About
Aemon builds an autonomous AI research engineer for R&D and computational teams in fields such as quant finance, biotech, logistics, and materials science. Its differentiator is self-accelerating research: it reads frontier papers, generates and tests thousands of approaches inside customers’ technical environments, and validates improvements against trusted evaluations.
Market
Aemon competes in agentic AI for scientific and technical R&D, spanning autonomous research, optimization, and software-engineering agents. It positions itself as a forward-deployed research engineer rather than a stand-alone literature or coding assistant: it starts from a customer-defined evaluation, studies frontier work, experiments inside the customer codebase, and iterates toward validated production improvements. Its stated differentiation is long-horizon, expert-steerable search at machine scale, supported by a claimed sub-$10 circle-packing result that beat Google DeepMind’s AlphaEvolve result.
Technical R&D teams—especially AI/ML, software, and algorithm-heavy groups—working on difficult, measurable problems in real or proprietary codebases. Likely buyers are CTOs, heads of engineering, or research leads at startups, scale-ups, and enterprise R&D organizations; the available evidence identifies R&D teams but does not establish a narrower company-size segment.
At a Glance
Problem
Aemon addresses a core bottleneck in technical R&D: teams can define the objective and how success should be measured, but often do not know which approach will achieve it. That uncertainty makes R&D expensive and slow; the stakes are significant given the company’s cited estimate of $3.1 trillion in annual global R&D spending. The main use case is a hard scientific or engineering problem—particularly optimization, machine learning, or other problems where many possible approaches must be tested—such as Aemon’s claimed result on the NP-hard circle-packing problem.
Product / Service
Aemon is positioned as a forward-deployed, autonomous AI R&D engineer. It converts a technical goal into a trusted evaluation with datasets, metrics, baselines, failure cases, and production constraints; studies relevant research and implementations; then generates, tests, and evolves large numbers of approaches inside the customer’s codebase. Each experiment is recorded so successful approaches, failures, and constraints improve the next research cycle, while human experts remain able to steer the system.
The delivery model combines software with an applied research engagement. An Aemon engineer works with the customer to harden the evaluation, after which Aemon connects to the team’s codebase and research context and runs self-accelerating research loops. The promised benefit is faster discovery of validated, production-ready improvements, with the company stating that it aims to move from the first call to technical breakthroughs in roughly two weeks.
Market
Aemon is competing in an emerging category of autonomous AI research engineering: B2B systems that do more than assist with coding by understanding a technical objective, designing evaluations, running experiments, and iterating toward a breakthrough. Adjacent alternatives include general-purpose AI platforms and coding agents from OpenAI, Anthropic, and Factory, as well as automated research systems such as The AI Scientist. Google DeepMind’s AlphaEvolve is a notable technical benchmark and adjacent capability; Aemon claims to have surpassed one of its 2025 results on circle packing, rather than merely matching a standard coding-assistant workflow.
The company is very early but has meaningful technical and initial-market signals. YC lists Aemon as an active Winter 2026 company founded in 2025, with a three-person team, while the founders say Aemon is already in the hands of R&D teams. The public evidence also includes the claimed world-record optimization result achieved with less than $10 of compute and a reported pre-seed round. No named customers, pricing, or revenue are disclosed in the reviewed material, so for diligence purposes Aemon is best treated as pre-revenue or revenue-undisclosed, with early product and technical validation rather than demonstrated commercial scale.
Founders & Leadership
Funding History
Y Combinator
Recent News
Firecrawl reported that Aemon, which builds autonomous AI research engineers, observed similar results in its benchmark of scientific and technical retrieval systems. The mention appeared in coverage of Firecrawl's specialized research index.
A Forbes Councils profile identified Richard Zhou as Aemon's CEO and co-founder and described the company as developing an autonomous AI research engineer. The profile also stated that Aemon is backed by Y Combinator, Nexus Venture Partners, SV Angel, Liquid2 Ventures, Afore Capital, and notable angels including Paul Graham.
Forbes featured Aemon AI among Y Combinator's Winter 2026 batch, describing its systems as designed to automate scientific discovery by exploring large solution spaces and testing ideas faster than human teams. The article stated that each company in the batch receives $500,000 in funding.
Active Roles
1Business Model
Aemon appears to use a direct, high-touch B2B model in which its AI research engineer is deployed to customer teams and supported through discovery, evaluation design, and integration with their codebases. The company does not publicly disclose pricing or a specific recurring-revenue structure.