When will the robotics bull run start?
In our robotics market deck, you will find everything you need to understand the market
SUMMARY
The robotics bull run has already started in venture capital, but the broader commercial and public-market phase is most likely to become obvious in 2027–2028.
The strange part is how far capital has run ahead of deployments. Robotics startups had already raised $18.8 billion by late June 2026, while North American robot unit orders were growing only 2% in the first half.
Humanoid production is no longer the bottleneck it was a year or two ago. More than 22,000 humanoids shipped globally in the first half of 2026, but manufacturing and logistics still accounted for only about 18% of that volume.
That makes shipment mix more important than headline shipments. A 50,000-unit humanoid market dominated by entertainment, research and data collection would be less commercially meaningful than a smaller market where factories and warehouses keep reordering robots.
The most important threshold has already been crossed: humanoids are doing real work for real customers. Figure at BMW, Digit at GXO and UBTECH's industrial deployments show that the category has moved beyond stage demos, even though customer fleets remain small.
Traditional robotics is giving the boom a much stronger base than the humanoid narrative alone suggests. Industrial robots, cobots, autonomous mobile robots and specialized warehouse systems already operate at million-unit scale, and demand is broadening beyond automotive.
China has the clearer manufacturing advantage today. It installs more industrial robots than any other country, has a huge local component ecosystem and dominates current humanoid shipment rankings.
The US is making a different bet: that the most valuable part of robotics will be the intelligence layer. Figure, Skild AI, Apptronik, Nvidia and other US-linked companies are pouring capital into models, compute, simulation and physical-data systems that could improve whole fleets through software.
Robot economics only need to work on a narrow set of repetitive jobs to create a very large market. Warehouses and factories are the obvious starting point because labor costs are measurable, utilization can be high and the same movements repeat thousands of times.
The main risk is not that robotics fails. It is that useful deployment takes several years longer than valuations assume. If factories can produce humanoids faster than customers can find profitable jobs for them, the sector could have a serious correction even while the underlying technology keeps improving.
The cleanest proof of a durable bull run will be repeat orders and customer fleet expansion. Hundreds of robots at the same customer, a majority of humanoid shipments going into productive work, and several public robotics companies posting sustained high revenue growth would change the story from speculation to commercial scale.
Our base case is therefore a three-stage market: the speculative robotics bull run is already here, the commercial bull run can broaden sharply in 2027–2028, and mass adoption comes later. The next two years are when today's huge valuations should either start to look prescient or very premature.
This market map, featured in our robotics market deck, highlights top companies and startups in the robotics market
Has the robotics bull run already started?
The robotics bull run has already started in venture capital, while robot deployments are still catching up.
The gap is unusually large. Crunchbase counted $18.8 billion of global robotics startup funding by late June 2026. That was already more than the $15 billion raised during all of 2025 and the $14.1 billion raised during the 2021 venture peak. Investors did not wait for mass adoption.
The size of individual rounds tells the same story. Skild AI raised $1.4 billion at a valuation above $14 billion in January 2026. Apptronik's Series A eventually reached $935 million, taking the company's valuation to about $5.3 billion. Figure had already raised more than $1 billion at a $39 billion post-money valuation.
Real robot demand looks much calmer. North American companies ordered 17,995 robots worth $1.17 billion during the first half of 2026, according to the Association for Advancing Automation. Unit orders grew only 2% from the previous year, while spending increased 6.6%.
So the financial bull market is already here. The harder question is when customers start buying robots fast enough to justify it.
If you want more recent data on this point, please see our latest robotics market report.
Why is everyone suddenly betting on robotics now?
Robotics is attracting so much money because AI has started attacking the hardest part of the robot problem: teaching machines what to do when the environment changes.
Old industrial robots can be incredibly productive, but they usually work inside carefully designed environments. Change the object, workstation or task and somebody may need to reprogram the system.
The new robotics companies are trying to make learning much more reusable. Figure's Helix system combines vision, language and physical control. Skild AI is developing one general-purpose robot brain that can work across different machines. Nvidia's GR00T platform gives robot developers foundation models, simulation tools, synthetic data and computing infrastructure.
That changes the potential economics. Better software can improve an entire fleet after the robots have already been manufactured. A warehouse robot that becomes more capable through software updates starts to look very different from an old piece of factory machinery whose capabilities were mostly fixed when it was installed.
Investors are clearly paying for that possibility today. Nvidia, Google, SoftBank, Amazon-linked funds, Mercedes-Benz, Salesforce and other large technology or industrial groups have all put substantial money into physical AI companies.
We still do not know how general these robot models will become. But robotics suddenly has the same software-scaling story that drove the earlier AI boom, and money is arriving much faster than revenue.
As this chart shows, and as featured in our robotics market deck, search interest in robot costs has increased significantly
Are humanoid robot shipments finally exploding?
Humanoid robot shipments are exploding, although most of those robots still are not doing factory or warehouse jobs.
Counterpoint Research estimated that more than 22,000 humanoids shipped worldwide during the first half of 2026, nearly 300% more than a year earlier. It expects annual shipments to exceed 50,000 units. AGIBOT alone shipped about 9,700 robots and captured more than 43% of the market.
That is a real manufacturing jump. A market that was recently measured in prototypes and hundreds of units can now produce tens of thousands.
The catch appears when we look at where the robots went. Entertainment and performance together with data production and research still represented more than 60% of first-half shipments. Intelligent manufacturing accounted for 13%, and warehousing and logistics another 5%.
That 18% combined share for manufacturing and logistics is the number we care about most. These are the environments where humanoids can start replacing measurable amounts of labor and where customers can calculate a return on investment.
Shipments could therefore pass 50,000 this year without creating a 50,000-robot workforce. Still, the mix is gradually moving toward productive applications, and Counterpoint expects more industrial pilots to enter larger deployment phases during the second half.
| Humanoid metric | Current picture | What we think it means |
|---|---|---|
| H1 2026 shipments | >22,000 | Manufacturing scale is appearing |
| Year-on-year growth | Nearly 300% | Adoption is accelerating from a tiny base |
| Manufacturing share | 13% | Real industrial use is growing |
| Warehousing/logistics share | 5% | Commercial use remains early |
| Entertainment + data/research | >60% | Most shipments still do not prove labor economics |
| Expected 2026 shipments | >50,000 | Six-figure annual volume is getting much closer |
Are humanoid robots actually working in factories now?
Humanoid robots are doing real factory and logistics work today, and Figure, Agility Robotics and UBTECH have moved well beyond one-off stage demos.
Figure gives us one of the cleanest examples. Figure 02 operated at BMW's Spartanburg factory on 10-hour shifts, Monday through Friday. Over the deployment, Figure reported more than 1,250 operating hours, more than 90,000 parts loaded and involvement in the production of more than 30,000 BMW X3 vehicles.
BMW has since brought Figure 03 into the same plant for a more complicated logistics sequencing task. The robot needs to manipulate components while repositioning its body and pulling a wheeled cart. That progression is useful because BMW is moving from a tightly defined pick-and-place operation toward work requiring more whole-body coordination.
Agility Robotics has produced another meaningful data point. Its Digit robot has moved more than 100,000 totes at GXO's logistics facility in Georgia. Agility also says Digit has been deployed at customer sites connected with GXO, Schaeffler, Toyota Motor Manufacturing Canada and Mercado Libre.
China is producing similar evidence. UBTECH said its full-sized embodied humanoid business shipped 1,079 units during 2025 and generated RMB820 million. More than 80% of those robots were deployed in manufacturing, logistics, electronics, semiconductor, aviation and industrial-data applications.
The industry has crossed a basic threshold: humanoids can perform repetitive commercial jobs for extended periods.
What we have not seen yet is hundreds of robots running inside the same customer operation for years. That is the scale that could turn successful pilots into a genuine labor market.
This chart, featured in our robotics market deck, shows annual venture capital investment in robotics startups
Is traditional industrial robotics booming too?
Industrial robotics is growing steadily rather than explosively, but the customer base is becoming much broader.
The International Federation of Robotics counted roughly 542,000 new industrial robot installations worldwide in 2024. The installed global fleet reached about 4.66 million robots, up 9% from the previous year. Annual installations are expected to move toward 700,000 by 2028.
North America's latest numbers are even more interesting when we look below the headline growth rate. Robot unit orders increased only 2% in the first half of 2026. Automotive manufacturers, historically one of the biggest buyers, actually cut orders by 25%.
Other industries more than compensated. Semiconductor and electronics orders rose 35%. Pharmaceuticals and life sciences increased 32%. Automotive components grew 24%, while food and consumer goods rose 17%.
More than half of second-quarter robot orders came from non-automotive customers.
That gives us a healthier robotics market than one driven by a single car-investment cycle. Warehouses, electronics plants, pharmaceutical companies, food producers and smaller manufacturers are all becoming relevant buyers.
Teradyne's results show the same trend from the vendor side. Its robotics business, which includes Universal Robots and Mobile Industrial Robots, generated around $100 million of second-quarter revenue, about 33% more than a year earlier.
Traditional automation probably will not produce the spectacular shipment curves that humanoid startups promise. It does give the robotics boom something much more useful: millions of machines already proving that companies will spend serious money when automation works.
Is China already in a robotics bull market?
China is already deep into a robotics boom, and its strongest advantage comes from manufacturing scale rather than humanoid hype.
China installed around 295,000 industrial robots in 2024, according to the International Federation of Robotics. That represented roughly 54% of all new industrial robot installations worldwide. Its operating industrial robot fleet has now passed 2 million machines.
The Financial Times recently reported that China's robotics industry generated more than RMB300 billion of revenue in 2025 after growing by more than 20% annually on average over five years. China also became a net exporter of industrial robots.
Humanoids are being built on top of that existing supply chain. Counterpoint's latest shipment ranking is dominated by Chinese companies, including AGIBOT, Unitree, Galbot, UBTECH and Leju. The top five humanoid vendors together controlled 86% of global first-half shipments, with Chinese groups occupying all five positions.
Government support is also getting more practical. Chinese authorities have launched programs designed to move humanoids and embodied AI into continuous real-world operation, while a proposed standards framework aims to establish at least 100 key humanoid standards by 2028.
This combination is difficult to reproduce quickly: huge industrial robot demand, local component suppliers, fast hardware iteration, government-backed deployment programs and increasingly large humanoid production runs.
The flashy humanoid videos get most of the international attention. China's larger advantage is that it already has the deepest robot manufacturing and deployment ecosystem in the world.
If you want more recent data on this point, please see our latest robotics market report.
This chart, featured in our robotics market deck, breaks down Figure’s playbook in robotics
Could the US still win the most valuable part of robotics?
The US could still capture the most valuable part of the robotics market if American companies control the AI models and computing layer that make robots useful.
The capital concentration is already striking. Figure is valued at $39 billion. Skild AI is above $14 billion. Physical Intelligence has moved into roughly $11 billion territory. Apptronik is worth about $5.3 billion after raising almost $1 billion.
Those valuations are concentrated around software-heavy robotics companies rather than component manufacturers.
Figure's latest spending shows where the bet is going. In early September 2026, the company announced a strategic agreement with Nscale for an initial $3.5 billion of Nvidia Vera Rubin computing infrastructure, with potential commitments above $6 billion. The infrastructure is expected to start coming online in 2027.
Figure also launched Index, its physical-data system, and says it can now collect the equivalent of about 35 minutes of training data every second. Skild is pursuing a general-purpose model that can transfer between different types of robots. Nvidia is building a common development stack through GR00T, Isaac and Cosmos.
China currently has the clearer advantage in physical production volume. American companies are making an enormous bet that robot intelligence will capture more value than robot assembly.
We have seen that pattern before in semiconductors and smartphones, where the company making the finished device or physical component does not necessarily earn the highest margins.
If robot foundation models eventually transfer well across manufacturers, the US could dominate a very valuable software layer even while China ships more machines.
Are robot economics finally good enough for mass adoption?
Robot economics are getting much more attractive, especially in warehouses and factories where labor is expensive and tasks repeat thousands of times.
US warehouse wages reached roughly $26.85 an hour by mid-2026, according to data recently cited by The Wall Street Journal. That was around 5% higher than a year earlier and 41% above the level a decade ago.
Automation costs are moving in the other direction. Robot hardware is getting cheaper, computer vision is improving, deployment software is easier to use and some suppliers now offer robots through leasing or subscription models.
Humanoids remain expensive, but the long-term cost curve is heading down too. Goldman Sachs estimated in 2024 that humanoid manufacturing costs had already fallen roughly 40% from earlier expectations as components and supply chains improved.
The comparison with human labor gets more interesting once utilization rises. A machine that can cover several shifts does not need to cost less than one worker's annual salary to make sense. Maintenance, charging, supervision and integration still have to be included, of course, and poor reliability can wreck the calculation.
That explains why warehouses and structured manufacturing jobs are becoming the first serious proving grounds. Moving totes, loading components, sorting products and transferring material create thousands of similar movements every day. Customers can measure exactly how much labor is saved.
The economics become much harder in homes, restaurants or other messy environments where every task looks slightly different.
Mass adoption does not require humanoids to become as capable as people at everything. Enough cheap, reliable performance on a few high-frequency jobs could already create a very large market.
This chart, featured in our robotics market deck, shows annual funding in robotics startups
What is still stopping humanoid robots from scaling?
Reliability, battery life and training data are still the biggest things holding humanoid robots back from mass deployment.
Robot demonstrations have improved so quickly that walking is no longer the interesting question. Companies can show machines running, jumping, carrying objects, folding laundry or performing increasingly long sequences of actions.
Factories care about a much harsher standard: whether the robot gets the same boring movement right several thousand times without stopping production.
Figure's earlier BMW deployment targeted placement success above 99%, an 84-second cycle time and zero human intervention. Requirements like those explain why moving from an impressive demonstration to a production line can take so long. Even a small failure rate becomes expensive when a task is repeated all day.
Battery life creates another practical headache. Industry estimates still put many humanoids in roughly the two-to-four-hour operating range. A factory that wants continuous coverage either needs additional robots, downtime for charging or battery-swapping systems.
Then we have data. Language models could train on an internet containing trillions of words. There is no equivalent internet-scale dataset showing robots how to manipulate every physical object in every situation.
Companies are now trying to manufacture that missing data. Figure's Index project collects large amounts of human and robot interaction data. Nvidia uses simulation and synthetic environments to generate training examples without requiring every movement to happen in the real world. Thousands of deployed robots should eventually create another source of data automatically.
That feedback loop is becoming more important: larger fleets create more physical experience, more experience improves the models, and better models make larger deployments possible.
How quickly that loop starts working in practice may decide whether the strongest part of the bull run arrives in 2027, 2028 or several years later.
Will cheaper specialized robots beat humanoids?
Specialized robots will probably take most robotics revenue for years, while humanoids expand the market into jobs that are awkward to automate with fixed machines.
Amazon has already deployed more than 1 million robots across its logistics network. Most are optimized machines designed for specific warehouse jobs rather than humanoids.
China's more than 2 million operating industrial robots tell the same story on an even larger scale. Fixed arms, autonomous mobile robots, cobots and specialized machines already automate huge amounts of economically valuable work.
They often have straightforward advantages. Wheels use less energy than legs. A fixed arm does not need to balance itself. A machine designed around one task can be cheaper and faster than a general-purpose robot.
Humanoids become interesting where the environment was built for people and would be expensive to redesign. Factories contain stairs, carts, shelves, doors, tools and workstations designed around the human body. Warehouses have similar constraints. Homes push that problem even further.
We expect the robotics market to grow in layers. Specialized automation keeps spreading through structured environments, AI makes those machines easier to deploy, and humanoids gradually open jobs where traditional automation has struggled.
A robotics bull run does not depend on humanoids taking over the market. Industrial arms, warehouse robots, surgical systems and autonomous mobile machines can drive large revenue pools while humanoid technology matures.
If you want more recent data on this point, please see our latest robotics market report.
This chart, featured in our robotics market deck, compares the main business model options for warehouse AMR robotics providers
Are robotics stocks already too expensive?
Some robotics companies are already priced for enormous future success, so a sector bull run can happen alongside brutal losses in individual names.
Figure's $39 billion valuation is the obvious example. The company has proven real factory work and is scaling manufacturing, yet its present commercial fleet is tiny compared with what that valuation eventually requires.
Skild AI is valued above $14 billion while the category of general-purpose robot foundation models is still being created. Apptronik reached roughly $5.3 billion before Apollo had reached anything close to mass deployment.
Public markets can become even more aggressive when a scarce robotics pure play appears. Unitree's Shanghai listing produced an extraordinary first-day repricing, showing just how much investor demand can build when direct humanoid exposure is limited.
At the same time, established robotics companies show what mature economics can look like. Intuitive Surgical generated $2.89 billion of revenue in the second quarter of 2026, up 19% year over year. Its installed da Vinci fleet reached 11,710 systems, and instruments and accessories alone generated $1.73 billion during the quarter.
One part of the robotics market already produces billions of dollars of recurring revenue. Another part is being valued on the expectation that equally large markets will emerge later.
We would therefore expect enormous dispersion during a robotics bull run. Robotics can become a huge industry while plenty of robotics stocks still turn out to be terrible investments.
| Robotics exposure | What investors get today | Main risk |
|---|---|---|
| Figure / Skild / Apptronik | Private physical-AI upside | Valuations far ahead of revenue |
| Humanoid pure plays | Direct humanoid exposure | Commercial scale remains unproven |
| Intuitive Surgical | Mature robot revenue and recurring usage | Much less humanoid-style upside |
| Teradyne | Cobots and mobile robots | Robotics is only part of the company |
| Nvidia | Robot computing and AI infrastructure | Robotics remains small inside a huge AI business |
| Industrial automation groups | Existing profitable robot demand | Slower growth than humanoid startups promise |
Could Tesla's Optimus start the next robotics stock boom?
Tesla's Optimus could become the biggest public-market catalyst for humanoid robotics if Tesla finally shows investors measurable production and useful work.
Tesla has said Optimus Gen 3 is being designed for mass production and has been installing production equipment with an eventual capacity ambition of 1 million robots per year.
That target is enormous compared with today's humanoid industry. Counterpoint expects the entire global market to ship just over 50,000 humanoids during 2026.
Tesla also gives investors something Figure, Apptronik and many other leading US humanoid companies cannot currently offer: a large, liquid public stock.
A credible Optimus ramp could therefore change the public-market narrative very quickly. We would want to see the kind of information Tesla already gives investors about cars: robots produced, robots deployed, hours worked, intervention rates, useful tasks completed, manufacturing cost and eventually revenue from external customers.
Production capacity alone tells us very little about demand.
If Tesla reaches thousands of productive Optimus deployments and publishes convincing operating numbers, investors will suddenly have a visible public benchmark for the whole humanoid industry. Figure, Unitree, Apptronik and the rest of the sector would be valued against something much more concrete.
That could easily produce another leg higher in robotics stocks.
If you want more recent data on this point, please see our latest robotics market report.
This chart, featured in our robotics market deck, breaks down revenue across customer segments in the robotics market
What would prove the real robotics bull run has arrived?
The real robotics bull run will be obvious when productive deployments start growing almost as quickly as robot funding and robot shipments.
Humanoid volume is the first number to watch. Global shipments are already heading above 50,000 units, but manufacturing and logistics currently account for only about 18% of them. We would become much more bullish when productive industrial work represents the majority of shipments.
Customer fleet size is even more useful. Five robots inside a factory mostly tell us the technology can be tested. Fifty begin to tell us something about integration. Hundreds at the same customer would show that the economics survived contact with real operations.
Repeat purchases matter just as much. BMW, GXO, Toyota, Mercedes-Benz and other large companies have already participated in robotics pilots or deployments. A second or third major order after a customer has measured the first fleet would be much stronger evidence than another new partnership announcement.
Traditional robotics should also keep broadening. North America's latest figures are encouraging because robot demand rose even while automotive OEM orders fell 25%. Semiconductor, pharmaceutical, automotive-component and food companies filled the gap.
Finally, public-company revenue needs to catch up. A genuine sector-wide bull market becomes much easier to sustain when several robotics companies can report 30%, 50% or higher revenue growth rather than asking investors to value prototypes ten years into the future.
| What we would watch | Where robotics is now | What would convince us |
|---|---|---|
| Humanoid annual shipments | >50,000 expected | >100,000 |
| Manufacturing + logistics share | ~18% | >50% |
| Large customer fleets | Mostly small deployments | Hundreds per site |
| Repeat commercial orders | Starting to appear | Common among major customers |
| Public pure plays | Still scarce | Several liquid robotics names |
| New robotics revenue | Uneven and often small | Fast growth across multiple vendors |
What could kill or delay the robotics bull run?
The biggest risk to the robotics bull run is a long gap between manufacturing capacity and customers willing to pay for productive robots.
Humanoid companies are investing heavily in factories while most shipped robots still go into entertainment, research, data collection and other early applications. If companies can manufacture 100,000 humanoids before industry has profitable jobs for 100,000 humanoids, prices and valuations could fall sharply.
Reliability could create the same delay. A robot that succeeds 95% of the time may look impressive in a research lab but become unusable when a factory needs thousands of uninterrupted cycles.
Battery constraints add another layer of cost. Two-to-four-hour operating periods force companies to think about charging downtime, battery swaps or spare robot capacity.
Valuations leave little room for a slow transition. Figure at $39 billion, Skild above $14 billion and several other physical-AI startups already carry expectations of very large future markets.
The dangerous scenario is quite mundane: robots keep getting better, but customers take five more years than investors expect to deploy them at scale.
That would still produce a huge robotics industry eventually. It could also produce a nasty robotics bear market along the way.
This chart, featured in our robotics market deck, shows how home cleaning robot technology has evolved over time
When will the robotics bull run really start?
The broad robotics bull run is beginning now, but 2027–2028 looks like the strongest window for it to become a genuine commercial and public-market boom.
Capital has already crossed the line. Robotics startups raised $18.8 billion by late June 2026, more than the sector raised during all of 2025 and more than during the 2021 venture peak.
Robot production is crossing another line. Global humanoid shipments grew almost 300% in the first half of 2026 and are expected to exceed 50,000 for the full year.
Commercial deployment is the lagging piece. As we saw above, manufacturing and logistics still represent only around 18% of humanoid shipments. Figure's work at BMW, Digit's 100,000-plus tote movements at GXO and UBTECH's industrial deployments show that useful humanoid work is real, but fleet sizes remain small compared with the market investors are pricing in.
The next two years contain several plausible accelerants at once. Figure is scaling Figure 03 production, building far larger AI-compute infrastructure and collecting physical training data. Tesla is pushing Optimus toward manufacturing. Chinese humanoid suppliers are moving into larger production runs. Traditional robot demand is spreading outside automotive. Warehouse labor keeps getting more expensive.
We therefore see three phases.
The speculative robotics bull run is already here.
The commercial robotics bull run has a good chance of becoming much broader in 2027–2028 if pilots convert into hundred- and thousand-unit customer fleets.
Mass adoption will take considerably longer. Humanoids still need better reliability, longer operating time, cheaper hardware and far more real-world training data before they become ordinary pieces of equipment across the economy.
For investors asking when robotics becomes one of the next major technology trades, waiting for the distant mass-adoption phase would probably be too late. The trade has started. The evidence needed to turn today's excitement into a durable robotics boom should arrive — or fail to arrive — over the next two years.
If you want more recent data on this point, please see our latest robotics market report.
OUR METHODOLOGY
We approached the question by breaking a loosely defined "robotics bull run" into the things that would actually show one taking shape, rather than relying on market excitement or a single headline metric.
We looked across capital formation and valuations, robot production and shipment growth, the mix of commercial versus experimental use, real-world deployments, traditional automation demand, unit economics, advances in robot intelligence and infrastructure, and the evidence coming from public-company revenue and customer behavior.
We prioritized the freshest available evidence and combined several relevant measures instead of allowing one company announcement or forecast to drive the conclusion. Measured deployments, customer orders, repeat usage, installed fleets and reported revenue carry more weight in our judgment than announced capacity, partnerships or long-term production targets.
Broad industry data was used alongside company-level evidence so that unusually strong individual examples did not become proxies for the entire market. That is why the article pairs humanoid shipment data and venture funding with industrial robot orders, customer-side deployments and public-company revenue.
We then looked for convergence. The key question was whether capital, production, customer adoption, economics and enabling technology were beginning to move in the same direction, and which pieces were still lagging.
The 2027–2028 timing judgment does not come from one shipment forecast or one company target. It is the period in which several of the commercialization tests identified above should become much easier to judge: larger customer fleets, repeat orders, a higher share of robots doing productive work, broader revenue growth and evidence that manufacturing ambitions are being matched by demand.
Key industry sources include Crunchbase News on global robotics venture funding, Counterpoint Research on humanoid shipments and application mix, the International Federation of Robotics on industrial robot installations, and the Association for Advancing Automation on North American robot orders.
For commercial deployments and company-level evidence, we used BMW Group on Figure's factory deployment, BMW Group on the Figure 03 logistics use case, Agility Robotics on Digit at GXO, and UBTECH on its humanoid shipments and industrial use.
For capital, software infrastructure and future production capacity, we used Figure's Series C disclosure, Skild AI's Series C announcement, Apptronik's Series A disclosure, Nvidia's GR00T platform documentation, Figure's Nscale compute partnership, and Tesla investor materials on Optimus production ambitions.
For mature robotics economics and cost comparisons, we also used Amazon on its fulfillment robot fleet, Goldman Sachs Research on humanoid cost trends, Teradyne's second-quarter robotics results, Intuitive Surgical's second-quarter results, and Chinese government reporting on humanoid standards and deployment policy.
In our robotics market deck, we identify pain points entrepreneurs should prioritize
Related blog posts
