microagi
A data research lab building end-to-end physical AGI with a focus on reliable real-world deployments.
Website: https://www.microagi.ai/
Cover Block
Public sources
| Field | Detail |
|---|---|
| Company | microagi |
| Tagline | A data research lab building end-to-end physical AGI with a focus on reliable real-world deployments [microagi, Unknown] |
| Headquarters | Munich, Germany [Tech.eu, July 2026] |
| Founded | 2025 [Crunchbase] |
| Stage | Seed [SiliconANGLE, July 2026] |
| Business model | Hardware + Software |
| Industry | Deeptech |
| Technology | AI / Machine Learning |
| Geography | Western Europe |
| Growth profile | Venture Scale |
| Founding team | Co-Founders (5): Bercan Kilic, Yoan Iliev, Anton Poletaev, Nico Nussbaum, Artjem Weissbeck [SiliconANGLE, July 2026] [The Next Web, July 2026] [Business Insider, June 2026] [RobotToday] |
| Funding label | $50M+ |
| Total disclosed funding | Approximately $55,000,000 [SiliconANGLE, July 2026] [Crunchbase] |
Links
Public sources
- Website: https://www.microagi.ai/
Executive Summary
PUBLIC microagi is building a business around one of embodied AI's hardest bottlenecks: collecting task-specific real-world data and turning it into deployable robot behavior for factories, a proposition that drew unusual investor attention after its $55 million seed round in July 2026 [The Next Web, July 2026] [SiliconANGLE, July 2026]. Founded in 2025 in Munich, the company emerged from a team that mixes Formula One engineering, applied AI research, and prior startup experience, with CEO Bercan Kilic previously at Red Bull Racing, co-founder Yoan Iliev from Mercedes-AMG Petronas Formula One, Anton Poletaev from the Alan Turing Institute, and Artjem Weissbeck previously a co-founder of Charles [SiliconANGLE, July 2026] [The Next Web, July 2026] [Business Insider, June 2026] [RobotToday].
The core offer is Atlas, which the company positions as a platform for capturing human work data and using it to adapt existing robotics models to specific factory tasks, rather than trying to build a general-purpose robot stack from scratch [The Next Web, July 2026] [microagi] [LinkedIn]. That framing matters because early evidence suggests microagi is selling into a concrete deployment problem in manufacturing, with The Next Web reporting five companies collecting data through Atlas as of July 2026 and one customer preparing a factory-line robot deployment, although customer names and production outcomes remain undisclosed [The Next Web, July 2026].
The financing is the headline event: Hummingbird led the $55 million seed, with Northzone, LocalGlobe, Village Global, and redalpine also named, and multiple outlets described it as the largest seed round raised by a German company [SiliconANGLE, July 2026] [Semafor, July 2026] [The Next Web, July 2026]. Business model signals point to a mix of software and deployment work around industrial robotics, with an adjacent data-supply angle as Semafor reported that some datasets have been licensed exclusively to frontier AI labs [Semafor, July 2026].
Over the next 12 to 18 months, the key issue is not whether microagi can attract capital, but whether it can convert data collection into repeatable production deployments and durable revenue across manufacturers [The Next Web, July 2026] [Semafor, July 2026]. The public record supports genuine early momentum, but several more expansive traction claims, including reported $100 million-plus revenue run-rate figures tied to LinkedIn and secondary database posts, are not independently verified and should be treated cautiously until corroborated by primary reporting or customer evidence [LinkedIn, Retrieved 2026] [dealroom.co] [robotico.market].
Lightly corroborated -- Core company, team, and funding claims are corroborated by multiple public sources, but traction details and some product specifics rely partly on company descriptions or single-source reporting.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | Hardware + Software |
| Industry / Vertical | Deeptech |
| Technology Type | AI / Machine Learning |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding | $50M+ total disclosed, approximately $55,000,000 [SiliconANGLE, July 2026] [Crunchbase] |
How the Company Got Here
PUBLIC
microagi presents itself as a data research lab focused on end-to-end physical AGI and reliable real-world deployments, a positioning that places it at the intersection of robotics infrastructure and applied AI rather than pure software [microagi] [Crunchbase]. Public company materials and Crunchbase identify the business as founded in 2025 and headquartered in Munich, Germany [microagi] [Crunchbase].
The public record is still thin, which is typical for a company founded recently, but the chronology that is visible is straightforward. The company website establishes the current product and mission framing, while Crunchbase corroborates the 2025 founding year and the Munich base [microagi] [Crunchbase].
A first financing milestone appears in November 2025, when RobotToday reported a pre-seed round with undisclosed terms, but that source falls outside the evidentiary scope for this section and is not used here. Within the permitted source set, the next confirmed public milestone is Crunchbase's record of a seed financing in 2026, which aligns with the company's emergence into broader public view as a venture-backed robotics and AI startup [Crunchbase]. The legal entity name is not established in the cited materials used for this section, so it is better left unstated.
Independently corroborated -- Confirmed by Crunchbase and the company website.
Product and Technology
Sources and analysis
microagi is positioning itself less as a robot manufacturer than as a data and deployment layer for physical AI, and that distinction matters for how the product should be read. Public reporting describes the company as capturing human work with cameras and sensor-equipped gloves, converting those demonstrations into training material for existing robotics models, and then supporting deployment for specific factory tasks in customer environments [The Next Web, July 2026]. The company frames itself more broadly as "a data research lab working on end-to-end physical AGI" focused on reliable real-world deployments, while its website and press coverage consistently point to industrial use cases first rather than a general-purpose robot platform [microagi, Unknown] [SiliconANGLE, July 2026].
The named product surface is Atlas, which company-linked materials describe as a platform for robotics deployment that captures and curates factory data, then uses that data to fine-tune robotics models and improve them through a reinforcement loop on the factory floor [LinkedIn, Retrieved 2026] [pulse2.com]. That loop is directionally plausible, but the evidence is still uneven: independent coverage more clearly supports the narrower claim that microagi records people performing work, adapts existing models to the task, and helps manufacturers prepare deployments on production lines [The Next Web, July 2026] [SiliconANGLE, July 2026]. As of July 2026, The Next Web reported that five companies were collecting data through Atlas and that one customer was preparing to deploy robots on a factory line, which suggests the product is past pure lab work but still early in commercial rollout [The Next Web, July 2026].
A second product surface appears to be Shift, a data-collection program tied to household tasks rather than factory processes. Several reports say Shift offers free home cleaning in New York City, with cleaners wearing cameras so the resulting footage can be used as training data for household robots, and other coverage describes a broader paid contributor model for recording chores and everyday tasks [Business Insider, June 2026] [arstechnica.com] [dealroom.co] [ai-market-watch.com]. Separately, Semafor reported that microagi has licensed some of its data exclusively to frontier AI labs, which implies the underlying asset may be useful both for internal model training and for external data monetization, though the counterparties and economics were not disclosed [Semafor, July 2026].
Single unverified source -- Material product claims combine independent press reporting with company and LinkedIn-sourced descriptions, and several important details about Atlas and Shift remain only partially corroborated.
Where the Demand Sits
PUBLIC The market matters now because industrial automation demand is colliding with a visible bottleneck in robot training and deployment, and microagi is positioned at that constraint rather than at the level of general-purpose robot hardware [The Next Web, July 2026].
Public sourcing does not support a direct TAM, SAM, or SOM for microagi’s exact category, so the cleaner approach is to anchor on adjacent industrial robotics demand and treat it as an analogous market rather than a company-specific sizing exercise. One reported data point, cited by pulse2.com, says China installed about 295,000 factory robots in 2024, or 54% of global installations, while Europe installed about 85,000, or 16% [pulse2.com]. Those figures are only single-source in this research set, but they are directionally useful because microagi is targeting manufacturers and positioning Atlas as a deployment layer for factory tasks rather than a consumer robotics platform [The Next Web, July 2026] [pulse2.com].
The immediate demand signal is narrower and more concrete than broad robotics adoption statistics. The Next Web reported in July 2026 that five companies were already collecting data through Atlas and that one customer was preparing a factory-line deployment [The Next Web, July 2026]. That does not establish category-scale penetration, but it does suggest buyers are willing to fund data capture and task adaptation before fleet-scale robot rollouts, which is consistent with a market that may first clear through workflow-specific pilots rather than full factory automation programs [The Next Web, July 2026] [SiliconANGLE, July 2026].
| Segment or signal | Public claim | Interpretation |
|---|---|---|
| China factory robot installations, 2024 | 295,000 [pulse2.com] | Indicates the largest current industrial robotics installation base in the cited set |
| Europe factory robot installations, 2024 | 85,000 [pulse2.com] | Relevant regional proxy for a Munich-based supplier selling into European manufacturing |
| Share of worldwide installations, China | 54% [pulse2.com] | Suggests scale is concentrated in Asia, even if microagi begins in Europe |
| Share of worldwide installations, Europe | 16% [pulse2.com] | Implies a meaningful but smaller installed base near the company’s home market |
| Companies collecting data through Atlas, July 2026 | 5 [The Next Web, July 2026] | Early evidence of workflow-level commercial interest |
| Customers preparing factory-line deployment, July 2026 | 1 [The Next Web, July 2026] | Points to movement from data capture into production use |
The table points to a market structure where installation volume is already large, but value may accrue to firms that reduce the cost and time required to adapt robots to specific tasks. For microagi, that means the addressable opportunity is likely better understood as a middleware and data layer riding industrial robotics capex, not as a direct read on total robot unit demand [The Next Web, July 2026].
Demand drivers in the public record are practical rather than thematic. Multiple reports describe microagi’s wedge as solving the training-data and deployment bottleneck by recording people performing work, adapting existing robotics models, and supporting production-line deployment for narrowly defined tasks [The Next Web, July 2026] [SiliconANGLE, July 2026]. That framing matters because it ties the company to a budget already familiar to manufacturers, factory automation, process optimization, and labor substitution in repeatable workflows, instead of asking customers to underwrite speculative humanoid programs [The Next Web, July 2026].
Adjacent markets broaden the opportunity set, but they also complicate the read-through from any single metric. Semafor reported that microagi has licensed some of its data exclusively to frontier AI labs, which places part of the business nearer to AI data infrastructure than to pure robotics software [Semafor, July 2026]. Separately, Business Insider and other cited sources describe Shift, a consumer-facing cleaning workflow in New York that captures household task footage for training, suggesting a second adjacent market in embodied-AI data collection outside factories [Business Insider, June 2026] [ai-tldr.dev] [dealroom.co].
The macro and regulatory backdrop is supportive in broad outline, though the present source set is thinner on formal policy analysis than on company reporting. Tech.eu reported a compute collaboration involving Google Cloud and NVIDIA, which is relevant because embodied-AI training is capital- and infrastructure-intensive, and access to cloud and accelerated compute can shape who is able to scale model improvement loops economically [Tech.eu, July 2026]. At the same time, the company’s own emphasis on de-identified datasets signals that privacy, worker-consent, and data-governance questions will matter, especially if collection spans factory floors and consumer environments [microagi] [Semafor, July 2026].
Single unverified source -- Section relies on a mix of independent reporting from The Next Web, SiliconANGLE, Semafor, Tech.eu, and one single-source adjacent market statistic from pulse2.com, with no directly cited third-party TAM study for microagi’s exact category.
Competitive Landscape
MIXED microagi is positioned less as a robot manufacturer than as a data and deployment layer for industrial automation, which places it in competition with incumbent robotics vendors, factory integrators, and newer embodied-AI data platforms rather than with a single direct peer set [The Next Web, July 2026] [microagi, Unknown].
The competitive map is easiest to read in three bands. First are the industrial incumbents, the established robotics OEMs and automation suppliers that already sell into automotive, logistics, and food production environments, the sectors microagi is also targeting [The Next Web, July 2026]. Those incumbents usually control installed hardware relationships and systems-integration budgets, which matters because microagi's offer appears to begin after a manufacturer has decided that a task should be automated. Second are deployment specialists and integrators, firms that adapt automation systems to a specific line or workflow. microagi's Atlas platform, as described publicly, overlaps with that layer because it captures task data, adapts existing robotics models, and supports deployment on a customer's production line [The Next Web, July 2026] [LinkedIn, Retrieved 2026]. Third are adjacent data suppliers to frontier model builders. Semafor reported that microagi has licensed some of its data exclusively to frontier AI labs, which suggests a second market where the company competes for scarce, high-quality embodied data rather than for factory automation budgets alone [Semafor, July 2026].
The edge today appears to rest on talent concentration, capital, and a narrow wedge around data capture. The founding group brings a mix of Formula One engineering, applied research, and startup operating experience, including CEO Bercan Kilic from Red Bull Racing, co-founder Yoan Iliev from Mercedes-AMG Petronas Formula One, Anton Poletaev from the Alan Turing Institute, and Artjem Weissbeck from Charles [SiliconANGLE, July 2026] [The Next Web, July 2026] [Business Insider, June 2026] [RobotToday]. The $55 million seed, led by Hummingbird with participation from Northzone, LocalGlobe, Village Global, and redalpine, gives the company more room than a typical European seed startup to finance data collection, on-site deployment work, and research hiring before a broad commercial base is visible [SiliconANGLE, July 2026] [Semafor, July 2026]. That edge is meaningful, but only partly durable. Capital can buy time, and specialist talent can build early credibility, yet a data advantage in industrial robotics is perishable if incumbent suppliers or large labs can gather equivalent task data through existing customer footprints.
The main exposure is channel control. Public reporting shows five companies collecting data through Atlas as of July 2026, and one customer preparing a factory-line deployment, which is a useful proof point but still small relative to the distribution reach that established automation vendors and integrators tend to hold [The Next Web, July 2026]. microagi also does not appear, from the cited public record, to own the core hardware stack; its differentiation rests on capturing human work, turning that into training material, and adapting existing robotics models to customer tasks [The Next Web, July 2026]. That makes the company vulnerable if hardware vendors internalize the same workflow, or if a well-capitalized model provider decides that exclusive data licensing is more strategically important than outsourcing collection to a third party [Semafor, July 2026].
The most plausible 18-month scenario is that the field bifurcates between companies that control deployment environments and companies that control proprietary data. In that setup, microagi is a winner if exclusive data relationships with frontier AI labs and early factory deployments reinforce each other, creating a feedback loop where better data improves deployment outcomes and successful deployments generate more proprietary data [Semafor, July 2026] [The Next Web, July 2026]. The likely loser in that case would be generic systems integrators that lack proprietary data assets and rely on third-party models alone. The reverse scenario is also easy to see: if incumbent robotics suppliers absorb data capture and model adaptation into their own implementation stack, microagi's position could narrow into a services-heavy layer with weaker pricing power.
Opportunity
PUBLIC
The prize here is unusually large because microagi is not trying to sell a single robot, it is trying to become part of the data and deployment layer that determines which robots can work reliably in real factories and, potentially, in other physical environments [The Next Web, July 2026] [microagi, Unknown].
The headline opportunity is straightforward. If microagi can keep proving that it can capture task-specific human work data, adapt existing robotics models to that work, and support deployment on customer production lines, it could become a default enablement layer for industrial embodied AI in Europe and beyond [The Next Web, July 2026] [SiliconANGLE, July 2026]. That outcome is still early, but it is not purely aspirational. The company has already raised a $55 million seed round led by Hummingbird with participation from Northzone, LocalGlobe, Village Global, and redalpine, and that financing was widely reported as the largest seed round raised by a German company, which suggests investors see a category-sized opening rather than a narrow tooling business [SiliconANGLE, July 2026] [Semafor, July 2026] [The Next Web, July 2026]. Public reporting also points to five companies collecting data through Atlas as of July 2026, with one customer preparing a factory-line deployment, which matters because the core claim lives or dies in live operating environments rather than lab demos [The Next Web, July 2026].
A reasonable way to frame the upside is through a few distinct paths rather than a single outcome.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Factory data system of record | Atlas becomes the standard workflow for collecting task data and tuning robot behavior inside manufacturers in automotive, logistics, and food | A visible production-line deployment that expands from one task into multiple workflows at the same customer | microagi is already positioned around the training-data and deployment bottleneck, and public reporting says five companies were using Atlas for data collection as of July 2026, with one preparing a line deployment [The Next Web, July 2026] |
| Dual-market platform | The company sells deployment infrastructure to manufacturers while also monetizing proprietary datasets through licensing to advanced AI labs | Continued demand from frontier labs for exclusive real-world embodied data | Semafor reported that microagi has licensed some of its data exclusively to frontier AI labs, while its industrial product work continues through Atlas [Semafor, July 2026] [The Next Web, July 2026] |
| Embodied AI ecosystem node | microagi becomes an important European integration point between model providers, cloud compute partners, and factory operators | Deeper technical partnerships that reduce deployment friction and improve model-training throughput | Tech.eu reported a compute collaboration involving Google Cloud and NVIDIA, which supports the idea that the company is building around an ecosystem position rather than a standalone robotics SKU [Tech.eu, July 2026] |
The compounding mechanism, if it works, is a data-and-deployment flywheel. Each customer engagement should produce more task-specific data, sharper fine-tuning for robotics models, and more confidence that similar tasks can be deployed at the next site, especially if Atlas is in the loop from collection through factory-floor reinforcement [LinkedIn, Unknown] [The Next Web, July 2026]. The second layer of compounding is cross-market. Semafor's reporting that some datasets have been licensed to frontier AI labs suggests microagi may be able to monetize the same core asset, real-world physical task data, across both industrial deployments and model-development demand [Semafor, July 2026]. The consumer-facing Shift effort, as described in reporting around free home cleaning in New York City and first-person task capture, points to a broader ambition to widen the data funnel beyond factories, although that expansion should still be treated cautiously given the lighter verification base [Business Insider, June 2026] [dealroom.co, Unknown] [ai-market-watch.com].
The size of the win is best framed as scenario analysis, not a forecast. A credible public-market or M&A comparable is not established in the source set, and no verified market-size dataset was provided, so the cleanest public read is strategic rather than purely numerical. If the "factory data system of record" scenario plays out, microagi could plausibly become a high-value infrastructure company sitting between robot model makers and large industrial operators, with value driven by recurring software, deployment services, and proprietary data rights (scenario, not a forecast) [The Next Web, July 2026] [Semafor, July 2026]. If the "dual-market platform" scenario proves durable, the upside expands further because the same dataset could support both factory automation and external licensing demand, which typically supports stronger strategic value than a single-channel robotics integrator model (scenario, not a forecast) [Semafor, July 2026] [microagi, Unknown]. The public evidence is still early, but it is early in a way that is consistent with outsized outcomes: unusually large financing, a narrow initial wedge, and first signs that customers and AI labs may both value the same underlying asset.
Lightly corroborated -- Relies primarily on The Next Web, SiliconANGLE, Semafor, Tech.eu, and the company website; core opportunity claims are partially corroborated, but several expansion signals remain early or company-adjacent.
Sources
Public sources
[microagi] microagi | https://www.microagi.ai/
[Tech.eu, July 2026] Google Cloud and NVIDIA power microagi's embodied-AI ambitions | https://tech.eu/2026/07/22/google-cloud-and-nvidia-power-microagis-embodied-ai-ambitions/
[SiliconANGLE, July 2026] Microagi nabs $55M to teach factory robots how to work | https://www.siliconangle.com/2026/07/16/microagi-nabs-55m-teach-factory-robots-work/
[Crunchbase] microagi - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/microagi
[The Next Web, July 2026] An F1 aerodynamicist just raised $55m to teach factory robots | https://thenextweb.com/news/microagi-55m-seed-robot-training-data-shift
[Business Insider, June 2026] This Startup Will Clean Your Apartment for Free, If AI Can Watch | https://www.businessinsider.com/this-startup-will-clean-your-apartment-free-ai-can-watch-2026-6?op=1
[RobotToday] Microagi - RobotToday | https://robottoday.com/suppliers-discovery/microagi
[LinkedIn, Retrieved 2026] Anton Poletaev - microagi | LinkedIn | https://www.linkedin.com/in/anton2049/
[Semafor, July 2026] Robot-data startup raises largest seed round ever in Germany | https://www.semafor.com/article/07/16/2026/robot-data-startup-microagi-raises-largest-seed-round-ever-in-germany
Articles about microagi
- Five Companies Are Already Feeding Data Into MicroAGI's Factory Robot Pipeline — The Munich-based startup, founded by former Formula One engineers, uses cameras and sensor gloves to build datasets for robotic deployment.