Itoflow $2.5M Raises the Quant Bar for Mid-Size Funds

London-based Itoflow has closed a $2.5 million pre-seed round led by Balderton Capital to scale AI agents that turn an investment team’s own research process into continuous, governed workflows. Angel investors including Cleo founder Barney Hussey-Yeo joined the round. The 2026 startup already runs three pilots, one monitoring roughly $3 billion in assets.

The money will expand engineering and quantitative research headcount while the company moves past pilots into commercial and regulatory work with hedge funds, family offices, exchanges and data providers.

Plain English Becomes Continuous Research Agents

Itoflow does not impose a canned strategy. Teams write their approach, risk limits and review criteria in ordinary language. Agents then run quantitative research, backtesting and live portfolio monitoring across equities, bonds, ETFs, commodities and digital assets, flagging material changes with supporting evidence.

Clients set which actions need human approval. Institutional users can deploy inside their own infrastructure so proprietary strategies and data never leave the firm. The agent stack was built from scratch for long-running reliability as markets shift.

• Supports equities, fixed income, ETFs, commodities and digital assets in one workflow
• Backtesting and real-time monitoring from the same plain-language mandate
• On-premises option keeps sensitive signals inside the customer environment
• Human approval gates at every threshold the team defines

The same mandate that defines a backtest also drives live monitoring. That continuity matters because a rule written once can govern discovery, validation and overnight watch without a separate tooling stack for each stage.

Co-founder and CEO Aditya Jha said the long-term goal is self-improving agents that learn which approaches hold up over months and years, spot new signals and notice when an edge has decayed.

Self-improvement here stays bounded by the client’s stated limits. Agents may surface decayed edges or fresh signals, yet the approval gates still decide what moves from flag to action.

Six to Eighteen Months of Build Work, Compressed

Jha spent nine years inside sophisticated quant shops, most recently as business head of mid-frequency trading at Tower Research Capital. Before that he worked as a quantitative researcher at JPMorgan and RBC. Even at Tower, spinning up a new team still took six to eighteen months to assemble data pipelines, research tools, backtesting, monitoring and execution infrastructure.

One thing that struck me at Tower was how long it could still take a new team to become fully productive. Depending on the strategy, it could take six to eighteen months to assemble the data, research pipelines, backtesting, monitoring and execution infrastructure they needed. And this was at one of the most sophisticated quantitative firms in the world. It made me realise that the barrier for smaller firms often isn’t a lack of investment ideas, but the infrastructure and specialist talent required before you can even place your first trades.

Aditya Jha, co-founder and CEO, Itoflow, speaking to Tech Funding News

He founded the company in 2026 with engineers Abinash Meher and Dibya Jyoti Roy, who had built large-scale systems at Apple, Google and Microsoft. The bet is that the same capability set can now be delivered as a platform rather than a multi-year internal project.

That path from large-firm experience to product design shows up in the build sequence the founders lived through before incorporation.

1. JPMorgan and RBC – quantitative research roles that framed how institutional mandates are written and tested
2. Tower Research Capital – business head of mid-frequency trading, where new teams still needed six to eighteen months of infrastructure work
3. 2026 founding – Jha joined by Meher and Roy, bringing large-scale systems work from Apple, Google and Microsoft
4. Pre-seed close – $2.5 million led by Balderton Capital, with Barney Hussey-Yeo among the angels

The compression claim is mechanical. Data pipelines, research tools, backtesting, monitoring and execution once arrived as sequential internal projects. The platform aims to present them as one governed workflow described in plain language.

Three Pilots Already Watching Billions

Itoflow is live with an asset manager, an ETF provider and a mid-size hedge fund. The largest of those three has roughly $3 billion under the platform’s monitoring. More than 260 people have signed up on the self-serve side, yet near-term focus stays on converting the enterprise pipeline rather than retail volume.

Sivesh Sukumar, a principal at Balderton Capital, framed the economics directly: the most sophisticated firms spend years and heavy capital building the systems smaller shops simply cannot match. Itoflow changes that cost curve.

$2.5 million pre-seed, first disclosed round

3 active enterprise pilots

~$3 billion AUM under monitoring in the largest pilot

260+ platform sign-ups, enterprise conversion prioritized

Team size today is three co-founders, one employee and two consultants. Most of the new capital goes to hiring researchers and engineers.

The pilot mix itself is a stress test. An asset manager, an ETF provider and a mid-size hedge fund do not share identical review cycles or risk language. Running all three on one stack checks whether plain-language mandates hold across different institutional shapes.

Self-serve sign-ups above 260 show top-of-funnel interest. Conversion priority still sits with the live enterprise pipeline, where monitoring load and approval gates face real capital.

Who Gains When Quant Infrastructure Becomes a Product

The winners are mid-size hedge funds, family offices and asset managers that have solid investment ideas but lack the headcount or budget to staff a full quant research desk. They can now run continuous discovery and risk checks without expanding the research team in lockstep with assets under management.

Larger incumbent platforms still hold advantages in data depth and execution scale. Competitors already in the field include Boosted.ai (more than $61 million raised), Danelfin for stock scoring and Composer for retail no-code strategies. Itoflow’s pitch is that it encodes the client’s own process rather than selling a ready-made one, and that it covers digital assets alongside traditional markets.

Whether compliance teams accept long-running agents that act overnight remains an open question. Jha’s background is meant to close that gap faster than pure software teams can.

For mid-size teams the alternative is still a six-to-eighteen-month internal build or a permanent gap versus larger desks. Productized infrastructure does not erase data-depth advantages at the top end. It does change who can run continuous research without matching headcount to every rise in assets.

A Market Already Past $5 Billion and Climbing Fast

The broader category is moving quickly. According to The Insight Partners the AI in asset management market at $5.99 billion in 2025 is projected to reach $39.52 billion by 2034, a 23.31 percent compound annual growth rate. Portfolio optimization holds the largest application share; conversational and agent platforms are among the fastest-growing slices.

Europe already accounts for roughly a quarter of the market, with the UK a natural base for a London startup selling into hedge funds and family offices. Balderton’s check fits a pattern of early bets on agent infrastructure; the firm has also backed related security and enterprise AI names across Balderton Capital’s European early-stage portfolio.

Similar checks are landing in other verticals. Recent rounds include similar AI funding bets on the factory floor and another $25M AI infrastructure round in secondhand fashion, showing investors treating agent stacks as horizontal plumbing rather than single-sector toys.

Portfolio optimization’s 28-32 percent share shows where budgets already concentrate. Agent platforms grow from a smaller base, yet the same buyers fund both slices when continuous monitoring attaches to existing optimization work.

Where the $2.5 Million Goes Next

Jha says the bulk of the capital hires quantitative researchers and engineers. The balance funds product acceleration plus the commercial and regulatory work required once pilots convert. The company is already talking to exchanges, brokerages and financial data providers about deeper integrations.

On X the announcement drew mostly quiet syndication from specialist feeds rather than broad retail chatter. That tracks the enterprise focus: 260 self-serve sign-ups matter less than converting the three live pilots and the pipeline behind them. Early observers note the real test arrives when an agent surfaces a material change at 2 a.m. and the human approval gate has to decide whether the workflow still holds.

Hiring against a three-co-founder core, one employee and two consultants keeps the first wave narrow. Quantitative researchers and engineers come first because pilots already watch live capital. Commercial and regulatory capacity follows as those pilots convert and exchange or data-provider integrations harden.

Approval Gates Decide What Agents May Touch

Continuous agents only clear institutional scrutiny if humans still own the thresholds. Clients define which actions need approval. That design keeps overnight monitoring from becoming unsupervised trading.

The 2 a.m. flag is the practical case. An agent can surface a material change with supporting evidence across equities, bonds, ETFs, commodities or digital assets. The gate still decides whether the mandate allows a response, a hold, or an escalation to the desk.

On-premises deployment tightens the same loop. Proprietary strategies and data stay inside the customer environment while the agent stack runs. Reliability work built from scratch for long-running sessions matters here: markets shift while the process must remain governed, not merely available.

Jha’s years at Tower, JPMorgan and RBC are offered as a bridge to compliance readers who distrust pure software claims. The open question is unchanged. Acceptance depends on whether approval gates and evidence trails satisfy risk committees once pilots move to full commercial use.

Mid-Size Desks Inherit a Shorter Path to Live Research

The cost curve Sukumar described separates idea quality from infrastructure reach. Sophisticated firms spent years and heavy capital on systems smaller shops could not match. A platform that encodes the client’s own process shortens that gap without requiring a full internal quant desk.

Winners stay the same group named above: mid-size hedge funds, family offices and asset managers with solid ideas and limited headcount. They gain continuous discovery and risk checks without growing research staff in lockstep with assets under management.

• Plain-language mandates replace months of custom pipeline assembly
• One workflow covers backtesting and live monitoring across five asset classes
• Human gates and optional on-prem keep control with the existing risk team
• Enterprise pilots, not retail sign-ups, set the near-term proof points

Larger incumbents still lead on data depth and execution scale. Itoflow’s counter is process fidelity plus digital-asset coverage beside traditional markets. The $2.5 million round funds the headcount to test that counter under live monitoring loads already near $3 billion in the largest pilot.

The Edge That Used to Take a Decade Now Ships as Software

Itoflow’s claim is straightforward. The same systematic capabilities that once required years inside a Tower Research-scale firm can now be described in plain English and run continuously under the client’s own risk rules. If the agents prove reliable under live market stress, mid-size investment teams gain a research capacity they could never staff themselves. The $2.5 million is the first public bet that this compression works.