Better Forecasts, Better Decisions
Leading Preseen's seed round
Aditya Agarwal · July 30, 2026
Every large institution in the world runs on forecasts. Almost none of them keep score.
Funds take macro positions and reinsurers price political risk in countries nobody on the desk has visited. Intelligence analysts write up elections that shape what a government does next. Real money moves on those calls, and then nobody goes back to check who was right.
Venia Veselovsky and Theo Summer are building forecasting agents for global macro events. You hand them a question with a deadline, and fifteen minutes later you get a calibrated probability that keeps updating as the world moves.
I spent my career building systems, so I want to be precise about why this is a hard technical problem and not a wrapper. A language model out of the box is a poor forecaster. It reasons from priors frozen at training time, it inherits the biases of whatever it retrieves, and it will happily give you a confident number with nothing underneath it. Getting from there to a calibrated probability means solving several problems at once. You need fresh, structured data about the state of the world, which is an ingestion and normalization problem across thousands of heterogeneous sources, not a search query. You need to force genuine independence between lines of reasoning so errors stay uncorrelated, which Preseen does by running separate agents against primary records and reconciling their estimates the way you would aggregate a panel of analysts. And you need calibration, which is its own discipline: a system that says 80% has to be right about eight times in ten, measured across hundreds of resolved questions, or the number is decoration.
The deepest part of the moat is that forecasting is one of the rare AI problems with delayed ground truth. Every question eventually resolves, so every forecast becomes a labeled training example. Preseen has been accumulating that feedback since last fall, on top of a data layer Venia spent months building: crawlers and normalization pipelines across country APIs, satellite imagery, procurement records, and stranger signals like diplomatic flight patterns and international school enrollments. Last December that machinery flagged unusual Russian flight activity before it showed up anywhere you could read. A general-purpose model gets none of this for free. The loop has to be built, and it gets stronger with every resolved question.
The results are public. Their bot won the FutureEval tournament last fall, first out of 164. It came third out of 1,283 in the Metaculus Cup this spring, ahead of some of the best human forecasters alive. Then in Q2 it won Market Pulse outright, the first time a bot has ever beaten a field of humans to win a Metaculus tournament, and it won by a wide margin. Theo, before any of this was a company, turned $35 into roughly $1.9 million on Kalshi over seven months.
This category is crowded with people insisting their wrapper can see the future. Preseen publishes its scores, and other people keep them. That is the only kind of evidence that matters here, and they have more of it than anyone.
Theo's own path is the cleanest argument for why this works now. He built a forecasting tool on GPT-4 in 2022, decided the math errors made it worthless, and threw it away. In 2025 he built the same thing again, won Metaculus's Fall tournament with it, and quit his software job. The idea never changed.
Everything is still in front of them. The product is in early access, with pilots underway at a global financial data and media company, some of the world's largest hedge funds, and a major alternative asset manager, and demand is arriving faster than a two-founder company can absorb it. Human forecasters still hold a head-to-head edge for now, and the trend lines say that edge has a year or two left. Preseen is positioned to be the company standing there when it flips, with the best public record in the field and a data engine nobody else has bothered to build.
Preseen is hiring a founding engineer and a first go-to-market person, both in San Francisco. Send people to venia@preseen.com.
Everyone in this business will tell you they are good at predicting things. Venia and Theo published the receipts, and that is why I wrote the check.