Instacart Co-Founder Launches Hedge Fund Run by AI Agents
Apoorva Mehta, co-founder of Instacart, launched Abundance, a hedge fund that uses thousands of AI agents to autonomously handle portfolio functions including idea generation, research, stock selection, position sizing and trade execution. The fund raised $100 million in seed funding and began operations last year with a small team of quantitative researchers, engineers and AI specialists.
The approach represents one of the more ambitious attempts to replace human fundamental portfolio managers wholesale with AI agents and automation rather than incrementally automating individual tasks. Most asset managers have adopted automation in specific areas-signal generation, execution algorithms, risk overlays-but few have claimed to remove humans from core portfolio management decisions.
What the reporting shows and doesn't
Bloomberg provided the primary account of Abundance's launch, including the $100 million seed figure and team composition. Other outlets including GuruFocus and Economic Times published similar coverage without additional technical detail.
The available reporting does not disclose how Abundance handles operational controls that typically matter in production trading systems: data quality and lineage across multiple feeds, execution latency and integration, reproducible model training and backtesting, or real-time monitoring and kill switches. No public technical whitepaper, architecture diagrams, model names, data sources or governance specifics appear in the coverage.
Mehta has not publicly explained the fund's long-term roadmap or risk management approach in statements captured by the reporting.
What practitioners should monitor
For data scientists and ML engineers building trading systems, Abundance highlights practical questions that emerge when moving from research prototypes to live capital deployment. The case raises questions about how the fund manages data pipelines, maintains model reproducibility, and operates during market stress.
Watch for publicly disclosed performance results, regulatory filings that detail operational controls, and any technical write-ups or third-party audits describing model risk management. Also monitor filings that reveal fund structure, prime-broker relationships and whether systems operate under human oversight during volatile markets.
Follow industry reporting on whether other established quant managers attempt similar end-to-end agentic approaches or whether performance data from Abundance influences competitive responses.
The broader context
Asset management has trended toward greater automation across multiple layers for years. Generative AI and LLM systems now enable more sophisticated autonomous decision-making than earlier algorithmic approaches allowed. Abundance appears to extend that trend further into core portfolio management functions.
The deployment is notable because it moves agentic systems into live capital markets at scale. Without technical disclosure or performance data, however, readers should treat the launch as an observed deployment choice rather than evidence of specific safeguards or reproducible best practices.
Your membership also unlocks:
