# Generalist AI

> Source: https://aiwiki.ai/wiki/generalist_ai
> Updated: 2026-06-28
> Categories: AI Companies, Embodied AI, Robotics
> License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
> From AI Wiki (https://aiwiki.ai), the free encyclopedia of artificial intelligence. Reuse freely with attribution to "AI Wiki (aiwiki.ai)".

**Generalist AI** (branded **Generalist**, legally Generalist AI, Inc.) is an American artificial intelligence company building general-purpose [foundation models](/wiki/robot_foundation_model) for robots and other physical systems. Founded in 2024 and based in San Mateo, California, it was started by former [Google DeepMind](/wiki/google_deepmind) robotics scientists Pete Florence and Andy Zeng together with former [Boston Dynamics](/wiki/boston_dynamics) roboticist Andrew Barry, and it trains a single family of [embodied AI](/wiki/embodied_ai) models meant to run across many different robots rather than building any one robot of its own [6][8][14]. In June 2026 the company raised $400 million at a reported $2 billion valuation in a round led by Radical Ventures, with [Nvidia](/wiki/nvidia) and Bezos Expeditions among its backers, bringing total funding to more than $500 million [1][2][3].

The company describes itself as "a frontier AI research and product driven company building general intelligence for the physical world," with the tagline "We train robots that work" [14]. Its public models are named in the GEN series: GEN-0, released on November 4, 2025, and GEN-1, released on April 2, 2026 [11][9].

## What is Generalist AI?

Generalist AI is a robot foundation model startup: rather than designing a humanoid or a warehouse machine, it builds the software intelligence that is meant to control many kinds of robots from one trained model. The company positions cross-embodiment generalization, not hardware, as the central bottleneck in robotics, arguing that "the future of robotics is bigger than any single robot" [10][14]. This distinguishes it from companies that ship a specific robot and from the generic notion of a "generalist agent" such as DeepMind's Gato; Generalist AI's focus is a manipulation-capable physical foundation model trained on large-scale real-world interaction data [6][10].

## History

### Founding

Generalist was founded in 2024 by three researchers with deep backgrounds in robot learning [6][8]. Pete Florence, the company's chief executive, earned a PhD in computer science from MIT under Russ Tedrake and then worked as a senior research scientist at Google DeepMind, where he helped create [RT-2](/wiki/rt_2) (a vision-language-action model that transfers web knowledge to robotic control) and PaLM-E (an embodied multimodal language model) [6][16]. Andy Zeng, the chief scientist, earned a PhD at Princeton, then worked as a research scientist and tech lead at Google DeepMind, where he co-authored PaLM-E and contributed to methods for robots that "write their own code" and to low-cost handheld data collection for robots [15]. Andrew Barry, the chief technology officer, spent roughly five years as a senior roboticist at Boston Dynamics and was Florence's PhD classmate at MIT [6]. The broader team includes engineers from [OpenAI](/wiki/openai), Google DeepMind, and Boston Dynamics [8].

The startup spent its first year largely in stealth [6]. It first drew public attention in March 2025, when TechCrunch reported that Florence had left DeepMind and that Nvidia's venture arm had already backed his new company [6]. At that stage Florence framed the mission as making general-purpose robots a reality and ultimately driving "the marginal cost of physical labor to zero" [6]. Florence's DeepMind research was cited several times in DeepMind's own March 2025 robotics paper, underscoring how recently he had left frontier robotics research to start the company [6].

### Who founded Generalist AI?

Generalist AI was co-founded in 2024 by Pete Florence (CEO), Andy Zeng (chief scientist), and Andrew Barry (CTO) [6][8]. Florence and Zeng were the DeepMind robotics scientists behind RT-2 and PaLM-E, two of the most-cited embodied AI models, while Barry built robots such as Atlas and Spot at Boston Dynamics, giving the founding team a combination of frontier robot-learning research and production robotics experience [6][15][16].

### When was Generalist AI founded, and how is it funded?

Generalist's seed funding came together in 2024. Boldstart Ventures records its first investment in the company on March 24, 2024, alongside Nvidia as a co-investor, and Boldstart, Spark Capital, and NFDG were among the early backers [7][8]. Public databases also list a seed round dated to March 2025, reflecting the period when the previously stealth company became more visible [17].

On June 4, 2026, Generalist announced a $400 million funding round, reported by Bloomberg at a $2 billion valuation [1]. Radical Ventures led the round, with new investors 8VC, Union Square Ventures, Hanabi Capital, and Norwest, and returning investors including Nvidia's NVentures, Boldstart Ventures, Spark Capital, Bezos Expeditions, and NFDG [2][3]. Angel investors in the round included Zoom founder Eric Yuan, Xiaomi co-founder Bin Lin, AI researcher Fei-Fei Li, and entrepreneur Naval Ravikant [3][4]. The round brought Generalist's total funding to more than $500 million [1][2]. Some outlets characterized the financing as a Series B, while others described it only as a new round; the company said it would use the capital to build new models, scale its physical-data engine, expand compute and training infrastructure, and pursue commercialization partnerships [2][3].

| Date | Event | Amount | Lead / notable investors |
| --- | --- | --- | --- |
| March 2024 | Seed (first check) | Undisclosed | Boldstart Ventures, Nvidia, Spark Capital, NFDG |
| June 4, 2026 | New round (reported Series B) | $400 million | Radical Ventures; Nvidia NVentures, Bezos Expeditions, 8VC, Union Square Ventures, Norwest, Hanabi Capital |

Reported valuation at the June 2026 round was about $2 billion, and cumulative funding exceeded $500 million [1][2].

## What is Generalist AI building?

Generalist pursues a strategy that differs from most well-funded robotics companies, which tend to design a specific machine such as a humanoid or a warehouse robot. Instead, Generalist concentrates on the software, training one cross-embodiment foundation model meant to control many robot types [6][10]. The company frames this around the idea that "the future of robotics is bigger than any single robot," and its public materials stress that "goals are more powerful than methods" [10][14]. It has positioned its models as distinct from both conventional vision-language-action models ([VLAs](/wiki/robot_foundation_model)) and from world-model approaches, saying GEN-1 is trained from scratch rather than fine-tuned from an existing language or vision backbone [9].

Central to the approach is a large, proprietary dataset of real-world physical interaction. The company collects manipulation data across thousands of homes, warehouses, and workplaces worldwide, in part using wearable "data hands" (UMI-style handheld grippers) that capture human reflexes and micro-corrections [10][11]. By the time of GEN-0 the dataset stood at roughly 270,000 hours of real-world manipulation data, which the company said was growing by about 10,000 hours per week [10][11]. This emphasis on scaling real physical-interaction data places Generalist alongside other [robot foundation model](/wiki/robot_foundation_model) efforts such as [Physical Intelligence](/wiki/physical_intelligence) and [Skild AI](/wiki/skild_ai), though Generalist's particular claim is that robot performance follows predictable scaling laws as data grows [10][12].

## Models

### What is GEN-0?

GEN-0, announced on November 4, 2025, was presented as a new class of embodied foundation models that scale with physical-interaction data [11]. Generalist reported a power-law relationship between pretraining data and downstream task performance across more than 16 tasks, analogous to the scaling laws seen in [large language models](/wiki/large_language_model) [11][12]. The company also reported what it called the first observation of model "ossification" in robotics: models around 1 billion parameters showed early ossification (an inability to keep absorbing new information), while models of about 7 billion parameters and larger continued to improve as pretraining data grew [11]. Generalist tested GEN-0 across robots with 6, 7, and 16 or more degrees of freedom and discussed model sizes of 1B, 6B, 7B, and 10B-plus parameters [11].

A defining feature of GEN-0 is what Generalist calls "Harmonic Reasoning," in which the model is trained to think and act at the same time by processing asynchronous, continuous-time streams of sensing and action tokens [11][12]. Coverage in Andrew Ng's The Batch and elsewhere described the work as showing that training power laws translate from language to robotics [12]. Several outlets framed GEN-0 as a potential "ChatGPT moment" for embodied AI, though that comparison came from commentators rather than from independent benchmarking [10][12].

### What is GEN-1?

GEN-1, released on April 2, 2026, was described by the company as its first general-purpose model to reach "mastery of simple physical tasks," where Generalist defines mastery as "the combination of all of: reliability, speed, and improvisational intelligence" [9]. Generalist reported that GEN-1 "improves average success rates to 99% on tasks where previous models achieve 64%, completes tasks roughly 3x faster than state of the art" [9]. According to coverage, specific dexterous tasks ran far faster than before, for example box assembly in about 12.1 seconds (roughly 2.8 times faster than GEN-0's 34 seconds) [2]. The company said a new robot or task could be adapted with as little as one hour of robot-specific data, because the base model is pretrained primarily on human-demonstration data and only the final stage of training uses the target robot hardware; in some tests it matched GEN-0 with about 10 times less task-specific data [2][9]. Training data was reported to have expanded from 270,000 hours for GEN-0 to over 500,000 hours for GEN-1 [2][9].

Generalist demonstrated sustained reliability on repetitive real-world tasks, citing runs such as 200 consecutive box-folding repetitions, more than 200 robot-vacuum servicing repetitions, and more than 1,800 block-packing repetitions [9]. Florence described GEN-1's behavior in terms of reliability, speed, and "intelligent improvisation," pointing to unrehearsed corrections such as a robot using a second arm to shake a bag when an object snagged [9]. Comparing the moment to the early days of large language models, Florence said: "What's happening now with robotics parallels when people opened GPT-3 and asked it to write a completely new limerick" [9]. The company said GEN-1 marked a transition from research prototype toward commercial viability and was being made available to early-access partners [9].

| Model | Released | Training data | Headline metric | Notable feature |
| --- | --- | --- | --- | --- |
| GEN-0 | Nov 4, 2025 | ~270,000 hours | Power-law scaling across 16+ tasks; ossification below ~7B params | Harmonic Reasoning (think and act simultaneously) |
| GEN-1 | Apr 2, 2026 | 500,000+ hours | ~99% average success vs ~64% prior; ~3x faster | Trained from scratch; ~1 hour data to adapt to a new robot |

## How does Generalist AI differ from a generalist agent like Gato?

The term "generalist agent" predates this company and is most associated with DeepMind's Gato, a single network trained to perform many tasks across modalities. Generalist AI is a separate organization whose goal is narrower in domain but deeper in the physical world: a robot foundation model for real-world manipulation that generalizes across robot bodies (cross-embodiment), trained on hundreds of thousands of hours of physical-interaction data and validated on physical hardware rather than primarily in simulation or on benchmark suites [6][10][11]. In short, "generalist agent" is a research concept, while Generalist AI is a venture-funded company building production [robot learning](/wiki/robot_learning) systems for [robot manipulation](/wiki/robot_manipulation).

## How has Generalist AI been received?

Generalist's emergence has been covered by Bloomberg, The Robot Report, SiliconANGLE, Silicon Republic, and robotics-focused outlets such as Humanoids Daily, and its research has been discussed in DeepLearning.ai's The Batch [1][2][3][4][9][12]. Reporting has emphasized the unusually strong investor lineup (Nvidia, Bezos Expeditions, Fei-Fei Li, Naval Ravikant) and the founders' direct lineage to influential DeepMind robotics work like RT-2 and PaLM-E [3][4][6]. Observers have placed Generalist within a wave of well-capitalized robot-foundation-model startups racing toward what some describe as "physical AGI" [5]. Independent, third-party benchmarking of GEN-0 and GEN-1 remains limited, and most performance figures originate from the company's own reports, a caveat that applies broadly across the robot-foundation-model field [9][12].

## Related

- [Robot foundation model](/wiki/robot_foundation_model)
- [Embodied AI](/wiki/embodied_ai)
- [Physical Intelligence](/wiki/physical_intelligence)
- [Skild AI](/wiki/skild_ai)
- [Google DeepMind](/wiki/google_deepmind)
- [RT-2](/wiki/rt_2)
- [Robot learning](/wiki/robot_learning)
- [Robot manipulation](/wiki/robot_manipulation)

## References

1. "Nvidia-Backed Robotics Startup Generalist AI Valued at $2 Billion." Bloomberg, June 4, 2026. https://www.bloomberg.com/news/articles/2026-06-04/nvidia-backed-robotics-startup-generalist-ai-valued-at-2-billion
2. "Generalist raises $400M to scale its general-purpose AI models." The Robot Report, June 4, 2026. https://www.therobotreport.com/generalist-raises-400m-to-scale-its-general-purpose-ai-models/
3. "Generalist AI raises $400M at $2B valuation to build general intelligence for robotics." SiliconANGLE, June 4, 2026. https://siliconangle.com/2026/06/04/generalist-ai-raises-400m-2b-valuation-build-general-intelligence-real-world/
4. "Nvidia, Fei-Fei Li back Generalist's $400m round to scale AI robotics." Silicon Republic, June 2026. https://www.siliconrepublic.com/start-ups/nvidia-fei-fei-li-back-generalists-400m-round-to-scale-ai-robotics
5. "Generalist AI Raises $400M in New Funding to Develop Physical AGI." The AI Insider, June 4, 2026. https://theaiinsider.tech/2026/06/04/generalist-ai-raises-400m-in-new-funding-to-develop-physical-agi/
6. "A key DeepMind robotics researcher left Google, and Nvidia has already backed his stealth startup." TechCrunch, March 19, 2025. https://techcrunch.com/2025/03/19/a-key-deepmind-robotics-researcher-left-google-and-nvidia-has-already-backed-his-stealth-startup/
7. "GeneralistAI, When Robots Start to Improvise, Welcome to boldstart." Boldstart Ventures, 2026. https://boldstart.vc/news/generalistai-when-robots-start-to-improvise-welcome-to-boldstart/
8. "Generalist." Boldstart Ventures company profile, 2026. https://boldstart.vc/companies/generalist/
9. "Generalist AI Unveils GEN-1: The Quest for Robot Mastery and 'Intelligent Improvisation'." Humanoids Daily, April 2026. https://www.humanoidsdaily.com/news/generalist-ai-unveils-gen-1-the-quest-for-robot-mastery-and-intelligent-improvisation
10. "Generalist AI Unveils GEN-0, Claims Scaling Laws for Robotics Backed by 270,000 Hours of Real-World Data." Humanoids Daily, November 2025. https://www.humanoidsdaily.com/news/generalist-ai-unveils-gen-0-claims-scaling-laws-for-robotics-backed-by-270-000-hours-of-real-world-data
11. "GEN-0 / Embodied Foundation Models That Scale with Physical Interaction." Generalist AI, November 4, 2025. https://generalistai.com/blog/nov-04-2025-GEN-0
12. "Training power laws translate to robotics." The Batch, DeepLearning.ai, November 2025. https://www.deeplearning.ai/the-batch/training-power-laws-translate-to-robotics/
13. "Generalist AI Introduces GEN-theta: A New Class of Embodied Foundation Models." MarkTechPost, November 5, 2025. https://www.marktechpost.com/2025/11/05/generalist-ai-introduces-gen-%CE%B8-a-new-class-of-embodied-foundation-models-built-for-multimodal-training-directly-on-high-fidelity-raw-physical-interaction/
14. "Generalist AI." Official website, 2026. https://generalistai.com/
15. "Andy Zeng." Personal homepage, 2026. https://andyzeng.github.io/
16. "Pete Florence's Home Page." Personal homepage, 2026. https://www.peteflorence.com/
17. "Generalist AI, 2026 Company Profile, Team, Funding and Competitors." Tracxn, 2026. https://tracxn.com/d/companies/generalistai/__TaFrISSyLZQX6OYhyWKASir1Itlg9FAM-PKgk6cKzYE