SAN FRANCISCO — A San Francisco startup's bold claim of creating the world's first 'uploaded animal' — a digital version of a fruit fly's brain controlling a virtual body — has ignited a firestorm of excitement and skepticism across social media and scientific circles.
Last week, Eon Systems, a company focused on developing 'digital human intelligence,' shared videos on X, formerly Twitter, showcasing what it described as a virtual embodied fly exhibiting behaviors like walking, eating, and rubbing its legs. The posts, dated March 8, 2026, quickly went viral, amplified by AI enthusiasts and high-profile figures. Elon Musk commented simply 'wow,' while longevity advocate Bryan Johnson called it 'this is amazing,' and futurist Peter Diamandis declared 'this is a living being…online.'
CEO Michael Andregg posted one of the clips, writing, 'We've uploaded a fruit fly. We took the @FlyWireNews connectome of the fruit fly brain, applied a simple neuron model (@Philip_Shiu Nature 2024) and used it to control a MuJoCo physics-simulated body, closing the loop from neural activation to action.' Cofounder Alexander Wissner-Gross shared another version, labeling it the 'world’s first embodiment of a whole-brain emulation that produces multiple behaviors' and alluding to an approaching technological singularity.
Eon Systems, based in San Francisco, aims to build a full digital emulation of a mouse brain within the next two years, an ambitious timeline according to observers. The company's blog post, titled 'How the Eon Team Produced a Virtual Embodied Fly,' provides some technical details but stops short of a full scientific paper. It describes combining the FlyWire connectome — a detailed map of a fruit fly's neural connections — with a basic neuron model from a 2024 Nature paper by Philip Shiu, and integrating it into a MuJoCo simulation for the fly's body physics.
Andregg has claimed the simulation achieves '91% behavior accuracy,' though he offered limited explanation of the metric in his X thread or the blog. In a message to reporters, he insisted, 'This is, in our view, a real uploaded animal.' When pressed further, Andregg told The Appleton Times that the team and its academic collaborators believe 'this fly is conscious in a limited sense, it can smell, see, taste, etc.' He described the project as a 'MVP,' or minimum viable product, of an uploaded animal, acknowledging 'lots of limitations.'
However, experts in neuroscience and AI have raised significant doubts about the claims' validity and scope. Shahab Bakhtiari, a professor leading the systems neuroscience and AI lab at the University of Montreal, said the initial X posts 'obscured critical' details about the work. While praising the blog for providing more context, he noted it 'arrived a bit late, and remains insufficient to fully validate the claims.' Bakhtiari emphasized the need for a detailed technical report including software, code, and simulation environments to allow reproduction by other scientists.
Alexander Bates, a research fellow in neurobiology at Harvard Medical School who studies fly brains, echoed those concerns. He described Eon's output as 'under-delivered' for such a monumental assertion. According to Bates, the team stitched together existing projects like the FlyWire map, a simulated fly body, and interaction models, but 'for a claim of this magnitude, I would expect something that should spell out the whole approach in specifics.' He also pointed out that the virtual fly's behaviors require evaluation against real data with 'clearly defined metrics,' and noted the 91 percent figure remains unexplained. Bates added wryly, 'Also, the fly does not fly.'
Bates acknowledged that 'strong framing and hype can matter for fundraising,' but stressed that Eon's assertion of a 'real uploaded animal' lacks credibility. Aran Nayebi, a professor of machine learning at Carnegie Mellon University, went further, stating the group is 'not even close' to a full brain emulation. He explained that while connections between cells are shown, key elements like neurotransmitters and synaptic strengths are missing. Nayebi said the motor system isn't a 'true upload' either, adding, 'We are not even faithfully simulating its brain in silico.'
The debate extends beyond technical feasibility to fundamental questions of what constitutes an 'uploaded animal.' Eon's blog avoids directly calling the simulation a 'real fly,' but Andregg has maintained the claim with caveats. In follow-up communications, he conceded the work 'isn’t a perfect replica of a fly' and that Eon never claimed it was. He described uploading as 'not a binary concept,' with 'different levels,' and admitted uncertainty about how much biology is needed to capture essential information. 'There is a lot more work to be done to achieve the level of upload that we may want for ourselves someday,' Andregg said.
Philosophers and neuroscientists question whether the term 'uploaded animal' even applies. Jonathan Birch, a philosopher at the London School of Economics, was direct: 'I don’t think we should ever say ‘uploaded animal,’' he said. Birch argued that Eon is pursuing 'whole-brain emulation,' which leaves out the rest of the animal's biology. Tom McClelland, a philosopher at the University of Cambridge, agreed that biology is crucial to behavior, stating, 'So at best they’ve uploaded some of the fly’s mind and thereby uploaded some of the fly.'
Bakhtiari called it an 'open question' whether a 'real uploaded animal' is possible at all. The simulation, drawn from multiple flies' data, represents a composite rather than a single organism. Critics note it's a copy, not a transfer, raising implications like the ability to duplicate the digital fly endlessly — a point the hype around the project has largely overlooked.
The buzz began with a handful of X posts and videos showing the digital fly in a simulated environment, but no independent verification has emerged. Content creators and AI-focused accounts on X and Reddit repackaged the clips as groundbreaking news, speculating on human applications and invoking science fiction like The Matrix. Yet, without peer-reviewed evidence, the achievement remains unconfirmed.
Eon Systems' work builds on public resources like the FlyWire project, a collaborative effort to map the fruit fly brain's 139,255 neurons and over 50 million synapses. The 2024 Nature paper by Philip Shiu and colleagues introduced a neuron model that Eon adapted. MuJoCo, an open-source physics engine, handles the virtual body's movements. While these components are legitimate, experts say combining them doesn't equate to a conscious, uploaded entity.
As the company pushes toward more complex emulations, like a mouse brain by 2028, the fly project highlights tensions between technological ambition and scientific rigor. Andregg's vision of digital intelligence could accelerate AI development, but skeptics warn against overhyping preliminary simulations. For now, the virtual fly walks in code, but whether it's truly 'alive' or just a sophisticated model divides opinions sharply.
The episode underscores broader challenges in the AI and neuroscience fields, where rapid advancements often outpace validation. With endorsements from figures like Musk fueling investment, Eon may secure funding, but earning scientific consensus will require transparency and reproducible results. As Andregg put it, this is just the beginning of a long road toward digital minds.