Genesis AI, a Palo Alto-based robotics startup with backing from Khosla Ventures, has secured $105 million in funding to develop what it describes as a foundational AI model for robotics. The company has simultaneously unveiled its first model, designated GENE-26.5, along with a general-purpose humanoid robot called Eno that runs on that model.
The funding round and product announcements position Genesis AI within a crowded but fast-moving segment of the robotics industry: companies attempting to build general-purpose AI systems that can operate across multiple robot form factors rather than being hard-coded for a single task or machine.
GENE-26.5 is the company's first publicly disclosed foundation model. Genesis AI claims the model provides robots with "human-level physical manipulation capabilities," a phrase that appears in the company's own materials and should be understood as a vendor claim rather than an independently verified benchmark. The model is designed to absorb large volumes of data and operate across varied environments, according to the company.
What makes the announcement notable is not just the model itself but the surrounding hardware and data infrastructure. Genesis AI has developed a sensor-loaded glove that functions as a real-world counterpart to its robotic hand. The glove is intended to collect manipulation data from human operators performing everyday tasks, which can then be used to train the model. The company describes this as a way to "unlock unlimited amounts of data" — again, a vendor claim — and to train GENE-26.5 at scale.
The glove's practical value, according to Gervet, a former research scientist at Mistral AI who now serves as Genesis AI's president, is that it allows data collection during normal work rather than in a dedicated lab setting. Gervet said the company is in talks with potential customers, and that the glove could be worn by lab technicians in pharmaceutical or manufacturing settings while they perform their daily duties. This approach, if it works as described, would represent a shift from the more common practice of collecting robot training data through teleoperation or scripted demonstrations in controlled environments.
The company also introduced Eno, its first general-purpose humanoid robot. Eno operates using the GENE foundation model and is described by Genesis AI as a "true physical agent" that can reason, adapt, and take responsibility for outcomes beyond pre-defined tasks. Notably, Eno will be available in a version with an optional screen that displays a cognitive interface, showing what the robot is thinking and doing in real time. This transparency feature appears designed to address trust concerns in human-robot interaction, particularly in settings where operators or bystanders may be uncertain about a robot's intentions or decision-making process.
Genesis AI describes itself as a "global full-stack robotics company" — meaning it develops hardware, software, and data infrastructure internally rather than relying on third-party components for each layer. The company's co-founder and CEO, Zhou Xian, emphasized this integrated approach, stating that the only path to a robot that delivers real-world value is through intentional design and a single, comprehensive system. Eric Schmidt, former CEO of Google and an investor in Genesis AI, characterized the work as "a fundamentally new model for extending human capability through advanced robotics."
The funding and product launch come amid broader investor interest in general-purpose robotics AI. In a related development covered by the same source material, Generalist AI Inc., a separate company founded in 2024, raised $400 million in new funding to scale its own general-purpose AI models for robotics. That round brought Generalist AI's total funding to more than half a billion dollars. The company released its GEN-1 model in April and describes its work as building embodied foundation models for general-purpose robots. Generalist AI attributes its progress to "thousands of decisions" across data, models, hardware, infrastructure, operations, and deployment, made by a team working at the frontier of AI and robotics.
The juxtaposition of these two funding events underscores a clear trend: investors are placing large bets on the idea that general-purpose AI for robots is a solvable problem with massive commercial upside. Whether those bets pay off depends on factors that remain unproven at scale — data collection efficiency, model generalization, hardware reliability, and real-world deployment economics.
Why it matters for European robot service
For European buyers, operators, and service providers in the robotics ecosystem, the Genesis AI announcement carries implications that extend beyond a single startup's product roadmap.
First, the sensor-loaded glove concept deserves attention from European industrial sectors that rely on skilled manual labor. The idea that a worker could wear a data collection device during their normal shift — in a pharmaceutical lab, a manufacturing line, or a logistics facility — and thereby contribute to training a robot that could eventually assist or take over certain tasks, is a fundamentally different model from the traditional approach of dedicated robotics engineers programming or teleoperating machines. If this approach matures, it could lower the barrier to robot deployment in small and medium-sized European enterprises that lack in-house robotics expertise. The glove would effectively turn existing workers into data contributors, creating a pipeline of task-specific training data that reflects real operational conditions rather than idealized lab scenarios.
However, European buyers should also consider the practical and regulatory questions that this data collection model raises. The General Data Protection Regulation (GDPR) imposes strict requirements on the collection and processing of personal data. A glove worn by a worker that records hand movements and manipulation data could potentially capture information that falls under GDPR's scope, depending on what exactly is recorded and how it is linked to identifiable individuals. The source material does not disclose whether Genesis AI has addressed GDPR compliance, how long data is retained, who owns the data, or whether workers would be informed that their movements are being used to train AI models. These are material questions for any European organization considering adoption.
Second, the humanoid form factor of Eno raises questions about suitability for European work environments. European manufacturing and logistics facilities are often older and more space-constrained than newly built facilities in other regions. Humanoid robots, by virtue of their size and shape, require adequate floor space, clearance, and safety infrastructure. The source material does not specify Eno's dimensions, weight, payload capacity, or safety certifications. European buyers should not assume that a humanoid form factor is inherently superior to fixed or mobile manipulators; the right choice depends on the specific tasks, environment, and regulatory context.
Third, the transparency feature on Eno — the optional screen showing the robot's cognitive interface — is a notable design choice that could resonate with European customers and regulators. The European Union's ongoing work on AI regulation, including the AI Act, emphasizes transparency and human oversight for AI systems deployed in high-risk settings. A robot that can display its reasoning process in real time could help organizations demonstrate compliance with transparency obligations, and could also help build worker trust in human-robot collaboration. That said, the source material does not specify what exactly the cognitive interface displays, how detailed the explanations are, or whether the screen is a genuine window into the model's decision-making or a curated summary. Buyers should ask for a demonstration before making assumptions.
Fourth, the broader funding environment for general-purpose robotics AI is relevant to European service providers who are deciding which platforms to build their own services around. The entry of well-capitalized players like Genesis AI and Generalist AI into the foundation model space could accelerate the commoditization of certain robotics capabilities — manipulation, navigation, task planning — that were previously custom-built for each deployment. For European system integrators and robot service providers, this could mean a shift from building bespoke AI solutions to configuring and deploying foundation-model-based systems. That shift carries both opportunity and risk: opportunity to reduce development costs and time-to-market, and risk of becoming dependent on a small number of large AI providers whose pricing, licensing, and data policies may not align with European preferences for data sovereignty and open standards.
Fifth, the source material indicates that Genesis AI is in talks with customers, but it does not disclose which customers, in which sectors, or in which geographies. European buyers should therefore treat the company's claims as pre-commercial. There is no disclosed evidence in the source material of a production deployment, a reference customer in Europe, or a track record of reliability in industrial settings. The company's claims about "human-level physical manipulation" and "unlimited amounts of data" are promotional statements, not verified performance metrics.
Finally, the involvement of Eric Schmidt as an investor, and the backing of Khosla Ventures, signals that Genesis AI has access to significant capital and influential networks. That does not, by itself, validate the technology, but it does suggest the company has the resources to sustain a long development cycle. For European buyers, that may be a relevant consideration when evaluating the risk of adopting a platform from a startup that could pivot, be acquired, or run out of funding.
What buyers and operators should know
For organizations in Europe that are evaluating Genesis AI's technology — or the broader category of general-purpose robotics foundation models — the following points are worth keeping in mind, based solely on what the source material discloses and does not disclose.
**What is known:** Genesis AI has raised $105 million, backed by Khosla Ventures. The company has released a foundation model called GENE-26.5 and a humanoid robot called Eno. The model is claimed to provide human-level physical manipulation capabilities. The company has developed a sensor-loaded glove for data collection. Eno can be equipped with an optional screen showing a cognitive interface. The company describes itself as a full-stack robotics company. Gervet, the company's president, is a former Mistral AI research scientist. Zhou Xian is co-founder and CEO. Eric Schmidt is an investor. The company is in talks with customers. The source material also reports on Generalist AI, a separate company that raised $400 million and released a model called GEN-1 in April.
**What is not disclosed:** The source material does not specify when the funding round closed, beyond the general timeframe of the reporting. It does not disclose the valuation of Genesis AI. It does not provide technical specifications for GENE-26.5 — such as parameter count, training data volume, compute requirements, or benchmark results against other models. It does not provide specifications for Eno — such as height, weight, degrees of freedom, payload, battery life, or operating environment. It does not disclose pricing for either the model, the glove, or the robot. It does not provide availability dates, deployment timelines, or target markets. It does not disclose any customer names, pilot programs, or production deployments. It does not provide safety certifications, such as ISO or CE markings, which are critical for European deployment. It does not address data privacy, GDPR compliance, or data ownership for glove-collected data. It does not specify whether the cognitive interface on Eno is a standard feature or an optional add-on, nor what it costs.
Practical considerations for European buyers:
1. **Verify claims independently.** The phrase "human-level physical manipulation capabilities" is a vendor claim. Ask for benchmark data, third-party evaluations, and reference deployments. If the company cannot provide these, treat the claim as aspirational.
2. **Clarify the data pipeline.** If you are considering the sensor-loaded glove, ask specific questions: What data is recorded? Where is it stored? Who has access? How is it used for training? Can your organization opt out of having its data used for the company's general model training? What happens to the data if you terminate the relationship?
3. **Assess the humanoid form factor.** Humanoid robots are not automatically the right choice for every task. Evaluate whether a humanoid form factor is genuinely necessary for your use case, or whether a simpler, cheaper, and more reliable fixed or mobile manipulator would suffice. The source material provides no evidence that Eno outperforms other form factors.
4. **Plan for the transparency feature.** The optional cognitive interface on Eno could be valuable for building worker trust and for regulatory compliance. However, you should ask to see exactly what the interface displays, how it handles edge cases, and whether it can be customized for your organization's needs.
5. **Consider platform risk.** Foundation models for robotics are an emerging category. The companies building them are well-funded but largely unproven in production. If you build your operations around a specific foundation model, you are exposed to that company's roadmap, pricing changes, and financial health. Consider whether you need portability or an abstraction layer that allows you to switch models if necessary.
6. **Monitor the competitive landscape.** The source material notes that Generalist AI raised $400 million for similar work. This suggests the market for general-purpose robotics AI is becoming competitive, which could be good for buyers in terms of pricing and innovation, but also means the landscape is still consolidating. It is too early to tell which platforms will become standards.
7. **Do not assume European availability.** The source material does not state when or whether Genesis AI products will be available in Europe. Regulatory approvals, safety certifications, and local support infrastructure are all unresolved. European buyers should not make procurement decisions based on announcements alone.
8. **Request a demonstration.** The source material mentions a video demonstrating GENE-26.5's performance with fluid, human-like dexterity. Videos can be edited and may not reflect real-world reliability. Ask for a live demonstration in your own facility, with your own tasks, before making any commitments.
9. **Budget for integration.** A foundation model and a humanoid robot are not turnkey solutions. You will likely need integration work, safety assessments, worker training, and ongoing maintenance. The source material does not disclose any of these costs.
10. **Keep expectations realistic.** The robotics industry has a long history of impressive demonstrations that failed to translate into reliable, cost-effective production systems. The claims made by Genesis AI are consistent with that pattern. Until there is independent evidence of sustained, reliable performance in real-world conditions, a prudent approach is to treat this as an emerging technology with potential, not a proven solution.
In summary, Genesis AI's $105 million raise and the launch of GENE-26.5 and Eno are significant developments in the field of general-purpose robotics AI. The company's full-stack approach, its sensor-loaded glove for data collection, and its transparency-focused cognitive interface are all noteworthy design choices. However, the source material leaves many critical questions unanswered — from technical specifications and pricing to regulatory compliance and deployment timelines. European buyers and operators should follow the company's progress, but should not make procurement decisions based on this announcement alone.
Sources
Genesis AI brings in $105M to build universal robotics foundation model
Published by Vigla Media OÜ (Estonia).