AI 热点

Genesis AI 完成 5 亿美元 A 轮融资,推进通用机器人落地

Unite.AI更新于 11月17日 22:54

Genesis AI Raises $500M Series A to Accelerate Robot Deployment

AI 导读

Genesis AI 于 2026 年 11 月 17 日宣布完成 5 亿美元 A 轮融资,用于加速通用机器人的研发与部署,先从制造和物流场景起步,并扩充团队。公司此前于 2025 年 7 月 1 日结束隐身状态,获得由 Eclipse 和 Khosla Ventures 联合领投的 1.05 亿美元种子轮。其技术栈包括机器人基础模型 GENE、机器人硬件 Eno 和部署运营三部分,公司称三者自有可让真实场景中的失败反哺模型训练与硬件迭代。

推荐理由

AI 评分 61

融资细节之外,原文还交代了 GENE 模型、Eno 硬件与部署三线并行的思路,以及 20 步各 99% 成功率只剩 82% 的可靠性算例。

主题

标签

正文

Genesis AI announced a $500 million Series A on November 17, 2026, to accelerate development and deployment of its general-purpose robots, beginning in manufacturing and logistics, and to expand its team.

The financing follows its July 1, 2025 emergence from stealth with $105 million in seed funding co-led by Eclipse and Khosla Ventures. Genesis is now directing capital toward both its underlying robotics technology and the industrial deployments intended to make that technology commercially useful.

Genesis’s strategy spans GENE, its robotics foundation model; Eno, its robot hardware; and the deployment operation connecting the machines to customer workflows. The company says owning all three allows real-world failures to inform model training and subsequent hardware, while coordinating hardware and model development and using manufacturing scale to improve reliability and reduce costs.

That design is intended to address reliability across complete jobs rather than isolated demonstrations. Genesis’s reliability example multiplies the risk across a workflow: when each of 20 steps has a 99% success rate, the probability of an error-free run falls to 82%. The example explains why the company treats intelligence, hardware and deployment as a shared reliability problem, with one weak component potentially limiting the entire system.

Human Data and Simulation

The model’s training pipeline begins with a human-scale robotic hand and a data-collection glove fitted with tactile-sensing electronic skin. In its May 2026 GENE-26.5 announcement, Genesis described a 1:1:1 mapping among the human hand, glove and robot hand, designed to turn human skills into training data. People wearing the glove during ordinary work can generate job-based data, supplemented by egocentric video from camera-wearing people and human-focused internet video.

Genesis reported that the glove’s hardware cost was 100 times lower than typical options and that internal testing showed up to five times greater data-collection efficiency, with better data quality than traditional teleoperation.

The company’s GENE-26.5 demonstrations spanned cooking, pipetting and liquid transfer, wire bundling and grasping several objects simultaneously. GENE serves as a shared robotics foundation model, while Genesis World extends the training loop into simulation. The simulator exposes failures across millions of scenarios, using them to guide additional data collection and model changes. Once Eno is deployed, real-world failures can also become simulation cases, training data and hardware revisions.

Eno’s Deployment Roadmap

Genesis unveiled Eno on June 16, 2026, as its first general-purpose robot. A wheeled base supports jointed panels that adjust height and reach and fold for storage, while its two arms use the company’s human-scale dexterous hands. The design prioritizes mobility and manipulation over a fully human appearance.

Genesis describes Eno and GENE as an integrated system capable of managing multi-step workflows, including keeping production lines stocked and preparing facilities for the next shift. An optional screen is planned to show nearby workers the robot’s intent, reasoning and operational state.

At that unveiling, Genesis said it planned to begin production and targeted customer deployments by the end of 2026. Manufacturing, logistics and laboratory customers were first in line, followed by hotels and hospitals, with home and outdoor applications later. The Series A roadmap maintains that sequence, with expansion beyond factories and warehouses dependent on improvements in capability and affordability.

A separate June 16 partnership announcement with LG CNS provides a potential deployment route. The plan calls for evaluating and validating robots for manufacturing and logistics at LG operations in the United States and within LG CNS’s industrial customer base, initially in the U.S. A subsequent pilot is intended to create a deployment playbook covering workflow documentation, data collection, further model training and validation before a broader rollout.

The new funding makes those industrial deployments the critical test of Genesis’s approach. Its broader expansion depends on converting dexterity demonstrations into reliable daily workflows, then using the resulting failures and operational data to improve subsequent models and robots.

来源

Unite.AI