What is the real market size of digital twins?
Everything we know about the digital twin market, in one place. Get our report whether you build or invest.
Summary
The real digital twin market is about $40 billion globally today, with a defensible broad-market range of roughly $35 billion to $50 billion and a core software opportunity closer to $20 billion.
The biggest source of confusion is scope, not forecasting skill. Published estimates are often counting different mixes of simulation, PLM, industrial IoT, cloud, hardware and implementation under the same digital-twin label.
The 2025 reports cluster around roughly $30 billion, even though individual estimates stretch from below $20 billion to more than $36 billion. That center of gravity is more useful than any single decimal-point figure.
Pure software is materially smaller than the broad market. A reasonable working estimate is around $20 billion once systems integration, consulting, sensors, edge hardware and other enabling infrastructure are stripped out.
Digital twins are also not a cleanly new software category. A large part of the spending comes from older simulation, CAD, PLM and industrial software that has become more connected to live operating data.
Manufacturing is the clearest commercial center of gravity, with energy, infrastructure and aerospace close behind. Those sectors have expensive assets, measurable downtime costs and enough operating data to make digital-twin economics tangible.
Adoption has moved well beyond demos in industrial settings. Published projects from Electrolux, Ferrero, Siemens, Pringles and others show measurable gains in commissioning time, capacity, development speed and operating efficiency.
The vendor revenue test also supports a market worth tens of billions, but not a $40 billion pure-software figure. Siemens, Dassault Systèmes and Bentley already generate billions from relevant engineering software, yet only part of those portfolios is genuinely attributable to digital twins.
Growth forecasts near 50% a year look aggressive once the market is already this large. Something closer to the high-20s or low-30s is easier to reconcile with both current market estimates and the growth rates of the major software vendors.
AI is expanding the use case, especially through factory simulation, robotics and physical AI. The important distinction is that AI infrastructure can drive digital-twin demand without turning every GPU, cloud workload or robot into digital-twin revenue.
A $100 billion-plus market around the end of the decade is credible under a broad definition. Reaching roughly $150 billion requires both very strong adoption and continued expansion of what researchers count inside the category.
The practical takeaway is to keep three layers separate: roughly $20 billion for core software, about $40 billion for the broader digital-twin market today, and a much larger surrounding technology ecosystem that digital twins influence without fully owning.
Written by the New Market Pitch teamWe track new markets so founders and investors can move faster. About us
In our report covering the digital twin market, we always try to gather several sources and make estimates that make sense, so you can work with solid numbers.
Why can't anyone agree on the size of the digital twin market?
The digital twin market is clearly worth tens of billions of dollars today, but published estimates still disagree by so much that quoting one precise number gives a false sense of accuracy.
Look at 2025 alone. MarketsandMarkets puts the global market at $21.14 billion. Fortune Business Insights' latest estimate is $24.48 billion. IMARC is around $29 billion. Mordor Intelligence reaches $36.19 billion. Grand View Research is almost identical at $35.8 billion.
That leaves credible estimates for the same year roughly $15 billion apart.
The disagreement gets even bigger when we look at current estimates. Fortune Business Insights has the market at $33.97 billion, while Mordor Intelligence and Grand View Research are both around $49 billion. We are talking about a gap of roughly $15 billion before trying to forecast anything.
Different growth assumptions explain part of it, but the bigger problem is what researchers decide to count. Some focus on digital-twin software and directly related services. Others capture more of the simulation software, IoT infrastructure, cloud platforms, engineering tools, connected hardware and implementation work around a twin.
There is a real market underneath these numbers. We just have to define its boundaries before pretending we know its exact size.
What actually counts as a digital twin?
A digital twin is a continuously connected digital representation of a real asset, process or system, which already rules out a lot of things that now get marketed as digital twins.
A CAD model of a turbine is useful, but the model alone does not tell us what is happening inside the real turbine. A simulation can tell engineers how that turbine might behave under different conditions, but it can still exist without any live relationship with the physical machine.
The digital-twin idea becomes much clearer once operating data from the real turbine updates the digital representation and engineers use that connection to monitor performance, predict failures or test changes.
The Digital Twin Consortium puts synchronization at the center of its definition. NIST also emphasizes the connection between the physical system and its digital representation. ISO's manufacturing work follows a similar logic.
The commercial boundary is much messier because major vendors rarely sell a box labelled "digital twin."
Siemens can bring together Teamcenter, Simcenter, Tecnomatix, industrial automation and operational data. Dassault Systèmes builds virtual-twin workflows across CATIA, SIMULIA, ENOVIA and the 3DEXPERIENCE platform. Bentley connects engineering models and live infrastructure information through iTwin.
Digital twins have therefore become a capability spread across several products. That makes the technology easier to sell into existing industrial software stacks, while making the revenue much harder to isolate.
In our research about digital twins, we don't want to repeat what boring consulting reports say. Instead, we show you the real and latest market signals that matter.
How big do the main digital twin market reports say the market is?
The most credible reading today is that digital twins were roughly a $25 billion to $35 billion market in 2025, with some broader definitions already pushing above that range.
The useful insight comes from comparing estimates rather than choosing the largest or newest one.
MarketsandMarkets remains near the bottom at $21.14 billion. Fortune Business Insights' latest update gives $24.48 billion. IMARC is around $29 billion. The Business Research Company has also been around the high-$20-billion range. Grand View Research and Mordor sit close to $36 billion.
Those estimates produce a center of gravity around $30 billion for 2025.
The current-year estimates expose the definition problem even more clearly. Fortune Business Insights reaches $33.97 billion, while Mordor Intelligence is at $49.2 billion and Grand View Research at $49.5 billion.
A $34 billion market and a $49 billion market cannot both be precise measurements of exactly the same economic activity. They can both be reasonable estimates once the boundaries change.
| Research firm | 2025 estimate | Current / next-year estimate | Longer-term forecast | Forecast growth |
|---|---|---|---|---|
| Global Market Insights | $18.9B | $26.4B | $428.1B by 2034 | 41.7% CAGR |
| MarketsandMarkets | $21.14B | - | $149.81B by 2030 | 47.9% CAGR |
| Fortune Business Insights | $24.48B | $33.97B | $384.79B by 2034 | 35.4% CAGR |
| The Business Research Company | $28.9B | $39.75B | $122.24B by 2030 | 32.4% CAGR |
| IMARC | ~$29.3B | - | $223.6B by 2034 | 25.3% CAGR |
| Grand View Research | $35.8B | $49.5B | $328.5B by 2033 | 31.1% CAGR |
| Mordor Intelligence | $36.19B | $49.2B | $228.46B by 2031 | 35.95% CAGR |
So what is the real size of the digital twin market today?
We would currently treat digital twins as roughly a $40 billion market, while keeping a reasonable broad-market range of about $35 billion to $50 billion.
That estimate starts with the roughly $30 billion center of gravity we find across 2025 reports and then accounts for another year of very fast growth.
It also fits the three useful current estimates reasonably well. Fortune Business Insights is near $34 billion, while Mordor and Grand View Research sit just below $50 billion. Calling the market exactly $49.2 billion would give us precision the underlying evidence cannot support. Calling it approximately $40 billion says almost everything we can defend.
The lower end probably excludes part of the implementation, integration and enabling software that customers genuinely buy to make digital twins work. The upper end captures more of those surrounding technologies.
For most strategic work, we would use approximately $30 billion as the 2025 baseline and around $40 billion as the best current working figure.
That estimate could easily move several billion dollars depending on what we decide to do with simulation, PLM, industrial IoT and services. The order of magnitude is much more robust than the decimal point.
If you want more recent data on this point, please see our latest digital twin market report.
In our market report about digital twins, we like to show you the big picture. We know your time is precious, so we make every page quick to read and packed with as much learning as possible per minute.
How much of the digital twin market is actually software?
Core digital twin software is probably a $15 billion to $25 billion market today, and roughly $20 billion is a sensible working estimate.
Global Market Insights estimated software at 65% of its 2025 market. Applied to its $18.9 billion total, that works out to approximately $12.3 billion.
Mordor gives us the other end of the range. Its "solutions" category accounted for 62.85% of a $36.19 billion market in 2025, or about $22.7 billion. That category also includes connected hardware, so $22.7 billion should be treated as an upper bound for pure software rather than a clean software figure.
If we apply a roughly 60% share to the broader $30 billion 2025 market we derived above, we land close to $18 billion.
The vendor numbers make that scale believable. Dassault Systèmes generated €5.64 billion of software revenue across its entire portfolio in 2025. Siemens' Digital Industries software business generated €6.17 billion. Bentley produced $1.50 billion of total revenue. Those businesses include plenty of engineering work outside digital twins, so only part of each total belongs here.
A core market around $20 billion leaves room for Siemens, Dassault Systèmes, Ansys, PTC, Bentley, AVEVA, Autodesk, the cloud platforms and smaller specialist suppliers without pretending that every CAD, PLM or simulation subscription has suddenly become digital-twin revenue.
| Spending category | Count in broad market? | Count in core software market? | Why |
|---|---|---|---|
| Dedicated digital-twin platforms | Yes | Yes | Direct digital-twin revenue |
| Twin-specific simulation | Yes | Yes | Central part of many twins |
| General CAD software | Sometimes | Only attributable use | Most CAD work has other purposes |
| PLM software | Often | Only attributable use | Frequently shared with non-twin workflows |
| Industrial IoT platforms | Often | Twin-related portion | Enables synchronization |
| Sensors and edge hardware | Sometimes | Usually no | Infrastructure around the software |
| Cloud compute and storage | Sometimes | Direct workloads only | Shared with many other applications |
| Systems integration | Yes | No | Real project cost but service revenue |
| Consulting and model creation | Yes | No | Often essential to deployment |
Is digital twin technology partly old software with a new name?
Yes. Quite a lot of digital twin spending comes from established simulation, engineering, PLM and industrial software that companies now connect in more useful ways.
Simulation existed decades before "digital twin" became a popular software category. So did CAD, product-lifecycle management, industrial control systems and equipment monitoring.
What changed is how tightly those pieces can now work together.
An engineering model can remain connected to the asset after production. Operational data can flow back into the model. Engineers can compare real behavior with simulated behavior. Maintenance teams can test a change before touching the machine. Factory planners can simulate a new line using the same underlying product and production data.
That is why Siemens talks about a comprehensive digital twin spanning product design, electronics, software, manufacturing and operations. Dassault Systèmes increasingly uses "virtual twins" across entire industrial lifecycles. Bentley brings engineering and operational infrastructure data into the same environment.
The category is commercially meaningful because those connections create new workflows and new spending. At the same time, some of the revenue now presented around digital twins would previously have appeared under simulation, CAD, PLM, IoT or industrial software.
That overlap is one reason the market can grow quickly without every dollar representing genuinely new IT expenditure.
If you want more recent data on this point, please see our latest digital twin market report.
In our document detailing everything you need to know about digital twin, we turn what is vague and complex into simple, useful visuals. Market maps like this one show you at a glance where every startup stands.
Which industries are actually spending money on digital twins?
Manufacturing is still the biggest real digital twin market today, while energy, infrastructure and aerospace provide the other clearest large-scale uses.
Mordor estimates manufacturing at 35.1% of global digital-twin spending in 2025. That makes sense when we look at where the technology already has years of industrial deployment behind it.
Factories can use digital twins before production starts, during commissioning and once the line is running. Product teams simulate equipment. Manufacturing teams test layouts and robot movements. Operators compare actual production with expected production. Maintenance teams look for anomalies before equipment fails.
Energy has similarly good economics because turbines, grids, refineries, pipelines and other assets are expensive enough for small performance improvements to matter.
Infrastructure is getting stronger as well. Bentley, Autodesk, Siemens and geospatial platforms increasingly connect engineering models with live information from buildings, transport networks, utilities and other physical systems.
Aerospace remains a natural fit because engines and aircraft combine high asset values, enormous amounts of sensor data and costly downtime.
Healthcare and city-scale twins receive a lot of attention these days, although their current commercial bases are still less mature. Both also create especially large measurement problems. A city twin can include BIM, GIS, IoT, mapping, traffic software and environmental data. A healthcare twin can touch medical imaging, AI, monitoring devices and clinical data platforms.
Counting all of those adjacent technologies quickly produces a huge market. It also moves us much further away from measuring digital-twin revenue itself.
Are companies actually using digital twins at scale now?
Digital twins have already moved well beyond pilot projects in manufacturing and infrastructure, although full enterprise-wide deployments remain much rarer than individual successful applications.
Capgemini surveyed more than 1,000 organizations working with digital twins and found an average 15% improvement across important sales and operational metrics, with system performance improving by more than 25%.
The more convincing evidence comes from individual projects where companies publish an operational result.
Electrolux reported development-time reductions of 20% to 30% and cost reductions of 15% to 20% from its factory digital-twin work. Ferrero said virtual commissioning helped cut commissioning time by 30%. Wipro PARI reported a 70% reduction on an automotive assembly-line project.
The use cases are also getting more practical. Companies are spending less time talking about generic "virtual replicas" and more time using twins for commissioning, factory layout, predictive maintenance, production optimization and robot simulation.
Foxconn, for example, is now working with NVIDIA on a Texas AI-server factory where an Omniverse digital twin is being used to simulate production lines and automation before deployment. That is a useful recent signal because it connects digital twins directly with the enormous buildout of AI infrastructure and physical AI.
Adoption is uneven, but the industrial use cases are already very real: mature applications have measurable economics, while many broader enterprise ambitions are still catching up.
| Organization | Digital-twin use | Reported result |
|---|---|---|
| Siemens Electronics Factory Erlangen | Production and operations | Labor productivity +69%, energy use -42%, time to market -40% over the measured period |
| Electrolux | Factory design and engineering | Development time -20% to -30%, costs -15% to -20% |
| Ferrero | Warehouse virtual commissioning | Commissioning time -30%, around 95% of errors detected before hot commissioning |
| Wipro PARI | Automotive assembly line | Commissioning time reduced by 70% |
| Pringles manufacturing | Process optimization | Production-line capacity increased 10% without additional hardware |
| Open Grid Europe | Hydrogen production | Optimization work indicated production-cost reductions of around 50% |
| Unilever | Product twins for content production | Workflows 2x faster, production costs reduced 50% |
| Krones | Manufacturing simulation | Some simulations cut from 3–4 hours to 5 minutes or less |
| Foxconn | AI-server factory | Production lines and robotics simulated before physical deployment |
Our report covering the digital twin market gets updated several times a month, so you always get the latest view of where things stand.
Do the big software companies really make enough money to support a $40 billion digital twin market?
Yes, the vendor numbers support a broad digital twin market around $40 billion today, although they point toward a much smaller figure for pure digital-twin software.
Siemens is a good example of why the distinction matters. Its Digital Industries business generated €17.8 billion in fiscal 2025, including €6.17 billion from software. Digital twins sit across that software and automation stack, but Siemens does not report a separate line called "digital twin revenue."
Dassault Systèmes generated €6.24 billion of total revenue and €5.64 billion of software revenue in 2025. Industrial Innovation, which includes major products such as CATIA, SIMULIA and ENOVIA, produced €3.13 billion.
Bentley generated $1.50 billion in 2025, with $1.38 billion coming from subscriptions. Its infrastructure products increasingly use digital-twin concepts, yet Bentley is still much broader than one digital-twin product.
PTC provides the same problem through PLM, CAD, IoT and industrial software.
Put these companies together with Ansys, Autodesk, AVEVA, Microsoft, AWS and specialist vendors, and a digital-twin ecosystem worth tens of billions is easy to believe.
The same numbers make a $40 billion pure software estimate look stretched. Too much of the revenue at the biggest suppliers still comes from products and workflows that serve a wider engineering market.
How fast is the digital twin market really growing?
A growth rate around 25% to 35% a year looks more believable for digital twins now than the nearly 50% CAGR found in the most aggressive forecasts.
MarketsandMarkets expects 47.9% annual growth between 2025 and 2030. Global Market Insights is also above 40%. At those rates, the market multiplies several times over in only a few years.
Other current estimates are less extreme. IMARC is around 25%. Grand View Research projects 31.1%. Mordor is at 35.95%. Fortune Business Insights is at 35.4%.
The companies selling the underlying software are growing more slowly. Bentley's revenue rose 11% in 2025. Dassault Systèmes grew total revenue by 4% in constant currencies. Siemens' Digital Industries software revenue actually fell slightly against a difficult prior-year comparison.
Digital twins can still grow much faster than those companies because the technology is taking a larger share of their portfolios, implementation services are expanding and newer suppliers are entering the market.
Still, sustaining 45% to 50% growth across a market already worth tens of billions would require an extraordinary amount of new spending every year.
For now, a model in the high-20s or low-30s is more defensible than assuming the most aggressive forecast plays out.
If you want more recent data on this point, please see our latest digital twin market report.
In our research about digital twins, you will always find fresh funding data, taken straight from our own verified funding database.
Is AI making the digital twin market much bigger?
AI is already making digital twins more useful, and the latest product activity suggests the two markets are becoming much more closely connected.
The clearest recent change is physical AI.
A factory twin can now serve as a training and testing environment for robots and AI agents before those systems operate around real workers and expensive equipment. NVIDIA is pushing this hard through Omniverse, Isaac and its newer AI-factory architecture.
Its current industrial-facility workflow combines real-time operational data, physically accurate simulation and synthetic data for robot training. NVIDIA has also made its Omniverse DSX Blueprint generally available for designing and operating large AI factories, with Siemens, Dassault Systèmes, PTC, Schneider Electric, Cadence, Jacobs and other industrial suppliers contributing to the ecosystem.
Dassault Systèmes is seeing early traction for AI-powered virtual twins as well, and its 3DEXPERIENCE and cloud revenue grew faster than the group's overall revenue in 2025.
This gives digital twins another job beyond monitoring physical assets. The same environment can help train machines that eventually interact with those assets.
That could become one of the strongest growth engines for the category over the next few years, especially in factories, warehouses, robotics and data centers.
Revenue accounting still needs discipline. A GPU used to simulate a factory also runs many workloads outside digital twins, and the same is true for cloud infrastructure and AI models. Physical AI can drive digital-twin demand without turning the entire AI infrastructure market into digital-twin revenue.
Can the digital twin market really reach $150 billion by 2030?
A $150 billion digital twin market by 2030 is possible, but it sits toward the aggressive end of what the current revenue base can support.
MarketsandMarkets reaches $149.81 billion by 2030. The Business Research Company is closer to $122 billion. Other firms reach similarly large figures a few years later.
Starting with a $30 billion 2025 market, getting to $150 billion in five years means multiplying the market by five. That works out to roughly 38% annual growth.
A current market around $40 billion makes the climb easier, but $150 billion would still require huge increases in deployment across manufacturing, energy, infrastructure, healthcare and other sectors.
There are credible ways to get there. Digital twins could become standard tools for factory design and commissioning. AI and robotics create more demand for simulation environments. Infrastructure operators can connect more physical assets to live models. Cloud deployment makes sophisticated twins easier to operate across multiple sites.
The risk sits in how much spending is genuinely incremental. If much of the future "digital twin market" is existing simulation, PLM, IoT and cloud spending being relabelled, the headline market can expand much faster than the economic opportunity available to dedicated twin vendors.
We would treat $100 billion-plus around the end of the decade as credible under a broad definition. Reaching roughly $150 billion needs a stronger adoption curve and continued expansion of what researchers count inside the category.
If you want more recent data on this point, please see our latest digital twin market report.
In our market report about digital twins, we look at what the winners do differently, so you can learn from the best instead of guessing.
How much bigger is the ecosystem around digital twins?
The technology ecosystem influenced by digital twins is already worth hundreds of billions of dollars, which is exactly why ecosystem figures should be kept separate from digital-twin revenue.
An industrial twin can depend on sensors, PLCs, industrial networks, edge computers, GPUs, cloud infrastructure, CAD, simulation software, PLM, AI models, cybersecurity, visualization tools and systems integrators.
A manufacturer building a new factory may genuinely see all of those expenses as part of one digital-twin project.
The suppliers record the revenue very differently. NVIDIA books compute revenue. Microsoft or AWS books cloud revenue. Siemens can sell automation equipment and software. Accenture can book consulting revenue. A sensor manufacturer records hardware revenue.
Adding all of it together tells us how much economic activity a digital-twin architecture can influence. It tells us much less about the actual size of the digital-twin product market.
That distinction will become even more important as AI factories and robotics expand. The surrounding infrastructure can easily cost far more than the twin software controlling or simulating it.
For investment or competitive analysis, we would keep three layers separate: roughly $20 billion for core software, around $40 billion for today's broader digital-twin market, and a much larger technology ecosystem influenced by digital twins.
What is the real market size of digital twins today?
The best current answer is about $40 billion globally, with roughly $20 billion of that representing the core software opportunity and a reasonable broad-market range of approximately $35 billion to $50 billion.
That conclusion is stronger now than it was from the older estimates because the latest market updates continue to disagree in a very revealing way. Fortune Business Insights currently sits near $34 billion. Mordor and Grand View Research are close to $49 billion. MarketsandMarkets still starts from a much lower 2025 base.
The disagreement itself tells us something useful: digital twins have become big enough to form a real technology market, while remaining too embedded in simulation, PLM, IoT, cloud and engineering software to produce one clean revenue number.
The vendor evidence points in the same direction. Siemens has more than €6 billion of annual Digital Industries software revenue. Dassault Systèmes has more than €5.6 billion. Bentley has passed $1.5 billion in total annual revenue. Digital twins touch significant pieces of those businesses, but nowhere near all of them.
At the deployment level, the technology is already producing measurable results in factories, infrastructure and energy. Lately, AI has added another major use case as companies use virtual environments to design AI factories, simulate robotics and train physical-AI systems before deploying them in the real world.
Approximately $40 billion is the number we would use when someone asks how large the digital twin market is today.
For a company selling dedicated digital-twin software, the more useful TAM is closer to $20 billion.
And when someone shows a digital-twin opportunity worth hundreds of billions of dollars, we would check immediately whether that figure has quietly absorbed the cloud, hardware, AI, IoT and engineering markets around it.
If you want more recent data on this point, please see our latest digital twin market report.
Our methodology
This analysis estimates the real size of the digital twin market by comparing market reports with different scopes, testing their implied totals against major software vendors, and separating dedicated digital-twin revenue from the much larger technology ecosystem around it.
We start with the market boundary. A digital twin needs an ongoing connection between a physical asset, process or system and its digital representation; a standalone CAD model or simulation is not enough. We used the Digital Twin Consortium, NIST and ISO 23247 as the main references for that distinction.
We did not mechanically average market reports whose definitions may differ. Instead, we compared the distribution of estimates from MarketsandMarkets, Fortune Business Insights, IMARC, Global Market Insights, Grand View Research, Mordor Intelligence and The Business Research Company, then looked at why the high and low figures diverge.
The 2025 estimates are used as the cleaner common baseline because they let us compare firms on the same year. The current working figure then reflects how the market moves into the next year without pretending that any one report provides a precise answer.
For the software estimate, we used reported software or solutions shares as anchors, then adjusted for the fact that some categories also include connected hardware or other enabling technologies. General CAD, PLM, cloud infrastructure, sensors and consulting only count when there is a defensible digital-twin connection.
Vendor revenues are used as a plausibility check, not as a direct market-share calculation. Siemens, Dassault Systèmes and Bentley all report large engineering-software businesses that support the scale of the market, but none reports a clean standalone digital-twin revenue line.
Deployment examples are used to show where the technology is already producing measurable operational results. We relied on first-hand or company-documented examples from Siemens, Electrolux, Ferrero, Pringles, Krones and NVIDIA/Foxconn, while treating those results as project evidence rather than universal ROI assumptions.
Growth forecasts are treated more cautiously than market-size estimates. We compare the published CAGRs with the actual growth rates of major suppliers and with the amount of new annual spending that a 40% to 50% growth path would require once the market is already worth tens of billions.
Key sources used for this analysis include: Digital Twin Consortium on the definition of a digital twin, NIST on digital twin definitions and the state of the art, ISO 23247-1:2021, MarketsandMarkets, Fortune Business Insights, IMARC, Global Market Insights, Grand View Research, Mordor Intelligence, The Business Research Company, Siemens' 2025 annual report, Dassault Systèmes' FY2025 results, Bentley Systems' full-year 2025 results, Capgemini's digital-twin research, NVIDIA's Foxconn digital-twin case study, and NVIDIA's Omniverse DSX announcement.
If you build or invest in the digital twin market, you need to know which business models pay off now and which ones are fading. That's why our report covering it has a whole section on it.
Written by the New Market Pitch teamWe track new markets so founders and investors can move faster. About us
