Changing the game of risk
In a game of chess, you castle to protect the king. Done right, it hides your king safely behind a shield of pawns and brings your more powerful rook to the center of the board. It changes the game as you swap positions and move your most vital asset out of harm's way.
"That was very much our inspiration for the company," said Bruno Felix Castillo, co-founder and president of Castle Technologies. "You should protect your company – your king. You should hedge external risks."
Castle is a New York-based startup that came out of stealth in August 2026. Founded by four best friends who met at Stanford, the company is helping businesses "hedge the unhedgeable."
Castle identifies risks for businesses based on current events: an impending tariff that could raise prices, a regional conflict that could disrupt supply chains, a new bill that could drive up costs for an entire industry. Then it approaches the businesses that stand to lose money and offers them a way to hedge that exposure.
It can be a little hard to understand in the abstract. CNBC recently covered Castle's work with Tim Arrowsmith, who runs a goat grazing business in northern California that clears dry brush to reduce wildfire risk. A legislative reinterpretation stripped out the state's longstanding wage framework for goat herders. If lawmakers don't restore the old rules, Arrowsmith's labor costs for his eight herders could roughly quadruple.
Castle built him a market in collaboration with Susquehanna and Kalshi. Arrowsmith paid a $50,000 premium for a contract that pays up to $500,000 if California hasn't fixed the wage rules by the end of September.
"It's all about locking in a future outcome," explained Arjun Pandey, co-founder and CTO. "There's some future scenarios that you can control, but there are others you can't: Is the AI bubble going to pop? Is the stock market going to crash? All of that affects your business." Bruno chimed in: "I think hedging is like paying for peace of mind."
Letting the market do the actuarial work
Even though Castle uses prediction markets and agentic client sourcing, there's something ancient about how the company operates.
When fires could ravage the densely packed, wooden cities of medieval Europe, guild coffers were created to pool funds for materials and labor after a blaze. Further back, 3,700 years ago, the Code of Hammurabi let a merchant borrow against a shipment and walk away from the debt if the cargo was lost at sea. No claim, no adjuster, just an event and a payout.
"We're trying to redefine risk management from first principles," Bruno explained. "Insurance, risk management, and hedging are grounded in history, but they've evolved surprisingly little over the past three-and-a-half centuries. We recognize that history, but we're also leaning into the tech-forward aspect and recognize this is a space ripe for disruption."
Alex Michael, co-founder and CSO, built an AI agent named Percival to scrape the Internet for discussions from disgruntled business owners: it could be a Reddit thread where wholesalers are discussing their worries about a tariff, or a forum where oil riggers are talking about a permitting delay. Percival surfaces those conversations and reports back to Castle with a filtered list of the potential clients with the strongest fit. Castle then reaches out to those businesses directly and explains the opportunity for a hedge.
"It's the wild west," Bruno said. "We've seen things from submarine cables being sabotaged in the Strait of Hormuz, to carbon credits in Ghana, to oil rigs in West Texas, Colombian presidential elections, FIFA World Cup results. We've dipped into all sorts of different fields."
The insurance industry is built on actuarial work, or risk modeling. It's a highly specialized field that generally requires large teams of researchers, mathematicians, and statisticians to reach a defensible read on the probability of an outcome. Castle's founders realized that prediction markets, aside from being an investment tool, function like risk models, too.
"What prediction markets enable is, through the wisdom of the crowd, for us to get a relatively well calibrated probability for any uncertainty," Bruno explained. There's data behind the claim: Kalshi's own research arm published findings on its market calibration and concluded that "prices behave like genuine probabilities, and increasingly so as resolution approaches."
Once Castle begins working with a client, they move into the legal work and market making. If a suitable market already exists for hedging a risk on a prediction market, they can move straight to execution.
But the markets that exist often don't line up with what corporations actually want to protect against, so Castle has to translate what is inherently an insurance problem into a markets problem, limiting basis risk for the customer while maximizing liquidity from the market makers.
For Arrowsmith, that took two and a half weeks. Castle went to Susquehanna, who established pricing and contract terms with Kalshi. The contract spelled out every way California might resolve the wage question. If the state does none of them by Oct. 1, 2026, Arrowsmith gets paid.
"A lot of the work at that stage is a back and forth on designing a contract that is best suited to target the risk our customer is facing," Arjun explained. "The work is designing the triggers and resolutions that make it so there's no ambiguity when it comes to a resolution."
Two quants, a computer scientist, and a hedge fund manager
Castle's four co-founders have the résumés of a team that was recruited. Bruno and CEO Lucas Cavalieri both studied math and went into quantitative trading. Arjun studied computer science and worked as an engineer. Alex runs a hedge fund on the side.
But the four of them were friends long before they were founders. They met in their first week at college, before classes even started, and spent the next four years living together under one roof, daring each other into the hardest classes and the hardest problem sets on offer. The company came later.
One of the group's first real projects together came out of a machine learning class. They built a classification model to predict how many yards NFL quarterbacks would throw for, looked for gaps between their numbers and what the sportsbooks were offering, and traded around them. "We didn't make money," Bruno laughed. "But we tried."
During their senior year, the four of them sat down together in November to make a plan. They asked each other where they wanted to be in 10 years, and what steps they would take to get there. By February, they had already started building the Castle walls. Their senior spring put things on pause, but once Bruno, Lucas, and Arjun landed in New York together, they picked up where they left off.
"It's something that happened very naturally. I think being people that are very intellectually curious, always working on a hard problem and just seeing very much eye to eye – if we have similar visions for the world, why don't we make that a reality together?" Bruno said.
Putting a price on uncertainty
Castle wants to be the first company businesses call when they identify a risk they don't know what to do with. The past six months have been about building the credibility to get there.
"When we go to a CFO and talk to them about hedging something like a change in law, they don't really believe it at first," Bruno said. "It's almost too good to be true. It's something that's never existed before."
Right now, Castle connects a client to whoever will make the market. The plan is to build its own forecasting and pricing models, put a number on the idiosyncratic risks nobody has priced before, and then start warehousing some of that risk on its own balance sheet.
Arjun wants systems that watch a company's entire risk portfolio and hedge it automatically, recommending contracts and generating new ones where none exist. He points out that most risks worth trading have no market yet.
Even though what Castle does is complex, they're still having fun with it. On the Castle website, the founders write about financial and mathematical theory under the pen names of storied knights: Arthur, Mordred, Galahad, and Lancelot. They're still working out of the offices of their investment partner, Ribbit Capital, but when they get their own space, the first purchase is a round table.
Castle spends its days figuring out what a business can't afford to lose. When the question is turned back on them, the founders come up with a short list.
"As one of our best friends Nico likes to remind us, there are three things that any startup needs," Bruno said. "One, obviously, the vision and the team. Two, the people who back you and the mentors you surround yourself with. And then third, your bank.* Slash has been an integral part of the process for us, so we definitely recommend it to everyone out there."
Apply in less than 10 minutes today
Join the 10,000+ businesses already using Slash.
