More institutional investors are turning to artificial intelligence not just for predictive power, but for transparency, a shift that QuantumStreet AI says it anticipated years ago. “We now have $8 billion in index assets powered globally,” said Art Amador, President and Co-Founder of QuantumStreet AI.
The firm, developed out of research at the University of California, Berkeley, has built an investment platform that blends macroeconomic data, corporate fundamentals, technical indicators and real-time news flow into portfolio signals. The system now underpins roughly $8 billion in index-linked assets globally, according to Amador.
“We leverage both macro data, fundamental data, technical data, and we combine that with news in order to drive better data investment decisions,” Amador said in an interview with AInvest's Capital & Power.
The firm’s technology sits behind institutional mandates and structured products distributed through banks including HSBC, Deutsche Bank and BNP Paribas, as well as the retail-facing exchange-traded fund AIEQ, which Amador described as “the world’s first AI powered exchange traded fund.”
His pitch is straightforward: markets are driven by shifting combinations of economic fundamentals and investor sentiment, and static models struggle to keep up. QuantumStreet’s system attempts to identify those shifts dynamically.
_3461000a1776794719245.jpg?format=webp&width=700)
“Markets sometimes trade on sentiment, other times they trade on fundamentals,” Amador said. “This is the power of AI. It’s an adaptive approach… as markets change, the AI models and systems can adapt and identify where markets are headed based on recognizing historical patterns.”
The company’s approach reflects a broader industry trend. Coverage by Reuters and Bloomberg News in recent years has highlighted how asset managers are increasingly incorporating machine learning techniques to process large datasets and uncover non-linear relationships that traditional factor models may miss. Bloomberg News reported that a machine-learning model could predict about 71% of mutual-fund trading decisions, underscoring how much of active management follows patterns that AI can learn.
Those systems often rely on alternative data, including news sentiment and macro indicators, to refine forecasts. QuantumStreet says its edge lies in scale and integration. As early as 2015 and 2016, the firm was processing roughly one million news articles per day alongside structured financial data, according to Amador.
Performance claims are central to the firm’s positioning. Amador said that “98% of the assets of the $8 billion actually outperformed the benchmarks” last year, with similar trends continuing into the current year. He pointed to allocations such as gold within multi-asset strategies and outperformance versus the S&P 500 in U.S. equities as contributors.
Institutional investors have raised concerns about the transparency of machine-learning models, particularly when portfolio decisions must be explained to investment committees.
That has driven interest in so-called explainable AI, or XAI, which aims to make machine-learning outputs interpretable. In a company-authored paper, Amador wrote that “a portfolio manager cannot say to an investment committee ‘we took a $200 million short position because our AI model recommended it.’”
QuantumStreet uses attribution frameworks such as SHAP (Shapley Additive Explanations) to decompose model forecasts into underlying drivers. “We make forecasts… and what we’re able to do… is signal attribution,” he said, describing how contributions from macro, technical and fundamental factors can be quantified.
The firm argues that such transparency is not optional. “Explainability is the bridge between statistical prediction and investment judgment,” Amador wrote in the paper, noting that institutional investors must be able to “explain performance and strategy clearly to investment committees.”
QuantumStreet’s models are designed to respond to market dislocations as well. Amador cited periods of heightened geopolitical tension and negative headlines, like the Iran war, saying the system can identify historical analogues and avoid reactive decision-making. “It’s not a time to panic, it’s actually a time to potentially even invest more,” he said, referring to conditions where fundamentals remain intact despite volatility.
While the broader AI investment landscape remains competitive, Amador emphasized that implementation differences—model selection, signal weighting and data quality—can lead to divergent outcomes even among firms using similar techniques. “AI is both science and art,” he said.
For retail investors, access to the firm’s strategies remains limited primarily to the ETF wrapper and bank-distributed products. Institutional mandates continue to account for the bulk of assets. And, as financial firms weigh the trade-offs between automation and oversight, QuantumStreet’s bet is that explainability will determine which AI strategies scale.