Robinhood co-founder Vlad Tenev is now looking to conquer a new territory: artificial intelligence. Though Robinhood reported a robust $927 million in Q1 revenue and its stock has jumped 154% year-to-date, Tenev has been quietly co-founding a new company, Harmonic AI, which has already begun attracting investor buzz.
Although he does not take part in the day-to-day activities of this project, his influence has led to the company raising a new $100 million in Series B funding, placing its valuation at $875 million within less than two years. Top-tier investors such as Kleiner Perkins, Sequoia Capital, Paradigm, and Index Ventures are all supporting the venture.
Harmonic’s Vision: Mathematical Superintelligence Over Language Models
Harmonic AI isn’t trying to outdo OpenAI’s ChatGPT or Google’s Gemini. Instead, it’s aiming to solve a very different problem, AI hallucinations, where systems generate false or misleading information with high confidence. The company’s focus is on developing Mathematical Superintelligence (MSI), a different path from language-based models. While others are trying to make machines talk like humans, Harmonic wants them to think and reason like mathematicians.
Embedded in the heart of Harmonic’s technology is Aristotle, a system of artificial intelligence created to compute sophisticated math problems. Unlike large language models (LLMs), however, Aristotle does not depend on formal reason, cutting down the risk of hallucinations. It’s something Tenev thinks will prove to be vital in making AI systems both auditable and trustworthy.
Practical Applications and Industry Impact
Harmonic’s math-first methodology can transform domains of precision dependence, engineering, theoretical physics, finance, blockchain, and aerospace, and many more. As Aristotle does not rely on language but mathematical reasoning, it does not harbor the biases typically baked into LLMs. According to Harmonic, this enables the AI to deliver verifiably correct answers with unambiguous, interpretable chains of reasoning.
Aristotle already looks promising. With MiniF2F, a measure of solving complex mathematical problems, the accuracy of the system improved from 83% to 90% within a single month last July, proof of rapid iteration and promise. Although there has been no follow-up since, the momentum indicates that Aristotle is moving steadily towards Tenev’s ultimate goal.
Racing Toward Superintelligence
The wider AI community has long hoped to produce superintelligence, AI that is stronger than human intellectual capabilities. Harmonic’s MSI may be a first step to achieving that. Elon Musk said last year digital superintelligence might appear as early as this year or the following one, emphasizing how close to this great breakthrough the tech sector feels.
But with superintelligent AI comes cautions too. Geoffrey Hinton, a pioneer in the technology, recently estimated there’s a 10% to 20% possibility that AI will entirely replace humans, highlighting how serious these developments are.
A Math-Driven Future Without Hallucinations
For Tenev, the path forward is clear: math is the antidote to AI’s biggest flaws. His background in mathematics, before his entrepreneurial ventures, fuels his belief in the value of MSI. In a statement, he explained, “By incorporating mathematics as a primitive in our models… the output will be verifiably correct and auditable, with no hallucinations.”
With Harmonic AI, Tenev is not creating another technology firm. He’s setting the stage for a whole new type of intelligence that emphasizes reasoning over mimicry, and truth over deception.
