A London startup barely three months old has secured $2.5 million to build AI agents that do what hedge funds typically pay armies of quantitative researchers to accomplish: sift through market data, generate investment ideas, and execute trades. ItoFlow announced its funding on August 25 after Balderton Capital led its pre-seed round, with backing from Barney Hussey-Yeo, who founded the personal finance app Cleo, UK Investor Magazine reported.
The timing reflects a broader rush among financial firms to deploy so-called agentic AI. Unlike chatbots that simply answer questions, these systems take actions, make decisions, and operate with varying degrees of autonomy. The question hanging over the industry is how much autonomy investors will tolerate when real money is at stake.
ItoFlow's answer: not much, at least not yet.
The platform automates research workflows but insists on human approval before executing any live trade, the company says on its website. It converts investment questions into what it describes as reproducible work records, pulling from global price feeds, fundamental data, regulatory filings, and web research. Strategies run on paper first. Each simulated decision gets logged with its rationale, risk factors, and performance metrics.
That caution extends to the architecture itself. The system employs what ItoFlow calls a "Guardian" review agent, a second AI that scrutinizes research plans before work begins and vets completed analysis before release. Vacslav Glukhov, the founding research scientist, wrote in a July blog post that the company uses formal contracts to constrain what the agents can do. "A formal, machine-checked contract is how that idea becomes something the agent must prove," he explained.
The founding team brings credibility from the kinds of institutions ItoFlow hopes to sell to. CEO Aditya Jha previously ran the mid-frequency trading desk at Tower Research Capital and worked in quantitative roles at RBC and J.P. Morgan. CTO Abinash Meher built infrastructure for Google Maps and worked on Apple's object storage systems. Founding engineer Dibya Jyoti Roy spent nine years on Microsoft's Azure team. Glukhov holds a PhD from Stanford and served as AI Research Director and Head of EMEA e-trading quant research at J.P. Morgan.
Companies House records show ItoFlow incorporated on May 29, with Jha and Sivesh Sukumar listed as directors. Jha and Meher each hold between 25 and 50 percent of the company, filings aggregated by Corpium indicate.
The startup plans to deploy the fresh capital on engineering hires, particularly in quantitative research, and on product development. It also faces the less glamorous but essential work of navigating financial regulation as it moves from pilots to live trading. Job postings on the company's careers page seek a founding full-stack engineer and a compliance lead.
ItoFlow targets hedge funds, asset managers, wealth managers, and family offices. Its platform covers US, UK, European, and Indian markets, handling stocks, ETFs, bonds, crypto, and funds. UK Investor Magazine reported the company is running pilots with hedge funds and family offices, though ItoFlow declined to identify specific clients. Discussions are underway with exchanges, brokerages, and data providers.
Competition in agentic AI for finance is intensifying. Arcesium, the post-trade technology firm spun out of D.E. Shaw, launched Arcesium Intelligence in May with more than 100 pre-configured agents. Orbit Financial Technology released an Agent Builder for institutional research in June. Citi is constructing a centralized platform called Arc, Axios reported in April.
The market opportunity seems obvious: automate tedious research tasks, compress analysis timelines, scale human expertise. But Jha strikes a note of caution rarely heard in startup pitch decks. "A quant agent that always finds an edge is not impressive; it's dangerous," he wrote in a research blog post. The comment hints at the regulatory and ethical minefields ahead for firms deploying autonomous trading systems.
Three weeks before announcing the funding, ItoFlow published BuySideBench 1.0, a public benchmark of its system. The move suggests the startup understands it will need to prove performance in a skeptical industry where track records matter more than demos.
For now, the business model remains straightforward. ItoFlow offers a free tier with $30 of monthly AI usage, a Pro plan at $50 per month with $100 of AI usage, a Max plan at $100 per month with $250 of AI usage, and enterprise options that include what the company describes as "ZDR-compliant AI endpoints." The pricing page provides few details on what enterprises actually pay. A privacy policy last updated on August 14 notes the platform uses Anthropic's large language models and that SOC 2 certification is in progress.
Jha told UK Investor Magazine the long-term goal is building self-improving investment agents. That ambition, if realized, would represent a significant leap beyond today's assisted workflows. Whether regulators, compliance departments, and chief risk officers will embrace truly autonomous trading AI remains an open question. For now, ItoFlow is making a more modest bet: that hedge funds will pay for automation as long as a human still holds the keys.
