M+ profiles by real skills
The US remains the clear majority at seven of twelve fellows, with five continents represented across the full class. What stands out is not the US share itself, but the intentionality of the five international founders: each is building in or for their home market, rather than defaulting to San Francisco.
* Galen Mead (University of Toronto background, Chapel Hill, NC residence)
We researched each fellow's career trajectory, prior ventures, educational background, and social presence. Here's what we found.
Cavalla is on a mission to get anything anywhere in under 5 hours. Starting by building autonomous forklifts, through to developing hypersonic highways.
Boyd co-founded Cavalla in a San Francisco hackerspace in late 2023 with CEO Mohammad Nafisi. His background spans the full robotics stack — hardware, software, and systems integration — which is what makes the hypersonic logistics moonshot credible rather than decorative.
| 𝕏 | cavalla.io |
Praso is building the new infrastructure for wholesale commerce, powering procurement, credit, and workflow tools for SMBs across underserved areas in Brazil.
Carvalho saw the gap first-hand at Stone, one of Brazil's largest fintechs, where foodservice SMBs were chronically underserved by existing infrastructure. He chose to build in Recife rather than Sao Paulo deliberately — a bet that the northeast's underserved market would compound faster than a crowded startup hub.
| 𝕏 | praso.com.br |
EveryTicker is democratizing institutional-grade financial research across the entire U.S. stock market, including the thousands of smaller companies Wall Street ignores.
Dobroshinsky conceived EveryTicker in 8th grade and built it using Anthropic's Claude with around 10 lines of his own code. He scaled to 250,000+ investors before finishing sophomore year, was profiled by the Wall Street Journal, and had 11 publicly traded companies cite his research — all before turning 16. The fellowship reportedly makes him one of the youngest in the program's 15-year history.
| 𝕏 | everyticker.com |
Juicebox is building an AI recruiter that helps companies make better hiring decisions. Agents that understand real skills and move hiring from guesswork to true meritocracy.
Gupta dropped out of Dartmouth CS after just 90 days — enough time to confirm that building was more urgent than studying. He built Juicebox with co-founder David Paffenholz, went through YC S22, and hasn't looked back since.
| 𝕏 | juicebox.ai |
Derpetual is building the infrastructure to create a market for any asset, with leverage.
Kiszka wrote his first code at 10 and launched a software house at 17 with the co-founder who still works beside him today. Born in the Polish countryside and now embedded in Warsaw's Solana ecosystem via Superteam Poland, he is among the first Polish founders in the fellowship's history.
| 𝕏 | kiszka.xyz |
Opt32 is building modern compute infrastructure to put AI onboard objects in the physical world, from robots to cars and drones.
On leave from Harvard CS/Philosophy, Lustig brings an unusually deep research trail for his age: ML accelerator architecture at Stony Brook's COMPAS Lab, parsing work at Michigan's Future of Programming Lab, and an end-to-end ML compiler he built from scratch targeting a hardware-agnostic ISA (MLISA) at SBU.
| 𝕏 | lustig.dev |
Standard Intelligence is building aligned general learners, pretraining large models to actively explore and learn from the Internet.
Mead comes out of the University of Toronto and is already shipping models on Hugging Face. What sets him apart from the applied AI crowd in this class is the research purity of the bet: rather than building on top of existing foundation models, he is rethinking how pretraining itself works.
| 𝕏 | nel.ag |
Swoop is building the super app for Africa, starting with food delivery in Nigeria and expanding into financial services across the continent.
Niederhoffer dropped out of UC Berkeley's Haas School of Business to move to Nigeria, having already spent four years running businesses across Southern Africa. When Thumo launched in Eswatini in August 2025, it signed 6,000 users within two months — over 1% of the country's internet users.
| 𝕏 |
Sentient Machines is a research lab building foundational models for robotics that generalize across tasks and environments.
O'Connor dropped out of high school in Ireland to chase humanoid robotics full-time, winning two BT Young Scientist awards (the same competition that launched Stripe's Collison brothers) before landing Emergent Ventures and Magnificent Grants fellowships and signing a $250K SAFE pre-seed almost immediately after.
| 𝕏 |
The Antifraud Company is a fraud bounty hunter defending American taxpayers with AI and investigative journalism.
A Phillips Andover alum, Shieh went viral at Brown with "Bloat@Brown," a DOGE-inspired site rating admin staff by efficiency, faced disciplinary proceedings, and was found innocent. He graduated cum laude from Hamilton (Government, History), had a brief stint at Palantir, and testified before Congress on government fraud in June 2025.
| 𝕏 |
Claire is building biologically accurate simulations of entire nervous systems, starting with C. elegans. A simulated brain that researchers can communicate with lays the foundation for brain-computer interface (BCI) technology.
Her methodology is what makes the work distinctive: custom DISPIM and SCAPE microscopes, RNN emulation models, and expansion sequencing under Davy Deng, combined with open-source contributions to robotics planning at MIT's Hadfield-Menell lab. C. elegans is the entry point because its 302-neuron connectome is fully mapped, but the architecture is designed to scale. If it does, the result is a simulated nervous system researchers can interrogate without ethical constraints, a foundational tool for the entire BCI field.
| 𝕏 | clairebookworm.com |
Action is an artificial intelligence company in stealth.
Wang is perhaps the most credentialed person in the class: Stanford CS, MTS at OpenAI on agent memory, Stanford AI Lab researcher, first engineering intern at Mercor, COO of Kinetic (NFT marketplace, $10M from Jump/Sequoia/Lux), founder of Spark Teen (global youth entrepreneurship nonprofit, Techstars-backed), and CS229 Best Project winner under Andrew Ng. The stealth label on "Action" only amplifies the intrigue.
| 𝕏 | kylerywang.com |
Patterns are emerging. Like the 2024-2025 cohorts we analysed last year, this class shows consistent traits: prior traction before applying (Juicebox Series B, Praso $14.5M raised, EveryTicker's 250K users), willingness to relocate radically (Niederhoffer to Lagos, O'Connor from Cork to San Francisco, Carvalho doubling down on Recife), and a research-first depth even for applied companies (Lustig's compiler work, Wang's neuroscience research, Mead's pretraining architecture).
The $250,000 grant, doubled from last year's $100K, signals that the Foundation is serious about attracting founders who have real alternatives. This isn't a grant for people who can't raise money. It's a signal for people who are choosing a different kind of capital: patient, network-rich, unconstrained.
What's particularly striking about 2026 is the global ambition without geographic privilege. Kiszka is Polish and building for global financial markets. Carvalho is from Recife, not Sao Paulo, and deliberately building in Brazil's northeast. O'Connor dropped out of an Irish high school and is now in San Francisco. The fellowship is explicitly rejecting the idea that geography is destiny.
Already researching the full Thiel alumni network? Our enriched dataset covers 300+ fellows across 200+ fields: career trajectory, company traction, funding, and more.
| Explore on Specter → | Request Access |
The 2026 class is defined by a quiet confidence in going where others won't. Three founders attacking physical robotics at the model and hardware layer simultaneously. Two building in markets, LATAM and Africa, that most Western VCs still treat as emerging. A 15-year-old who has already generated more financial research impact than most analyst teams. A Brown dropout filing federal fraud lawsuits with AI. A Harvard CS student building ML compilers for edge deployment. And one stealth company whose founder's resume suggests it's anything but a typical enterprise SaaS play.
The fellowships that age well tend to be the ones where you look back five years later and wonder how anyone could have missed it. We'll be watching all twelve.
