What are the fundraising trends in the physical AI market?
In our updated market reports, you will find everything you need
SUMMARY
We analyzed every publicly disclosed equity round raised by pure-play physical AI companies between January 1, 2024 and July 2, 2026. We only kept rounds of $300K or more and excluded companies where robotics, embodied AI, robot foundation models, or physical-world AI systems were not more than 80% of the business.
Over this period, fundraising in the physical AI market has accelerated sharply. The dataset contains 65 disclosed deals across 59 unique companies, with 25 of them landing in year-to-date 2026 for a combined $7.38B.
Physical AI has moved from an emerging robotics funding theme into an infrastructure-scale capital race. YTD 2026 funding is already almost double the full-year 2025 total of $3.77B and more than triple full-year 2024’s $2.24B.
The market is still highly concentrated, but less dependent on a single company than many other frontier categories. In YTD 2026, the largest deal accounts for 19.0% of all capital, while the top 3 deals account for 45.0%. That is concentrated, but far less extreme than a market driven by one single whale round.
Round sizes in physical AI are unusually large. The YTD 2026 median round is $165M and the average round is $295.3M, which means even the “typical” disclosed physical AI financing is now far above normal venture scale.
Humanoid Robots still lead the physical AI market by dollars, with 43.2% of YTD 2026 capital. But Robotic Foundation Models are gaining importance, taking 28.5% of capital from only 16.0% of deals.
General Purpose Robots are the second-largest category by deal count, with 6 of 25 YTD 2026 deals. But they capture only 19.9% of dollars, showing that deployment-oriented robotics companies are active but not priced as richly as humanoid or model-layer platforms.
North America remains the largest capital pool in physical AI, with 58.7% of YTD 2026 dollars. But Asia-Pacific and Europe together now account for 41.4%, making the market far more geographically balanced than it was in 2024.
First financings have almost disappeared from the physical AI market. Only 2 of 25 YTD 2026 deals are first financings, and they represent just 5.8% of capital. The market is now mostly recapitalizing companies that already passed an earlier credibility screen.
Repeat strategic investors are becoming central to physical AI. NVIDIA/NVentures appears in four YTD 2026 deals, Bezos Expeditions in three, and several corporate or industrial investors show up repeatedly across compute, manufacturing, robotics, and deployment-channel rounds.
Is more or less capital going into the physical AI market?
Much more capital is going into the physical AI market. Year-to-date 2026 has already produced about $7.38B across 25 disclosed deals, compared with $1.04B across 12 deals over the same year-to-date period in 2025 and $1.28B across 8 deals over the same period in 2024.
That means YTD 2026 physical AI funding is roughly 7.1x the comparable 2025 period and 5.8x the comparable 2024 period. This is not a marginal acceleration. It is a structural repricing of robotics, embodied AI, and robot foundation models as a major frontier-AI infrastructure category.
The comparison looks even stronger against full-year history. The physical AI market raised $2.24B in all of 2024 and $3.77B in all of 2025. By July 2, 2026, it had already reached $7.38B, meaning the market has more than doubled its prior full-year peak before the year is even complete.
The important nuance is that this is not simply one company distorting the picture. The largest YTD 2026 round is $1.4B, but excluding the largest deal still leaves $5.98B of capital. Excluding the top 3 deals still leaves $4.06B, which is already larger than full-year 2025.
The practical rule for reading physical AI funding right now is simple: the boom is real, but it is concentrated in companies that investors believe can become physical-world AI infrastructure. This is not a broad robotics funding recovery. It is a platform-selection cycle.
Is physical AI funding driven by more deals or larger rounds?
Physical AI funding in 2026 is being driven by both more deals and much larger rounds, but larger rounds are doing most of the work. Deal count rose from 12 in YTD 2025 to 25 in YTD 2026, a little more than doubling. Capital rose from $1.04B to $7.38B, more than a sevenfold increase.
The median round size tells the story even more clearly. YTD 2025 had a median physical AI round of $61M. YTD 2026 has a median round of $165M. The average round also jumped from $86.4M to $295.3M.
This is not a normal venture market where a few large rounds sit above a broad base of small financings. In YTD 2026, 21 of 25 deals are $50M or larger, and 18 are above $100M. That means 72% of all disclosed physical AI deals are already nine-figure rounds.
Compared with 2024, the shift is even more dramatic. Full-year 2024 had 15 deals and $2.24B of capital, with a $100M median round. Full-year 2025 had 25 deals and $3.77B, also with a $100M median round. YTD 2026 keeps the 25-deal pace but lifts the median to $165M and total capital to $7.38B.
For deeper benchmarks on physical AI deal sizes, medians, and round distributions, see our physical AI market deck. It breaks the funding cycle down with category, stage, geography, and investor-level context.
Is physical AI capital moving toward later-stage or earlier-stage companies?
Physical AI capital is moving decisively toward later-stage companies. In YTD 2026, Series B, Series C, and Growth Equity rounds captured $4.92B, or 66.7% of all capital. Unknown-stage rounds added another $1.05B, much of it also tied to companies that were clearly not new formations.
By deal count, the market is also follow-on heavy. Series B alone accounts for 9 of 25 deals, and Series C accounts for another 3. Seed has only 1 qualifying deal in the YTD 2026 dataset.
This is a major change from the YTD 2025 window, when Seed deals represented 50% of deal count and first financings represented 50% of deal count. The market began 2025 with formation and experimentation, but by 2026 it had shifted toward recapitalizing companies with existing teams, platforms, partners, and product roadmaps.
The caveat is that stage labels in physical AI are less useful than in software. A Series A in this market can be $320M, and a Seed can be $26M or even $320M if the company is attacking a simulation or data bottleneck. The better distinction is not Seed versus Series C. It is whether the company owns a credible robot body, robot brain, data loop, or deployment channel.
Is the physical AI market maturing or still experimental?
The physical AI market is maturing quickly by capital behavior, but it is still experimental by proof quality. The financing side now looks mature: 25 YTD 2026 deals, $7.38B of capital, a $165M median round, and 18 rounds above $100M.
But the product side is not yet equally mature. Many companies are still selling a roadmap toward general-purpose robots, transferable robot intelligence, or scalable physical-world autonomy. The funding market is underwriting platform potential before the category has produced widespread evidence on uptime, task success rates, unit economics, safety, and repeat deployments.
The strongest maturation signal is that capital is moving to companies with prior validation. Follow-on rounds represent 92% of deals and 94.2% of capital in YTD 2026. Investors are no longer funding many brand-new physical AI companies from scratch. They are doubling down on teams that already have some combination of technology, investor syndicate, customer channel, data access, or strategic partner support.
The strongest experimental signal is that the market is still paying for generalization before generalization has been fully proven. Humanoids, robotic foundation models, simulation training, and robot operating layers are being funded as if physical AI will become a scalable platform layer. That may be right, but the proof will come from deployments, not valuations.
Are new startups still entering the physical AI market?
New startups are still entering physical AI, but much less visibly than in 2025. Only 2 of 25 YTD 2026 deals are first financings, compared with 6 of 12 deals in the comparable YTD 2025 period.
The first-financing share fell from 50.0% of YTD 2025 deals to 8.0% of YTD 2026 deals. Capital going to first financings fell from 18.8% to 5.8%. That is a sharp signal that the market has moved from company creation toward recapitalization.
The two YTD 2026 first financings are also revealing. Lyte is an AI robot component and perception-infrastructure company, while General Intuition is a simulation and training-data platform. Neither is simply another full-stack humanoid manufacturer. New-company formation is still fundable, but it appears more fundable around bottleneck layers than around launching another capital-intensive robot-body company.
This matters because the entry bar in physical AI has become very high. A credible company now needs hardware depth, data strategy, safety credibility, manufacturing access, deployment partners, and AI talent. That makes pure startup formation harder, even while the category attracts more capital overall.
For the broader category view across new entrants, follow-ons, and subcategory formation, see our physical AI market deck. It gives more context on which company types are actually entering the market.
Are more investors entering the physical AI market?
More investors are entering physical AI, and the investor base is becoming broader and more strategic. YTD 2026 includes approximately 105 named investors, compared with 67 disclosed investors across full-year 2025 and 69 across full-year 2024.
The tier-1 investor count has also expanded. YTD 2026 has 36 unique tier-1 investors, versus 28 in full-year 2025 and 29 in full-year 2024. This is no longer a small robotics-specialist funding market. It now includes frontier AI funds, corporate strategics, sovereign-linked capital, industrial companies, semiconductor players, and global growth investors.
The repeat-investor pattern is especially important. NVIDIA/NVentures appears in four YTD 2026 deals, Bezos Expeditions in three, and Bain Capital Ventures, 8VC, Eclipse, Incharge Capital, Salesforce Ventures, HSG, IDG Capital, and Amazon-linked capital each appear more than once.
What this means is that physical AI is becoming a strategic-capital market, not just a financial-VC market. Investors are not only buying equity exposure. They are trying to secure influence over compute demand, industrial automation, robot platforms, data loops, manufacturing channels, and future deployment infrastructure.
Are top investors getting more or less active in physical AI?
Top investors are getting more active in physical AI, especially those connected to compute, frontier AI, industrial deployment, and strategic infrastructure. The clearest signal is NVIDIA/NVentures, which appears in four YTD 2026 qualifying deals.
Bezos Expeditions is another strong repeat signal, appearing in Skild AI, Generalist AI, and General Intuition. That pattern is not random. It shows repeated exposure to the robot-intelligence and data-layer side of the market, not just one-off interest in humanoid hardware.
The top investor list in 2024 was already strong, with Lux Capital, Samsung NEXT, OpenAI, Bezos Expeditions, EQT Ventures, Microsoft, NVIDIA, Thrive, Khosla, Sequoia, General Catalyst, and BOND all appearing more than once. In 2025, repeat investors included CRV, First Round, Alibaba, Geely, Khosla, NVentures, Intel Capital, and LG Technology Ventures.
By 2026, the repeat investor base looks more strategically aligned. The recurring names are less like general venture tourists and more like actors that benefit directly if physical AI scales: semiconductor companies, AI infrastructure backers, industrial corporates, sovereign-linked capital, and deployment-channel investors.
Which physical AI subcategories are gaining momentum?
Robotic Foundation Models, Humanoid Robots, and General Purpose Robots are the main physical AI subcategories gaining momentum. Together, they account for 91.6% of YTD 2026 capital and 76.0% of deals.
Humanoid Robots remain the largest category by both dollars and deal count. They captured $3.19B across 9 YTD 2026 deals, or 43.2% of capital and 36.0% of deal count. That is less dominant than YTD 2025’s 57.6% capital share, but the absolute dollar figure is far larger.
Robotic Foundation Models are gaining the most on a capital-efficiency basis. They captured 28.5% of capital from only 16.0% of deals, producing the highest capital-share-to-deal-share ratio in the market at 1.78. Investors are clearly treating the “robot brain” layer as one of the most valuable control points.
General Purpose Robots are also scaling. They account for 6 YTD 2026 deals and $1.47B of capital, helped by large financings for Bedrock Robotics, Mind Robotics, RoboForce, RobCo, and Pudu Robotics. The category is less richly valued than humanoids or foundation models, but it is producing real deal volume and large checks.
We cover this subcategory shift in more detail in our physical AI market report, including how robot bodies, robot brains, components, tooling, fleet management, and simulation platforms are moving at different speeds.
Which physical AI subcategories are losing momentum?
Developer Tools for Robotics, Robot Fleet Management, Embodied AI Software, and AI Robot Components are losing momentum in relative capital terms, even though they remain strategically important. Together, those four categories represent 20.0% of YTD 2026 deals but only 3.1% of capital.
That does not mean these layers are unimportant. In fact, robot components, deployment tools, embodied software, and fleet systems are essential bottlenecks. But the funding market is not yet pricing them like platform winners.
The contrast with 2024 is useful. Bright Machines and T-robotics gave Developer Tools for Robotics 5.0% of full-year 2024 capital. In YTD 2026, Vention is the only developer-tools deal, and the category falls to 1.5% of capital.
Robot Fleet Management also remains small. Formic raised in 2024, Starship raised in 2025, and Gather AI raised in 2026, but the category has not yet produced the kind of $300M-plus platform round that humanoids, robotic foundation models, simulation training, and general-purpose robots have produced.
The likely reason is narrative control. Investors are currently rewarding companies that claim to own the body, the brain, or the training data layer. Tools and fleet software may become critical later, but for now they are being treated as enabling infrastructure rather than the main prize.
Which regions are gaining momentum in physical AI funding?
Europe and Asia-Pacific are gaining the most momentum in physical AI funding, even though North America remains the largest region by dollars. In YTD 2026, North America captures 58.7% of capital, Asia-Pacific captures 21.0%, and Europe captures 20.3%.
Europe’s jump is especially striking. Europe had only 5.7% of full-year 2024 capital and 5.3% of full-year 2025 capital. In YTD 2026, it reaches 20.3%, largely because NEURA Robotics announced a Series C of up to $1.4B and RobCo raised $100M.
Asia-Pacific is gaining through breadth. APAC has 9 of 25 YTD 2026 deals, or 36.0% of deal count, across companies such as LimX Dynamics, AI² Robotics, RLWRLD, Galbot, RobotEra, X Square Robot, Pudu Robotics, and WIRobotics.
The difference between Europe and Asia-Pacific is important. Europe’s YTD 2026 position depends heavily on one giant NEURA round. Asia-Pacific’s position comes from a wider bench of humanoid and embodied-AI companies. Europe looks more concentrated; APAC looks more competitive.
For ongoing regional tracking across North America, Europe, and Asia-Pacific, see our physical AI market deck. It follows both where companies are founded and where strategic capital is showing up.
Which regions are losing momentum in physical AI funding?
Latin America, the Middle East, and Africa remain absent from the disclosed physical AI dataset under this strict definition. Across 2024, 2025, and YTD 2026, none of the qualifying disclosed equity rounds came from those regions.
That does not mean there is no robotics activity in those markets. It means there were no visible pure-play, publicly disclosed, equity-funded physical AI rounds above $300K that met the category definition and source-quality threshold.
North America is not losing momentum in absolute terms, but it is losing share. It had 88.1% of full-year 2024 capital, 61.2% of full-year 2025 capital, and 58.7% of YTD 2026 capital. The region is still dominant, but the market is becoming less exclusively North American.
The region most at risk of being overstated is Europe. Its YTD 2026 capital share looks strong, but it is mostly driven by NEURA. Without that one round, Europe would still look like a smaller physical AI funding market compared with North America and Asia-Pacific.
Is physical AI becoming more global or regionally concentrated?
Physical AI is becoming more global by deal count and somewhat more balanced by capital, but it is still concentrated in three regions: North America, Asia-Pacific, and Europe. No qualifying disclosed deals appear in Latin America, the Middle East, or Africa across the dataset.
North America’s share of capital has fallen from 88.1% in 2024 to 61.2% in 2025 and 58.7% in YTD 2026. That is a meaningful decline in concentration, even though North America remains the largest region.
Asia-Pacific has become the clearest challenger by company count. It produced 1 deal in 2024, 10 deals in 2025, and 9 deals already in YTD 2026. The APAC physical AI ecosystem is no longer a minor part of the dataset.
Europe is more volatile. It had 3 deals in 2024, 3 deals in 2025, and only 2 deals in YTD 2026, but the NEURA round makes it a major capital contributor. In other words, Europe is globalizing by deal quality more than by deal frequency.
The best interpretation is that physical AI is globalizing at the platform-contender level, but not yet at the long-tail startup level. The market still clusters around places with deep AI talent, robotics engineering, manufacturing access, compute capital, and strategic industrial investors.
Is physical AI capital moving toward proven winners or new opportunities?
Physical AI capital is moving overwhelmingly toward proven winners. In YTD 2026, follow-on rounds represent 92.0% of deals and 94.2% of capital. First financings are only 8.0% of deal count and 5.8% of dollars.
This is the opposite of the YTD 2025 market, where first financings represented 50.0% of deals and 18.8% of capital. The market has shifted from formation to selection.
Several companies now appear repeatedly across the multi-year dataset, including Figure AI, Skild AI, Physical Intelligence, Unitree Robotics, NEURA Robotics, Apptronik, Galbot, RobotEra, EngineAI, Dyna Robotics, RLWRLD, RoboForce, and Mind Robotics. Repeat fundraising is becoming the clearest survival signal.
The practical takeaway is that the market is no longer asking, “Can someone build physical AI?” It is asking, “Which existing companies can compound data, hardware, deployment, and manufacturing capacity fast enough to become platforms?”
Our physical AI market report tracks these repeat raisers over time, with more detail on which companies keep attracting follow-on capital and which still need to prove they can raise again.
Is the physical AI market becoming winner-takes-most?
The physical AI market is winner-takes-most by capital allocation, but not winner-takes-all. The top 10 YTD 2026 deals account for 79.1% of capital, while the bottom half of deals account for only 13.3%.
At the same time, the single largest deal accounts for only 19.0% of YTD 2026 capital. That is much less concentrated than YTD 2025, when the largest deal accounted for 33.8%, and YTD 2024, when Figure’s $675M round accounted for 52.8%.
The market is therefore becoming less dependent on one company but still extremely concentrated around a small group of platform contenders. Skild AI, NEURA Robotics, Apptronik, Mind Robotics, Generalist AI, Galbot, General Intuition, X Square Robot, Bedrock Robotics, and Sunday Robotics shape most of the capital story.
This distinction matters. Physical AI is not a single-company bubble. It is a multi-company platform race. But it is also not a broad-based market where the average startup meaningfully moves the funding total.
Is the next wave of physical AI winners becoming visible?
The next wave of physical AI winners is becoming visible, but mostly among companies that have already raised before. The YTD 2026 dataset is dominated by follow-ons, which means investors are selecting from an existing field rather than creating many new contenders.
The clearest candidate groups are humanoid platforms, robotic foundation models, and industrial general-purpose robot companies. These categories have the largest rounds, the strongest strategic investors, and the clearest claims to cross-customer scalability.
But the next winners will not be determined by funding size alone. The dataset already contains many companies with $100M-plus rounds. The differentiator will be evidence quality: robot operating hours, task success rates, deployment repeatability, safety performance, customer expansion, manufacturing readiness, and data-loop compounding.
The most interesting new entrants are not necessarily full robot companies. Lyte and General Intuition suggest that new winners may also emerge from bottleneck layers: perception foundations, simulation data, training environments, and systems that help robots generalize across real-world tasks.
For more context on the new cohort of physical AI startups and the signals that separate durable companies from one-off experiments, see our physical AI market deck.
Is the physical AI funding landscape fragmenting or consolidating?
The physical AI funding landscape is consolidating by capital and fragmenting by investor base. On the capital side, a small set of platform companies absorbs most dollars. On the investor side, the number and type of investors involved keeps expanding.
YTD 2026 has approximately 105 named investors, far more than the 67 disclosed investors in full-year 2025. But the top 10 deals still capture 79.1% of capital. That means more investors are entering the market, but the funding outcome remains concentrated around a few large rounds.
Category-level fragmentation is also increasing. In 2024, the market was dominated by Humanoid Robots, Robotic Foundation Models, and General Purpose Robots, with small contributions from developer tools, fleet management, and simulation. By YTD 2026, all eight categories have at least one deal.
The right way to describe the current state is selective expansion. The physical AI market is broadening across categories and investors, but the big dollars are consolidating around companies that can credibly claim platform control over robot bodies, robot intelligence, deployment data, or training infrastructure.
Where is investor attention shifting in physical AI?
Investor attention in physical AI is shifting toward three areas: robot foundation models, humanoid platforms with deployment paths, and data or simulation layers that can improve robot generalization.
Robotic Foundation Models are the clearest shift. The category captured $770M in full-year 2024, $783.3M in full-year 2025, and $2.10B in YTD 2026 alone. It now has the highest capital-share-to-deal-share ratio in the market.
Humanoids remain central, but investors appear to care less about the humanoid form factor by itself and more about whether the company can build a data flywheel. Apptronik, NEURA, Galbot, Sunday Robotics, RobotEra, LimX, and AI² Robotics are not just selling robot bodies; they are selling the idea of embodied intelligence that improves with deployment.
The third shift is toward enabling data layers. General Intuition, Lyte, Trener Robotics, Algorized, Vention, and Gather AI show investors looking at the perception, simulation, programming, deployment, and fleet layers that make physical AI usable outside demos.
The broader read is that investor attention is moving from “robots are back” to “which layer of the physical AI stack becomes the control point?” That is a much more strategic question, and it explains why corporate, compute, industrial, and sovereign-linked capital are increasingly visible.
For real-time tracking of how investor attention is moving across humanoids, robot brains, components, simulation, developer tools, and deployment infrastructure, see our physical AI market report.
INSIGHTS
The insights below come from reviewing every disclosed equity round in the physical AI market between January 1, 2024 and July 2, 2026.
- Physical AI funding has crossed from robotics venture capital into infrastructure-style underwriting. The YTD 2026 median round is $165M, and 72% of deals are above $100M, so normal seed-to-Series-A venture heuristics no longer describe the category well.
- The market is not being distorted by one outlier. Excluding the largest YTD 2026 round still leaves $5.98B of capital, which is larger than full-year 2025 and more than twice full-year 2024.
- Humanoids are still the largest category, but their dominance is weakening in share terms. Humanoids were 57.6% of YTD 2025 capital and 43.2% of YTD 2026 capital, while robotic foundation models rose from 15.3% to 28.5%.
- Robotic Foundation Models have become the highest-premium category in YTD 2026. They capture 28.5% of capital from 16.0% of deals, giving them the strongest capital-share-to-deal-share ratio at 1.78.
- The clearest funding premium goes to companies that can claim a compounding data loop. The largest rounds repeatedly reference robot foundation models, real-world training data, simulation, perception, robot gyms, or industrial deployment data.
- Physical AI is increasingly a follow-on market. First financings fell from 50.0% of YTD 2025 deals to 8.0% of YTD 2026 deals, which means the investor question has shifted from “what can be started?” to “which platforms deserve recapitalization?”
- New-company formation is more credible at bottleneck layers than at robot-body layers. The two YTD 2026 first financings, Lyte and General Intuition, are perception/data/simulation plays rather than new humanoid manufacturers.
- Stage labels are unusually weak indicators in physical AI. A Series A can be $320M and a Seed can be $26M, so diligence should focus more on data access, deployment evidence, manufacturing readiness, and strategic syndicate quality than nominal stage.
- North America remains the largest capital market, but its dominance is declining. Its capital share moved from 88.1% in 2024 to 61.2% in 2025 and 58.7% in YTD 2026.
- Asia-Pacific is the broadest challenger to North America by company count. It produced 10 deals in 2025 and 9 deals in YTD 2026, suggesting a deeper competitive bench than Europe despite lower average round size.
- Europe’s YTD 2026 surge is real but fragile. Europe reaches 20.3% of capital mostly because of NEURA’s $1.4B round; without that one event, Europe would remain a much smaller contributor.
- Strategic investors matter more in physical AI than in most software markets. NVIDIA, Amazon, Google, Qualcomm, Bosch, Mercedes-Benz, John Deere, Salesforce, Alibaba, Tencent, Meituan, Geely, Baidu, and industrial or sovereign-linked backers appear because commercialization depends on compute, manufacturing, data, and deployment channels.
- Funding size alone has become a weak validation signal. With 18 YTD 2026 rounds above $100M, the more useful test is whether the company can show real deployments, high uptime, repeat customers, safe operation, and a defensible data flywheel.
- Robot Fleet Management remains undercapitalized relative to its practical importance. The category has visible deals across years but still only 0.5% of YTD 2026 capital, implying investors prefer upstream platform narratives over operational fleet software.
- Simulation is strategically important but unevenly funded as a standalone category. Vsim appeared in 2024 and General Intuition in 2026, but much of the market still funds simulation inside larger robot or model companies rather than as a separate tooling layer.
- General Purpose Robots generate meaningful deal volume but lower capital premium. They are 24.0% of YTD 2026 deals and 19.9% of capital, suggesting investors see deployment value but less winner-takes-most upside than in humanoids or robotic foundation models.
- The market’s most important split is no longer hardware versus software. It is body, brain, data, and deployment loop. Companies that control more than one of those layers are likely to command larger rounds.
- YTD 2026 is more globally balanced than earlier years, but still not geographically broad. All qualifying capital is concentrated in North America, Europe, and Asia-Pacific, with no disclosed pure-play rounds in Latin America, the Middle East, or Africa.
- The strongest recurring investor signals in YTD 2026 are NVIDIA/NVentures and Bezos Expeditions. Their repeated presence across the robot-intelligence stack suggests frontier AI capital is trying to shape the physical-world AI control layer before deployment volumes are proven.
- The biggest unresolved question in physical AI is whether megaround-funded platforms can convert capital into reliable real-world performance. The market has already validated the financing narrative; it has not yet fully validated mass deployment economics.
OUR METHODOLOGY TO BUILD THIS TRACKER
We built this physical AI funding tracker by reviewing every publicly disclosed equity round raised by pure-play physical AI companies between January 1, 2024 and July 2, 2026. A company counts as pure-play when more than 80% of its activity is tied to AI systems, robots, robotic software, robotic foundation models, simulation, components, or robot fleet infrastructure that can perceive, decide, and act in the physical world.
We applied four filters to build the dataset. First, we only included equity rounds, so grants, debt, acquisitions, IPOs, SPAC transactions, and merger events are excluded. Second, we only counted rounds of $300K or more. Third, we only kept companies that fit one of the defined physical AI categories: Humanoid Robots, General Purpose Robots, Robotic Foundation Models, Embodied AI Software, Simulation Training Platforms, Robot Fleet Management, AI Robot Components, or Developer Tools for Robotics. And fourth, every entry had to be confirmed by a direct company announcement, a press release, a tier-1 media report, or a credible robotics-specialist source, with the source URL preserved for every row.
We also excluded adjacent cases that would have distorted the picture: autonomous vehicles, defense drones, surgical-only robotics, generic industrial automation, generic AI software, broad manufacturing SaaS, debt-only facilities, grants, undisclosed-size rounds, and companies where robotics or physical-world AI did not appear to exceed the 80% business threshold. The final dataset contains 65 disclosed deals across 59 unique companies, and every average, median, share, and concentration ratio is computed on that disclosed sample. Privately raised rounds that were never publicly announced are necessarily missing, and China/Korea private-market reporting is less standardized, which is a known limitation of any public-only physical AI funding tracker.
Related blog posts
- Which startups have raised the most funding in the physical AI market?
- Which startups are the most valued in the physical AI market?
- How strong is fundraising in the physical AI market right now?
