# Eon Systems **Entity class:** Company **Domain:** Brain Emulation / Embodied Simulation **Doc Type:** Company Node **Maturity:** Early-stage **Collection:** [[collections/Neurotech|Neurotech]] · [[collections/Consciousness Continuity|Consciousness Continuity]] **Primary Sources:** [[research/Research Inferences Integration Report - 2026-10-23|Research Inferences Integration Report — 2026-10-23]] ## Definition **Eon Systems is a San Francisco company pursuing whole-brain emulation, which announced on 7 March 2026 that it had joined a connectome-derived fruit-fly brain model to a physically simulated fly body.** Co-founder Alexander Wissner-Gross is its public voice on the claim; senior scientist Philip Shiu authored the underlying brain model, published in *Nature* in 2024. The company states an intention to emulate a mouse brain of about 70 million neurons within two years, and human-scale emulation as a longer-term goal. ## The demonstration The brain model carries over 125,000 neurons and about 50 million synaptic connections, wired from the [[wiki/FlyWire|FlyWire]] connectome, and drives a body simulated in the MuJoCo physics environment. Eon reports behaviors including grooming and feeding. Independent commentary notes two limits: the brain model itself was already published, and the novel body-coupling component had not been released at the time of the announcement. A separate group at Tsinghua reported a comparable result the same season (FlyGM, arXiv 2602.17997, March 2026), using the adult fly connectome as the controller architecture with the same MuJoCo body. ## Why it matters to the corpus Embodied connectome emulation converts the species ladder from a metaphor into a software pipeline with three stages: a connectome, connectome-constrained dynamics, and a neuromechanical body. The register's mammalian-emulation marker asks for a connectome-derived simulation of a mouse brain region that predicts held-out neural responses above a published threshold; Eon's stated mouse program makes it the nearest claimant, and the benchmark infrastructure to score it already exists. ## Open Questions - Published methods and code for the body-coupling component. - Validation criteria for the mouse program, and whether they meet the register's marker. - Funding, team size and investors; none were located in available reporting. ## Research Inferences These entries translate the forward-looking register in [[research/Research Inferences|Research Inferences]] into ordinary wiki prose. The tier labels apply to the inference, not automatically to every factual anchor inside it. Proposed IDs above INF-0500 come from the 2026-10-23 cycle: [[research/Research Inferences Integration Report - 2026-10-23|Research Inferences Integration Report — 2026-10-23]] · [[research/Research Inferences Integration Audit - 2026-10-23|Research Inferences Integration Audit — 2026-10-23]]. - **INF-0126 — Established.** Simulations built directly from the fly connectome have predicted behavioral responses without fitting to behavior, meaning structure alone carried enough information to generate function. This is the first empirical evidence that wiring plus reasonable dynamics reproduces behavior in a real animal. - **INF-0131 — Plausible.** If a connectome-derived model reproduces an animal's behavioral repertoire, functional equivalence has been demonstrated at that animal's complexity level. Each species crossed moves the demonstration up the ladder, and the ladder has no identified rung where the argument changes character. - **INF-0141 — Analytic.** The field is climbing an explicit ladder — worm, fly, zebrafish, mouse, primate, human — and each rung is a funded program with published milestones. Reading the ladder as a schedule rather than a metaphor produces a defensible estimate of when human-scale attempts begin. - **INF-0479 — Plausible.** A connectome-derived simulation of a mouse brain region predicting held-out neural responses above a published threshold is the marker for mammalian emulation feasibility. The benchmark infrastructure to evaluate it already exists. - **INF-0509 (proposed) — Plausible.** Whole-connectome fly controllers driving simulated bodies convert the species ladder into a software pipeline: connectome, connectome-constrained dynamics, neuromechanical body. The next rung is a mouse cortical region, and the first group to publish it will hold the INF-0479 marker. Proposed in the 2026-10-23 cycle. ## Relationships - **people:** [[wiki/Alexander Wissner-Gross|Alexander Wissner-Gross]] (co-founder); Philip Shiu (senior scientist, author of the underlying model). - **uses_connectome:** [[wiki/FlyWire|FlyWire]] · [[wiki/Connectome|Connectome]]. - **parallel_result:** FlyGM at [[wiki/Tsinghua University|Tsinghua University]] (March 2026). - **dataset_context:** [[wiki/Male Drosophila CNS Quantitative Summary|Male Drosophila CNS Quantitative Summary]] (the 2026 whole-CNS connectome). - **cluster_home:** [[wiki/Whole-Brain Emulation|Whole-Brain Emulation]]. - **benchmark_adjacency:** [[wiki/ZAPBench|ZAPBench]] · [[wiki/Mouse Visual-Cortex Foundation Model|Mouse Visual-Cortex Foundation Model]] · [[wiki/MICrONS|MICrONS]]. - **ecosystem:** [[wiki/Neurotechnology Ecosystem|Neurotechnology Ecosystem]] · [[collections/Neurotech|Neurotech]]. ## Sources - [How the Eon team produced a virtual embodied fly — Eon Systems, March 2026](https://eon.systems/updates/embodied-brain-emulation) - [Startup claims first full brain emulation of a fruit fly in a simulated body — The Decoder, 2026](https://the-decoder.com/startup-claims-first-full-brain-emulation-of-a-fruit-fly-in-a-simulated-body/) - [Whole-Brain Connectomic Graph Model Enables Whole-Body Locomotion Control in Fruit Fly — arXiv 2602.17997, March 2026](https://arxiv.org/pdf/2602.17997) As of 23 October 2026.