Every trading platform in 2026 is racing to slap an “AI” sticker on the box. Kosta Ouzas, CEO and founder of Minotaur Trading Systems, is doing the opposite, deliberately distancing his new automated Bitcoin trading product from the AI label.
Ouzas has more than a decade in quantitative system design and built Minotaur around capital protection and risk management.
He joined TheStreet Roundtable and argued that traders handing significant capital to AI models are trusting a black box, and explained that Minotaur’s edge is a fully human-built strategy where AI is the assistant, not the decision-maker.
Related: Stormrake CEO says not owning Bitcoin is the real risk, even for retirees
From algorithm to one-stop automation
Minotaur spent three years building out its core algorithmic IP, specialized in trading Bitcoin futures and built to capture the asset’s high volatility. The flagship Minotaur Maximizer runs on roughly 15,000 lines of code and 15 different indicators, with the firm claiming a 95% win rate on Bitcoin trades.
The last two years went into the automation layer.
“We’ve actually developed our own application. The application part of what we do now is the automation. (Now they’re) able to use our own platform to automate the strategy itself, which means all the trades, all the entries, all the exits will be done automatically by the trading bots,” Ousaz said.
The vision is a single platform that combines proprietary algorithm and automated execution, what Ouzas calls “a one-stop shop for algorithmic and automated trading.”
“Not an AI company”
Ouzas made clear that this is not AI trading trading bots.
“AI trading is when you’re relying 100% on the AI to predict the direction of the market. He might be right, he might be wrong, but it’s kind of a black box at this stage. We’re more of the old-fashioned way where we’ve built a strategy from the ground up. We’ve got devs and coders that have interpreted our way of trading and written that into code,” he explained.
Minotaur uses AI like a junior analyst, not a portfolio manager. The strategy logic is human-defined and traceable. AI and machine learning sit underneath as tools for code refinement and backtesting
More news:
AI trading bots are a “black box”
AI’s propensity to hallucinate, or make things up, is well documented. OpenAI has openly admitted that for its most recent models, GPT 5, hallucinations are close to 10%.
Ouzas stated that AI’s inability to discern between good and bad information pushed them to build their product using human intelligence, leading to better outcomes and giving investors a better understanding of why trades are made.
“What you’re really getting with us is a strategy that is human built, it’s human derived, and AI has just worked alongside us to help. You’re not getting a black box. You’re able to understand the logic and the reasoning behind the strategy itself. When you’re trusting something with money, significant amount of money, you really want to understand how it works. You don’t want an AI to be essentially guessing what the market’s going to be doing,” said Ouzas.
This story was originally published by TheStreet on May 26, 2026, where it first appeared in the MARKETS section. Add TheStreet as a Preferred Source by clicking here.