What are the top startups in the digital twin market?
In our digital twin market deck, you will find everything you need to understand the market
SUMMARY
Neara is the top digital twin startup today, with NavVis and Akselos forming the strongest group behind it and Willow, PassiveLogic and JuliaHub rounding out the broader top tier.
The market is moving away from generic digital twin platforms. The strongest startups own a narrow, expensive problem where the twin directly changes an engineering, maintenance, operating or clinical decision.
Neara stands out because its models are already used across electricity networks at unusual scale. The company combines engineering-grade physics, millions of infrastructure assets and measurable effects on capacity, reliability and capital spending.
NavVis leads the spatial-twin branch. Its advantage is less about simulating asset behavior and more about building the machine-readable spatial layer that factories, robotics systems and physical AI increasingly need.
Akselos is smaller, but its moat may be deeper than that of several better-funded rivals. Its reduced-basis finite-element technology turns high-fidelity structural simulation into something fast enough to inform real operating and maintenance decisions.
Buildings are becoming one of the clearest tests of where digital twins go next. Willow has stronger commercial proof today, while PassiveLogic is betting that the model itself can become the control system and operate the building autonomously.
JuliaHub shows how quickly the category is blending into engineering AI. Its Dyad platform gives agents access to physics-based models, pushing digital twins upstream into machine design, validation and control-code generation.
Reality-capture companies still matter, but they rank differently. Buildots, Exodigo and Cintoo create valuable representations of the physical world, yet customers often buy construction intelligence, subsurface intelligence or reality-data workflows rather than the twin itself.
Human digital twins have become their own branch of the market. Twin Health has the stronger commercial platform today, while Unlearn has the more unusual statistical approach through patient-level counterfactual models for clinical trials.
Funding is concentrating around vertical businesses, not another generation of universal twin platforms. Recent rounds for Exodigo, NavVis, PassiveLogic, JuliaHub, Neara and Twin Health all back a specific operational problem tied to the physical or biological world.
The larger strategic shift is toward physical AI. The most valuable digital twin startups increasingly look like world-model companies: they do not just reproduce reality visually, they help software reason about what the real system will do next and, in some cases, act on it.
What actually counts as a digital twin startup today?
A real digital twin startup today needs a model that stays meaningfully connected to a physical or biological system and helps someone understand what that system is doing, what it could do next, or what action to take.
That definition immediately removes a lot of noise. A 3D scan can feed a digital twin, but the scan alone is usually just reality capture. A simulation can become part of a twin, but a model that never reconnects with the real asset is closer to engineering software. And the label has become even looser lately: Twin1 AI, for example, recently raised $20 million to build AI “digital twins” of professionals. That may become a real software category, but it has very little to do with Neara modelling an electricity network or Akselos calculating fatigue on offshore infrastructure.
We use a practical test. Neara qualifies because its model represents real grid assets, applies engineering physics and changes what utilities do with those assets. Willow qualifies because live building data updates a structured model that facilities teams use to diagnose problems. Akselos continuously calculates the structural condition of industrial equipment. Twin Health models a person's changing metabolic state from real health data. Unlearn is slightly different because its patient twins forecast counterfactual clinical outcomes, but the model still represents an identifiable real participant.
NavVis, Cintoo, Buildots and Exodigo sit nearer the edge. Their technologies create extremely useful representations of physical reality, yet much of the value comes from capturing, mapping or comparing reality rather than continuously simulating how an asset behaves.
We include those companies in the broader market, while giving more weight to startups where the twin itself drives the decision.
Why are digital twin startups suddenly turning into physical AI companies?
Digital twin startups are increasingly becoming physical AI companies because the valuable product is shifting from seeing the physical world to letting software reason about it and eventually act on it.
NavVis gives us a very recent example. In its latest $85 million round, the company barely positioned itself as another 3D mapping vendor. It described its spatial twins as the data foundation for physical AI. More than one billion square metres were scanned, processed and distributed through NavVis during 2025 alone, and customers are already using that spatial data for robotics deployment and industrial AI.
PassiveLogic has made an even bigger jump in language. Its current pitch is “physical AI for buildings.” The company's physics-based world model feeds autonomous building controls, so the software can simulate possible actions and adjust equipment instead of waiting for a facilities team to interpret another dashboard.
JuliaHub has moved the same way from scientific computing toward agentic engineering. Its Dyad platform combines AI agents with physics-based system models so engineers can create, test and validate machines virtually. Akselos, meanwhile, now talks about AI for structural integrity and continuously calculating how equipment is performing from real operating data.
The pattern is pretty clear. Digital twins are becoming the world model underneath physical AI.
That also makes the market more interesting commercially. Companies can charge much more for software that changes a maintenance decision, releases grid capacity or controls a building than for software that mainly produces a better visual representation.
If you want more recent data on this point, please see our latest digital twin market report.
This market map, featured in our digital twin market deck, highlights top companies and startups in the digital twin market
Can digital twin startups still win against Siemens, Bentley and NVIDIA?
Digital twin startups can still win against Siemens, Bentley and NVIDIA, but the best route today is owning a narrow, expensive problem that the giant horizontal platforms do not solve deeply enough.
Trying to build another universal digital twin platform looks much harder. Bentley already offers iTwin for infrastructure. Siemens connects engineering software, simulation, industrial data and automation. NVIDIA provides Omniverse, accelerated simulation and much of the computing stack underneath physical AI. Autodesk, Schneider Electric and Dassault Systèmes add another layer of incumbent pressure.
The startups doing well have mostly chosen another path. Neara understands the physics of electricity networks. Akselos models structural fatigue on assets where a wrong maintenance decision can cost millions. Willow has built building-specific data models and operational workflows. Unlearn focuses on statistical counterfactuals inside clinical trials. None of those companies needs to replace every piece of Siemens or Bentley software to become valuable.
Cognite's current situation makes this even clearer. Schneider Electric agreed in June 2026 to acquire Cognite for $3.1 billion. Cognite had built a contextualized industrial data layer that helped companies connect equipment, engineering information, time-series data and AI. Schneider plans to combine that capability with its existing industrial software rather than rebuild the whole thing internally.
The startup opportunity remains large, but it sits increasingly in specialized intelligence rather than generic twin infrastructure.
Is Neara the best digital twin startup right now?
Neara is the strongest pure digital twin startup we can identify right now because it combines engineering-grade twins, network-scale deployment, strong customer evidence and a problem that is becoming more urgent.
Neara's latest Series D raised A$90 million and pushed the company above an A$1 billion valuation. Funding alone would not put it first. The deployment numbers do: Neara says it has modelled more than 15 million infrastructure assets across more than 3 million kilometres and works with close to 90% of Australia's electricity network utilities, alongside customers such as Southern California Edison, CenterPoint Energy, ESB Networks and ScottishPower.
The software is also being used for unusually consequential decisions. Neara has shown utilities that some existing lines can carry substantially more electricity than their old engineering assumptions suggested. Essential Energy found areas where available capacity could be roughly twice what it previously believed. Another Australian utility used the platform on work that could help unlock almost 3,000 MW of renewable generation.
The operational examples are just as strong. SA Power Networks used Neara's models during flooding to keep power on longer and re-energize lines in five days when the original expectation was around three weeks. One U.S. utility found that 84% of poles scheduled for replacement were structurally safe to remain in service, cutting a planned 25,000 replacements to roughly 4,000.
It is hard to find startup digital twin deployments where the software changes capital spending, reliability, engineering and emergency response at this scale.
Neara currently has the clearest claim to number one.
If you want more recent data on this point, please see our latest digital twin market report.
As this chart shows, and as featured in our digital twin market deck, search interest in digital twins has increased sharply
Has NavVis become the biggest spatial digital twin startup?
NavVis is the scale leader among independent spatial twin startups today, and its latest numbers put meaningful distance between it and most smaller reality-capture competitors.
The company says more than 1,500 customers in over 50 countries use NavVis, including BMW, Volkswagen, Toyota, Mercedes-Benz, ExxonMobil, BASF, Bosch, KION and Siemens. More than 150,000 people use its systems, while over 2.5 billion square metres have been captured through the platform since its creation.
The pace is accelerating. More than one billion square metres were scanned, processed and distributed during 2025 alone. That means roughly 40% of the company's cumulative spatial footprint was added in a single year.
Its recent $85 million Series D is another important piece of evidence. NavVis says the money will expand its spatial data engine and physical AI products. Customers already use the platform to document factories, plan shutdowns, monitor construction, prepare robotics deployments and train industrial models.
We rank NavVis behind Neara because NavVis often provides the spatial foundation rather than the deeper physics or operational decision itself. Still, that foundation is becoming more valuable as robots and physical AI systems need accurate machine-readable representations of factories, data centers and other real environments.
For spatial twins specifically, NavVis is currently the company to beat.
Who's ahead in building digital twins: Willow or PassiveLogic?
Willow is ahead in commercial building digital twins today, while PassiveLogic is taking the bigger technical swing by trying to make the twin control the building itself.
Willow already has serious operating evidence. Its customers include Walmart, Dallas Fort Worth International Airport, Brookfield, universities, hospitals and large property groups. At Walmart, Willow says its platform identified 842 issues in six months, reduced critical downtime costs by 20% and produced $1.4 million in savings. Georgia Southern University reported roughly $1 million of operating savings within nine months.
DFW gives us the clearest picture of scale. Willow's airport deployment brings together more than 170,000 assets, over 120,000 live data points and several previously separate building and maintenance systems. DFW's CFO has said the airport expects digital twin technology to reduce maintenance costs per square foot by 20% to 25% over five years.
PassiveLogic is building something more radical. Its Quantum world model represents building geometry, equipment, relationships and physics, while its controllers use that model to decide how the building should operate. PassiveLogic says its tools can turn weeks of controls design and commissioning into much shorter workflows and can continuously optimize HVAC and other systems.
Investors have put serious money behind that vision. The company's latest Series C was $74 million, with strategic backing from companies connected to real estate, controls and industrial technology.
For now, Willow ranks higher because its customer evidence is much easier to verify at enterprise scale. PassiveLogic could eventually become the more important company if autonomous building control works reliably across large, messy portfolios.
If you want more recent data on this point, please see our latest digital twin market report.
This chart, included in our digital twin market deck, shows annual VC investment in digital twin startups
Does Akselos have the strongest industrial digital twin technology?
Akselos probably has the deepest pure engineering technology among the industrial digital twin startups we reviewed.
Its core advantage comes from reduced-basis finite-element analysis developed through years of research associated with MIT. Traditional high-fidelity structural simulation can be far too slow for continuous asset monitoring. Akselos built its business around making that type of calculation fast enough to keep an engineering model useful during real operations.
Shell's Bonga FPSO shows what this looks like in practice. Akselos modelled around 15,000 structural fatigue locations and narrowed them to roughly 230 critical hotspots that deserved attention. According to Akselos' Shell case study, the resulting inspection and maintenance changes reduced operating costs by around 33%.
The technology has also been used to support life-extension decisions on aging offshore assets. More recently, Akselos has expanded its relationship with GRO, which made a follow-on investment in 2025 as the company pushed its Structural Performance Management software into wider industrial use.
There are larger startups in this article, and companies with far bigger funding rounds. Very few have a product where a competitor would need to reproduce both specialized physics technology and years of validation on safety-critical infrastructure.
That gives Akselos one of the strongest moats in the group and puts it near the top of our ranking despite its smaller funding footprint.
Is JuliaHub now a serious digital twin contender?
JuliaHub has become one of the most interesting digital twin startups in industrial engineering, especially after turning its physics stack into an agentic AI product rather than leaving it as specialist scientific software.
The company's latest Series B brought in $65 million from investors including Dorilton, General Catalyst and former Snowflake CEO Bob Muglia. JuliaHub launched Dyad 3.0 alongside that financing and says the product is already running in production with Fortune 100 customers.
Dyad gives AI agents access to physics-based models. Those agents can interpret specifications, build candidate models, run simulations, test constraints and generate control code. JuliaHub is essentially trying to bring the workflow advantages of AI coding agents into physical engineering, where an answer also has to obey thermodynamics, mechanics, electrical behavior and safety constraints.
The underlying ecosystem gives the startup an unusual distribution advantage. The Julia language is used across more than 10,000 companies and 1,500 universities, according to JuliaHub. The company also has partnerships around digital twins and scientific machine learning with companies such as Synopsys.
We would still rank JuliaHub below Neara, NavVis and Akselos because much of its current work happens before an asset exists physically. Its strongest use cases sit between digital twins, engineering simulation and AI-assisted product development.
But JuliaHub is moving quickly enough that leaving it out of a current top-startup list would already feel outdated.
If you want more recent data on this point, please see our latest digital twin market report.
This chart, included in our digital twin market deck, breaks down Neara's playbook in digital twins
Do Buildots and Exodigo really belong in the digital twin market?
Buildots and Exodigo belong in the wider digital twin ecosystem, although both have become strong businesses by solving specific reality problems rather than selling a conventional digital twin platform.
Buildots uses 360-degree site capture and computer vision to compare what has actually been built with project plans and schedules. The company recently raised $45 million, taking total funding to $166 million, and says it has moved from individual project deployments toward enterprise agreements, including several seven-figure contracts.
Its traction has also broadened. Buildots works with major contractors and owners including Intel, JE Dunn, Wates and Kier, tracks hundreds of millions of square feet of data-center construction, and recently acquired workforce and safety software company Genda. That acquisition pushes Buildots further toward productivity intelligence, connecting physical progress with labor activity.
Exodigo has built an equally distinctive business underneath the ground. Its hardware, sensors and AI reconstruct subsurface infrastructure without relying on large amounts of excavation. A $96 million Series B brought total funding to $214 million. Exodigo said more than 50 transit agencies, utilities and public organizations were already customers, including Amtrak, National Grid, LA Metro and California High-Speed Rail.
During one twelve-month period, Exodigo worked on projects connected to more than $75 billion of infrastructure investment across 18 U.S. states.
We keep both in the ranking because accurate, continuously refreshed knowledge of the physical world is becoming a core input to digital twins. Still, someone buying Neara or Akselos is buying a different kind of product. Buildots is primarily construction intelligence, while Exodigo is primarily underground intelligence.
Do Twin Health and Unlearn belong in the same digital twin market as industrial software?
Twin Health and Unlearn are genuine digital twin companies, but human digital twins have become distinct enough that we should compare them with each other before comparing them with power grids or factories.
Twin Health is the larger commercial story. Its latest financing raised $53 million at a reported valuation of roughly $950 million. The product combines wearable and health data with an AI model of each person's metabolism, then uses that model alongside clinical care to guide treatment for diabetes, obesity and related conditions.
The most useful recent evidence is economic rather than promotional. In a 2026 real-world employer analysis presented around ISPOR, Twin Health reported average annual savings of more than $9,000 per member with type 2 diabetes. Roughly $4,690 came from lower pharmacy spending, with another $4,357 linked to lower medical spending from fewer hospital and emergency visits.
Unlearn solves a completely different medical problem. Its models predict the control outcome of an individual clinical-trial participant, allowing researchers to extract more statistical information from a trial without simply enrolling more placebo patients.
AbbVie tested the approach retrospectively on its Phase 2 AWARE Alzheimer's trial. Unlearn says the resulting prognostic information could have reduced sample requirements while preserving statistical power. Its methods have also been evaluated with companies including Roche and Johnson & Johnson, and Unlearn has raised more than $130 million in total.
Twin Health ranks higher because it already looks like a large commercial health platform. Unlearn may ultimately have the more unusual technology.
Neither tells us much about who is winning industrial digital twins, so we keep human twins as a separate branch when forming the overall ranking.
This chart, included in our digital twin market deck, shows annual funding in digital twin startups
Where is digital twin funding actually going now?
Digital twin funding is concentrating around a small number of vertical companies, and the biggest rounds increasingly combine twins with physical AI, infrastructure or industry-specific workflows.
In our five-quarter deal set from Q2 2025 through Q2 2026, we counted 21 disclosed digital twin startup rounds worth about $534.9 million. The ten largest rounds represented 87.1% of that capital. Since that period ended, NavVis alone has added another $85 million round.
The companies receiving the biggest checks also tell us where investors think value is forming. Exodigo maps underground infrastructure. NavVis builds spatial twins. PassiveLogic runs buildings. JuliaHub simulates physical systems. Neara models power networks. Twin Health models metabolic health. Buildots observes construction.
There is no obvious cluster of venture capital flowing into generic “digital twin platforms.” Investors are backing vertical companies where the twin is tied to an expensive real-world problem.
The exits point in the same direction. CoStar completed its roughly $1.9 billion acquisition of Matterport in 2025. Schneider Electric has now agreed to buy Cognite for $3.1 billion. Beijing-based 51WORLD also became a listed company in late 2025.
Those three companies once appeared frequently on startup lists. Today they are better evidence of what mature digital twin assets can become than candidates for a ranking of independent startups.
| Company | Latest major round | Amount | What investors are backing |
|---|---|---|---|
| Exodigo | Series B | $96M | Underground infrastructure intelligence |
| NavVis | Series D | $85M | Spatial twins and physical AI data |
| PassiveLogic | Series C | $74M | Autonomous building control |
| JuliaHub | Series B | $65M | Agentic physics and industrial twins |
| Neara | Series D | ~US$64M | Physics-based utility infrastructure |
| Twin Health | Series E | $53M | Human metabolic twins |
| Buildots | Series D | $45M | Construction intelligence |
| Cintoo | Series B | €37M | Industrial reality-data management |
| VEERUM | Series B | C$12M | Industrial visual operations |
So which digital twin startups are actually leading today?
Neara is the best digital twin startup overall today, with NavVis and Akselos forming the strongest group behind it; Willow, PassiveLogic and JuliaHub round out the top tier once we broaden the market beyond infrastructure twins.
Neara takes first place because the twin itself drives engineering decisions across entire electricity networks and because the company now has both unicorn-level validation and unusually wide deployment. NavVis has the biggest spatial footprint we found among private companies and has just strengthened that position with another major financing. Akselos ranks third despite its smaller size because its physics technology is unusually difficult to reproduce and already influences safety-critical industrial decisions.
Willow currently beats PassiveLogic on commercial evidence in buildings, although PassiveLogic has a more ambitious autonomy architecture. JuliaHub has moved up quickly after its latest funding and Dyad launch. Twin Health dominates the human-twin branch commercially.
Buildots and Exodigo rank lower because their businesses are increasingly construction intelligence and subsurface intelligence respectively. Both may become larger companies than some higher-ranked names, but the digital twin is less central to what customers are actually buying.
Cintoo and VEERUM complete the group because each has built real industrial usage around spatial and visual representations of physical assets. Unlearn takes a higher strategic position than either because its patient-level counterfactual models are much harder to replicate, even though clinical digital twins remain their own market.
The ranking will probably keep changing as physical AI pulls digital twins closer to robotics, autonomous operations and AI agents. As of now, the strongest startups already look less like 3D-software companies and more like systems that understand how a specific part of the physical world behaves.
| Rank | Startup | Where it is strongest | Why we rank it here |
|---|---|---|---|
| 1 | Neara | Power grids and critical infrastructure | Best combination of digital twin depth, deployment scale and measurable operational value |
| 2 | NavVis | Spatial and industrial twins | Largest spatial footprint in the private market and strong new physical AI positioning |
| 3 | Akselos | Structural industrial twins | Deep physics moat and validated use on safety-critical assets |
| 4 | Willow | Building operations | Strongest verified commercial evidence in operational building twins |
| 5 | PassiveLogic | Autonomous buildings | Technically ambitious physics-based system with major upside if deployment scales |
| 6 | JuliaHub | Engineering and industrial systems | Fast-rising agentic physics platform with Fortune 100 production use |
| 7 | Twin Health | Human metabolic twins | Near-unicorn commercial health platform with increasingly strong economic evidence |
| 8 | Buildots | Construction intelligence | Large construction adoption, although the business has moved beyond a pure twin product |
| 9 | Exodigo | Underground infrastructure | Exceptional funding and customer traction in physical-world mapping |
| 10 | Unlearn | Clinical-trial twins | Highly differentiated patient-level counterfactual modelling |
| 11 | Cintoo | Reality data | Broad industrial usage and a strong 3D reality-data workflow |
| 12 | VEERUM | Industrial visual operations | Smaller company, but credible ROI in asset-intensive industrial environments |
If you want more recent data on this point, please see our latest digital twin market report.
This chart, included in our digital twin market deck, compares the main business model options for digital twin enterprise software platforms
OUR METHODOLOGY
We ranked digital twin startups by separating the market into the dimensions that best distinguish genuine leadership from category noise: product depth, real-world deployment, measurable customer outcomes, technical differentiation, commercial traction and recent market validation.
We gave more weight to companies where the twin itself drives a decision. Neara's grid models, Akselos's structural calculations, Willow's operational building models and Twin Health's metabolic models therefore receive more credit than products whose main value comes from reality capture, mapping or comparison.
Because the market spans infrastructure, buildings, industrial engineering, spatial data, construction and healthcare, we first judged companies against the signals that matter inside their own vertical. We then compared them through the common framework above rather than pretending that raw metrics from a power-grid platform and a clinical-trial company are directly interchangeable.
Recent evidence carried extra weight. We treated Neara's Series D, NavVis's $85 million financing and physical-AI positioning, PassiveLogic's autonomous-building architecture and JuliaHub's Dyad 3.0 launch as current indicators of competitive position rather than relying mainly on older reputation.
Deployment and customer outcomes mattered more than funding alone. Neara's utility use cases, Willow's results at Walmart and Dallas Fort Worth International Airport, and Akselos's Shell Bonga deployment were especially useful because they show the software changing engineering, maintenance or operating decisions with measurable consequences.
We included companies such as Buildots, Exodigo, Cintoo and VEERUM in the broader market because they create or maintain important representations of the physical world. They receive less ranking weight when customers are primarily buying construction intelligence, subsurface intelligence or reality-data workflows rather than a model that predicts or controls asset behavior.
Funding and M&A were used as market-validation checks, not as substitutes for product analysis. Recent rounds helped show where venture capital is concentrating, while CoStar's acquisition of Matterport and Schneider Electric's announced acquisition of Cognite show what mature digital twin assets can become strategically.
Key sources include Neara on its Series D and infrastructure expansion, NavVis on its $85 million Series D and physical AI strategy, Willow on the Walmart deployment, Willow on the DFW deployment, PassiveLogic on physics-based building control, Akselos on Shell's Bonga FPSO, JuliaHub on its Series B and Dyad 3.0, Buildots on its Series D, Exodigo on its Series B, Twin Health on its Series E, Unlearn on patient digital twins, Cintoo on its Series B, VEERUM on its Series B, CoStar on the completed Matterport acquisition, and Schneider Electric on the announced Cognite transaction.
This chart, featured in our digital twin market deck, illustrates how revenue is divided among customer segments in the digital twin market
