Poesis to Open Its Agentic Equity Strategy to Investors in 2027
Poesis, an AI-native asset manager based in Menlo Park, California, plans to open its flagship agentic equity strategy to external investors in early 2027, offering institutions a route into a fund where AI agents, rather than human analysts, carry out the daily research and build the trades. The strategy, named AI Clarity, has traded the firm's own capital since August 2026 and targets US-listed large-cap shares.
The firm was co-founded by Alex Popa, a former partner at Capital Group, and Charles Elkan, former global head of machine learning at Goldman Sachs. Their sales pitch rests on a design Poesis calls Dual Mandate Equity: capturing equity-market gains when markets rise, while limiting losses when they fall.
How does AI Clarity actually work?
A human signs off on every trade. Poesis describes an operating model in which its chief investment officer and team set the strategy, the risk framework and the parameters, after which AI agents conduct research, validate the data feeding the models, and generate trade proposals using quantitative tools. The investment team then reviews and approves each proposal before it is executed.
The firm says the same research, forecasting and risk process is applied systematically across sectors and market cycles. A dynamic beta overlay is designed to cut market exposure in stressed conditions and raise it in stronger ones, while machine intelligence informs which stocks the strategy holds. In practice, that places AI Clarity at the more conservative end of the autonomous-investing spectrum, with agents proposing and people disposing.
Who is behind the firm?
The founders bring institutional pedigree that most AI-native entrants lack. Popa spent more than 15 years at Capital Group, where he made partner and ran The Growth Fund of America, a US large-cap fund with roughly $350bn in assets and one of the largest actively managed equity funds in the country. Before his investment career he served as an economic adviser to Romania's prime minister and as secretary of state for foreign investments, and he holds a Harvard MBA earned as a Baker Scholar after graduating first in his engineering class at Stanford.
Elkan, the firm's chief scientist, was a managing director and global head of machine learning at Goldman Sachs, and earlier led machine learning across e-commerce and cloud at Amazon, where he became the company's first Amazon Fellow. A longtime computer-science professor at UC San Diego with a PhD from Cornell, he built AI Clarity's original signals.
How does this compare with other agentic funds?
Poesis is joining a field that is filling quickly but remains largely unproven at scale. As of early 2026, KPMG found that around a quarter of asset-management and private-equity firms had AI agents running in production, with roughly two-thirds more piloting them, so most of the industry still treats autonomous investing as an experiment rather than a live mandate.
Capital is already following the thesis elsewhere. Point72 launched its AI-driven Turion fund in October 2024, and, according to Reuters, it returned close to 14% in its first three months and grew towards $1.5bn in assets. Against that backdrop, Poesis's distinguishing feature is governance rather than autonomy: where some rivals hand full decision-making to the machine, Poesis keeps a human approval on every trade. This autumn the firm ran a one-week live experiment with ASUS and NVIDIA in which autonomous agents researched, managed risk and executed trades within set mandates, though that test used separate capital and a separate system and is distinct from AI Clarity.
What should investors weigh before 2027?
The track record is the thing to watch, and at present it is short. AI Clarity has traded proprietary capital for only about three months, has yet to manage outside money, and its central promise, protecting capital in a downturn, has not been tested through a genuine bear market.
Poesis is candid about the limits in its own risk disclosure. The firm notes that its de-risking mechanism can trigger after losses and miss subsequent recoveries, that its models can rely on inaccurate data or overfit historical patterns, and that some planned capabilities are not yet implemented. For now, then, the downside protection at the heart of the Dual Mandate approach is a stated objective rather than a demonstrated result.
Why This Matters to FinanceX Readers
For allocators, AI Clarity is an early test of whether elite human pedigree plus an agentic process can deliver the one thing systematic strategies most often struggle with: reliable downside protection. The founders' track records at Capital Group and Goldman Sachs give the launch more credibility than most AI-fund pitches, but credibility is not performance, and three months of proprietary trading is not a cycle.
The questions worth holding until 2027 are concrete: how the strategy behaves in its first real drawdown, whether per-trade human approval can scale as assets grow, and how fees compare with both passive large-cap exposure and established quant managers. Watch the first external-capital disclosures for live performance and risk figures before treating the bear-market claim as settled.
