Interview
The AI Company Simulating the Entire Economy | Simile Co-founder & CEO, Joon Sung Park
- Predicts a single simulation session priced at $100 million within two to three years, with a potential run cost of $10–20 million, while anticipating the market to warm up in one to two years and aggressive growth to begin by late 2026.
- Foresees synthetic panels surpassing the current human panel market size within three years and envisions a representation layer of society capturing global perspectives within five years.
- Projects that within a decade, every individual may possess a replicable twin capable of simulating their behavior, a capability expected to evolve into dating multiple simulated versions of oneself.
- Outlines a strategic roadmap to move beyond enterprise clients to serve broader society, potentially including the operation of a small hedge fund to align with world-modeling goals.
- Anticipates the market accelerating faster than expected, potentially compressing the timeline for raising Series A and subsequent funding rounds from two years to one year.
- Identifies AGI and perfect simulators as potential risks to stock market investability due to skewed advantages for providers with similar models.
- Emphasizes that a defensible data strategy and a data flywheel improving predictive accuracy against ground truth are fundamental requirements for AI company success.
- Notes significant capital flow into the robotics and inference/hardware chip layers while warning that neon labs lacking a clear vision for global impact may become non-viable resource projects.
- Positions simulation capabilities to act as a "GPU of intelligence units" for replicating collective societal processes, distinct from individual intelligence represented by a "CPU."