Facebook

Twitter

LinkedIn

Protecting Designers’ IP in the Age of 3D Modeling

We built a generative AI prototype for 3D models and scenes in 2 weeks using only 8,000 models.

Creating Equitable 3D Generative AI

Physna is a cloud-based software that understands 3D models based on their precise object geometry, turning a model’s unique geometry into code. 

For manufacturers with tens or hundreds of thousands of parts and disparate CAD, PLM, and PDM systems that aren’t always integrated with one another, finding the necessary parts is a hassle requiring legacy knowledge. That’s because simple, Google-like searching isn’t always possible due to poorly organized or labeled 3D data. 

Identify — one of Physna’s prediction models — automatically generates “smart tags,” or metadata, for every 3D model. 

Metadata is the data providing information about one or more aspects of the data. For 3D models, predictive metadata is information like understanding two differently named parts are the same, using the geometry of the part. 

Identify uses the geometry of an object to predict metadata about a model.

For example, if an engineer searches for a model of a wingnut, Physna will recognize the geometric similarly to another part that might be labeled “butterfly nut,” and suggest “wingnut” as a smart tag, so that when future users search either term, all relevant parts will return as search results.

This is true for individual parts, like a wingnut/butterfly nut, but Identify will also recommend smart tags for larger assemblies or products of which the wingnut/butterfly is a component. That way, when users search for “wingnut,” all similar parts and larger models with wingnuts or butterfly nuts will appear in search results. 

Identify can predict any metadata so long as 1) the company tracks the data for some of their models and 2) that data point is related to the geometry of the part or its components. 

Identify enriches 3D models with 2D data, making associations previously never possible, making it easier for anyone to 

Stop wasting time searching for models and parts you already have. 

Physna 101: Identify

Physna’s technology can significantly reduce design hours, lower procurement costs, and enable an enterprise to immediately identify models and parts using 3D models, text, or images.

For a manufacturer with tens or hundreds of thousands of parts, models inevitably contain common components, like screws. Manufacturers might have dozens of the exact same part called different things, built for different equipment, made at a different time, with different file names.

Redundancy in 3D models means duplicate parts. This strains design and manufacturing teams who have to differentiate between similar models without the intelligence to do so, or rely on manually inputted descriptions of the models.

It also exacerbates supply chain challenges with unnecessary duplicates in inventory. Beyond the supply chain, duplicative inventory can result in superfluous certification for duplicate parts, redundant training for employees, and more. 

The easiest way to think of Match is “find me other parts like this,” also known as a "part-to-part" comparison. After 

 a model, Match will find other 3D models that are a geometric match to the model. 

So if you search for a screw, Match will show you all other screw models that are geometrically similar to the screw you searched for, and how similar they are, expressed as a percentage. 

Match will find similar parts regardless of file name, file folder, or file type. If you have it, Physna will find it, quickly and accurately providing actionable results.

With Match, engineers and designers spend their time building, because they aren’t stuck looking for 3D models or creating duplicates. Supply chain and procurement operations are simplified and overall inventory is reduced by identifying duplicate parts and consolidating operations. 

Match makes manufacturing simpler by easily identifying alternative parts, giving everyone the confidence they’ll work as well as the original part. 

Physna can do more than help you identify alternative parts; our software can also tell you everywhere that part is used with 

 into the similarities and differences between geometrically similar models.

Physna 101: Match

Physna is a cloud-based software that understands 3D models based on their precise object geometry, turning a model’s unique geometry into code. Physna’s technology can significantly reduce design hours, lower procurement costs and enable an enterprise to immediately identify parts and how they can be used. 

For a manufacturer with tens or hundreds of thousands of parts, models inevitably contain common components. 

Take a model of a bearing. It’s challenging to know how many other 3D models, including their subassemblies, also include that — or geometrically similar — bearings. A manufacturer likely has similar bearings with different file names, built for different equipment, made at a different time.

By using Part Match for a bearing, you’ll see that Physna also identifies a turbocharger. That’s because the bearing is in the impeller axle, which is part of the turbocharger. 

Part Match also works the other way — you can start with an impeller axel and identify alternate bearings that would work in the impeller axle assembly.

Physna turns the unique geometry of a model into code, allowing Part Match to find 3D models in which your model is located (the forward match) and 3D models located in your model (the reverse weighted match). Let’s take a look at an example.

 with a 3D model of valve bottom, Part Match shows us two things.

The first is the forward match — the entire valve assembly has a 100% match, because the valve bottom is a 100% fit as a component of the valve assembly.

The second is the reverse weight match. The valve assembly is a 64.62% match to the valve bottom, because the valve assembly has more parts than just the valve bottom. 

This works without a parts list, bill of materials, or metadata. 

 with Jay Marshall, Physna’s Global Head of Value Engineering.

Part Match exponentially speeds up design time, allowing products to hit the market more quickly. Start with a partial 3D model in Physna and find subassemblies that will fit into your model. Let’s go back to the partial valve assembly. Physna finds the valve bottom with a 100% match, even if it’s not part of the valve assembly — it’s autocomplete for 3D designs. 

Part Match also gives engineers the easiest way to determine how parts work in conjunction with one another and provides alternative subassemblies for a larger 3D model. So, if you have a partial valve assembly, engineers can see all possible valve bottoms and how they’ll fit, not just the one with a 100% match. 

Beyond finding alternative assemblies, leverage 

Give your engineers the tool they need to 

 and get your products to market more quickly.

Physna 101: Part Match

Physna’s technology can significantly reduce design hours, lower procurement costs, and enable an enterprise to immediately identify models and parts using 3D models, text, or images. 

Manufacturers have thousands of 3D models and parts, some of which are duplicates or geometrically similar to one another. The exact same part might be called different things, designed for different equipment, made at a different time, with different file names.

Unintentional redundancy strains supply chains — companies are importing more than they need — and taxes logistics and procurement managers. When supply chain crises arise — and they will — understanding alternative parts already in stock that can be reused can reduce the crises’ impacts on businesses and consumers. 

However, engineers and designers struggle to easily identify replacement or alternative 3D models with confidence that they’ll fit in larger assemblies or subassemblies

Compare in Physna — great for engineers assessing the difference between models — allows users to easily see the differences between two, color-coded models. 

Physna has two primary compare functions.

First is Overlay Compare, which is best for assessing external differences. The similarities are shown in blue, and the differences in red.

Leveraging Overlay Compare is the easiest way to identify alternative parts for an organization’s entire 3D asset repository. 

For a deeper comparison, Physna also has Mesh Compare

 shows users how one 3D model fits into a larger assembly or subassembly, Mesh Compare is great for comparing 3D models with internal, part-to-part differences. Mesh Compare doesn’t highlight the differences. Instead, Physna creates a new model from scratch of what’s similar between the two 3D models being compared. 

For example, to understand how a bolt that’s inside a turbocharger relates to the turbocharger, watch the video below.

Leverage Mesh Compare (because the bolt is inside the turbocharger), and Physna will create a new model of the bolt – because that is the similarity between the two models. 

With Physna, engineers no longer need to guess how 3D models are related to one another and can easily identify what’s similar and different between two models. This knowledge — where and how 3D models are the same and different — makes identifying or designing replacement parts simple and easy.

Know with certainty how your models relate, and 

 with Physna’s Global Head of Value Engineering, Jay Marshall. 

Physna 101: Compare

For manufacturers with tens or hundreds of thousands of parts, collating parts into categories is either an arduous, manual process or not done at all.

At best, manufacturers have a team of people updating spreadsheets manually, labeling parts with categories, hoping they can categorize parts faster than new, uncategorized parts are added to disparate databases. At worst, parts are never categorized because the task is too monumental to do manually, or the data was incorrectly entered and can’t be trusted. 

Unorganized or messy 3D data isn’t easily searchable, which leads to frustration as well as wasted time and money. Imagine a new engineer is designing their first axle, and they need to find a 3D model for a butterfly nut. Without Physna, that part must be manually categorized as such or labeled as a butterfly nut. 

If it’s not — maybe it’s labeled as a wing nut, but it’s the same part — the engineer won’t find it. 

So they’ll design a new butterfly nut, which adds one more duplicative part to inventory, creating a new set of costs beyond the cost of creating a new design — new supply requirements, excess inventory, new product liabilities, and potentially new insurance needs. 

Classify is powered by reinforcement and supervised machine learning algorithms, helping organizations classify their parts, 

Classify can clean up enterprise’s data sets everywhere. Physna can connect to existing Product Lifecycle Management (PLM) or Product Data Management (PDM) systems, so as Physna classifies 3D models, it can automatically update those classifications in PLM or PDM software.

Classify is one of Physna’s predictions engines that determines the category of a part regardless of scale.

 3D models to Physna, and Classify categorizes models based on a number of factors. Users verify the categories, and upload more models. Based on users’ input, the machine learning algorithms’ classification improves. 

Over time and at scale, the classification process becomes essentially automated as the reinforcement and supervised learning algorithms learn about a company’s data set — its 3D models.

Find the 3D model you need when you need it

Classify enriches 3D data with 2D labels — like easily searchable text strings. For engineers, this means Google-like search for 3D models.

Let’s return to the butterfly nut example.

With Classify, if an engineer searches for “wingnut,” but the part is labeled as a butterfly nut, Classify will have grouped these parts together, allowing engineers to easily find the part they need.

Classify works with 3D models, too. Searching with a 3D model of a butterfly nut will return all results for classified butterfly nuts (like a wing nut), regardless of scale. 

Gone are the days of re-designing a part because an engineer couldn’t find it in archaic, unorganized software. Physna helps you eliminate duplicative models, reduce inventory, and simplify procurement and manufacturing operations. 

Want to start at the beginning? Learn the basics about 

Physna 101: Classify

Physna is a cloud-based software that understands 3D models based on their precise object geometry, turning a model’s unique geometry into code. Physna’s technology can significantly reduce design hours, lower procurement costs, and enable an enterprise to immediately identify models and parts using 3D models, text, or images.

3D models in assemblies can be siloed across different tools in a variety of file formats. Without an up-to-date or easily accessible assembly tree, it’s challenging for engineers to always know — with certainty — which components make up a larger 3D model. 

Component Finder can identify all the various parts of a 3D model and their locations, showing alternate parts and subassemblies that fit into it, without a bill of materials. Component Finder works without an assembly tree, and it can still identify the components, regardless of whether or not the components are in an assembly tree. 

Once you’ve put a model into Component Finder to see all the various components and their locations in a 3D model — with or without an assembly tree — you can take a component that you've located and use 

 to assess if another part from a different assembly would be a fit for the original model.

This means Component Finder can identify how parts in a 3D model are related without running a search and then running a comparison.

With the color coded heat map, you can see there are multiple ways to put the top of the water valve together, either as a single part or components that fit together. Physna can also assess differences between valve bottoms; two match almost completely, but the width of the stem doesn’t quite match (indicated by the color threading). 

Stop manually searching for components of models

Engineers can ditch the spreadsheets and stop manually hunting for parts of complex 3D models — product and model design becomes simpler by being able to easily identify a model’s components. 

Procurement and logistics managers have all the parts for all their products in one place, allowing them to compare components to find alternatives in inventory, or dive deeper to find alternative suppliers for a part. 

Repair technicians in the field seamlessly identify all the parts of a broken piece of equipment with just a photo, making part replacement a breeze. 

Physna 101: Component Finder

Physna is a cloud-based software that understands 3D models based on their precise object geometry, turning a model’s unique geometry into code. Physna’s foundation is its one-of-a-kind 3D model search capabilities. 

Before Physna, all search was based on text — like a part number or SKU — or occasionally, rudimentary shape-searching capabilities. Trying to find a part by its SKU or part number is, at best, time consuming and unintuitive. 

Engineers struggle to find what models they already have in clunky, siloed software, so they waste time on redesigning parts for which models might already exist. Supply chain leaders don’t have a transparent view of their inventory, often importing and manufacturing duplicate parts, further straining logistics operations. And maintenance professionals struggle to identify replacement parts, leading to downtime for manufacturers that impacts their bottom line.

Take a 3D model of a turbo charger for an engine. It can have dozens of subassemblies and parts that make up that model. Each subassembly can have a bevy of individual parts, down to something as small as a 3D model for a bearing.

Finding one bearing buried within assemblies and subassemblies has traditionally been incredibly challenging, but it’s a common problem. For a manufacturer with tens or hundreds of thousands of parts, models inevitably contain common components, like bearings. Manufacturers might have dozens of the exact same part called different things, built for different equipment, made at a different time, with different file names.

None of that matters to Physna. Gone are the days of remembering SKU or part numbers, or hoping files are correctly labeled, to find the model or part needed. 

Users can search for models in three primary ways:

Leverage an existing 3D model with Physna’s search engine and find that part and geometrically similar parts, meaning you’ll find your model and alternatives

Upload an image — a photograph or thumbnail — and find the model you need

Using the bearing as an example, let’s dive in. 

Searching by leveraging a 3D model of a bearing shows the bearing and geometrically similar models.

The search result under Part Match also shows a turbocharger, because the bearing we searched for is part of an impeller axle, which is part of the turbocharger. That’s because Physna’s technology can identify a 3D model within a 3D model within a 3D model, allowing users to search not only for a given model, but also understand everywhere that model is used. 

Search by uploading a photo or thumbnail of a bearing into Physna, and the result is the same — Physna will find that bearing and geometrically similar models, as well as every design in which that bearing is used.

Physna’s search capabilities reduce design time for engineers because they can find models more quickly, streamline logistics operations by giving procurement managers complete insight into what parts they have, how many, and where — regardless of file name or type — and improve maintenance speed by finding the part(s) needed with just a photo.

Physna’s search capabilities work right out of the box, supports all major CAD file types, and does not need a bill of materials, parts list, or metadata. 

 into how Physna’s search engine can help you reduce duplicate models and streamline inventory with 

Physna 101: Search

Whether we’re working with a Fortune 100 manufacturer, a robotics startup, or a small design firm, the first thing we do is ingest and index 3D models to make models searchable in the first place. 

Traditionally, technology that indexes images or 3D objects to make them searchable hasn’t worked well. Even the best computer vision software — technology that’s meant to recognize objects — is accurate but not precise. In practice, this means computer vision can differentiate between a dog and cat well enough, but is not good at identifying if the dog is a Doberman or a Rottweiler or a German Shepherd. 

Manufacturers and the mechanical engineers who design their products need accuracy and precision. It’s not good enough to search for a screw and find every single screw. Users want to find one specific screw, or screws that are geometrically similar enough to act as a functionally equivalent replacement in larger models. Without precision, there’s no guarantee an engineer would find the 3D model they need, or that the part in question would work in larger products. 

Physna solves for precision and accuracy when we ingest and index models, using our proprietary technology.

Physna ingests 3D models from CAD, PLM, or PDM software and segments each object into very, very small pieces. The software then takes each tiny piece and gives it a unique code. So much like letters in a certain order spell a word, Physna can recognize that unique codes in a certain order denote a specific geometry of a physical object — making physical objects indexable and searchable.

 functionality can both specifically identify discrete parts and larger 3D models of which that part is a component — we will find the unique geometry, based on the unique code, no matter where that code (or 3D object) is. 

The entire process is automated. Upload CAD models in Physna or integrate Physna into your workflow with our APIs, and Physna does the rest. Physna supports all major CAD file types, and does not need a bill of materials, parts list, or metadata. 

Learn more about Physna’s unique-in-market 

Physna 101: Ingest

 affects suppliers, consumers, and everyone in between who interacts with the physical world. Nearly every physical object — and the machines that build those products, like robots on an automotive assembly line — are all designed digitally as 3D models before the parts are shipped and the product is manufactured. 

For manufacturers who want to develop more resilient supply chains, start with optimizing for:

The people who design, manufacture, and service parts and products; and

Suppliers who provide parts for products. 

When manufacturers face supply chain challenges, it’s easy and accurate to look at the geopolitical landscape as the root cause of those disruptions. And while that’s true, it’s next to impossible for any single enterprise to control the geopolitical landscape or predict the future with any certainty. 

However, organizations can build a more resilient supply chain by focusing on the people who design and develop products, like engineers, supply chain leaders, maintenance professionals, and factory employees and managers. 

Often underestimated are the downstream impacts on supply chains as a result of how these different people and the tools they use do — or more often, do not — integrate well or permit easy collaboration.

Engineers use tools like AutoCAD, supply chain managers use software like Anvyl or Windchill, and employees on the ground in factories use everything from Oracle to spreadsheets. 

All of these tools break up the way technology collaborates because they view digital, 3D models — the foundation for all physical objects in the real world — differently. 

Manufacturers often have dozens of the exact same part called different things, built for different equipment, made at a different time, with different file names, but a lack of software interoperability means enterprises don’t know it. 

Physna brings the people who design and develop products together

Physna enables collaboration at the technology level based on 3D models by creating a single source of truth that’s accessible across all existing systems of record, significantly reducing design hours, lowering procurement costs, and increasing supply chain resilience.

Physna is the only technology that understands, finds, and classifies 3D models, turning their precise and unique geometry into code. With Physna: 

Engineers reduce design time because they can easily find the models they need and identify functionally equivalent parts to more easily reuse models, rather than needlessly designing duplicate models that complicate inventory management, procurement, and logistics operations. 

Supply chain leaders can identify common components between 3D models — even if those parts are built for different equipment, made at a different time, with different file names — streamlining inventory and consolidating suppliers.

Factory employees and maintenance professionals can seamlessly identify any part of a broken or deficient piece of equipment with just a photo, and share that information with engineering and supply chain counterparts. 

Make smarter procurement decisions. Save on procurement, product support, and life cycle costs and eliminate part redundancy and unnecessary costs. 

Let Physna’s Global Head of Value Engineering, Jay Marshall, 

part II of the supply chain resilience series

, where we’ll discuss supplier optimization

How to develop resilient supply chains: Part I

At Physna, we spend a lot of time talking about 3D. In fact, we spend virtually all of our time talking about 3D. It is at the heart of everything we do. One of the core concepts that we discuss with our customers is the need for deeper visibility and better understanding of their 3D data. This type of leverage ultimately leads to improved productivity, lower costs, and reduced downtime by making their 3D data searchable, accessible, and actionable.

At Physna, we spend a lot of time talking about 3D. In fact, we spend virtually 

 of our time talking about 3D. It is at the heart of everything we do. One of the core concepts that we discuss with our customers is the need for deeper visibility and better understanding of their 3D data. This type of leverage ultimately leads to improved productivity, lower costs, and reduced downtime by making their 3D data searchable, accessible, and actionable.

“3D” itself obviously isn’t a new concept. It has been around for decades and used by engineers and designers to develop the products we use every day. But with increasing innovation in the worlds of digital twin, augmented reality, and now even the metaverse, 3D is quickly becoming the epicenter of all of this innovation.

 - the creator of the digital twin concept - just released a blog post entitled “

Driving Digital Comparisons into Another Dimension

” that dives into the need for 3-dimensional search in this modern age of digital twin and the future of manufacturing. But to truly understand the differentiation of 3-dimensional search, we must first understand the difference between 2D and 3D data.

Some folks associate 3D data with animation and modeling programs used in film and gaming. Others may think of 3D CAD where the focus isn’t as much on finely-detailed textures and visual aesthetics, but rather the digital technical drawings developed by engineers that are required for the manufacturing of products. Both associations are correct. But for today’s discussion, we will be focusing on 3D models and objects associated with CAD use cases.

The difference between 2D data and 3D data

The key difference between 2D data and 3D data is in the way it’s structured and accessed.

When we discuss “2D data” with customers, we are referring to the standard systems of record within the existing enterprise, which also extends into file and object storage. The application and insights derived from this data via data warehousing, and more recently data lakes, also fall under this category of 2D. Although the data storage and retrieval strategies have evolved drastically over the last half century, the data itself remains largely aligned with the earliest days of binary as it pertains to business data. Strings, numbers, and other characters are stored to disk and retrieved via searches using strings, numbers, and other characters.

Access to 3D data does not work the same way.

For years our products and parts have been locked behind 2D search constructs like part numbers, SKUs, BOMs, and text strings to access them, as the 3D models themselves are simply binaries that only get unlocked with CAD software for use within that software. This has resulted in extreme inefficiencies that, unfortunately, extend beyond the design stage. Up until now, there has been no way to compare products, assemblies, and parts at the 3-dimensional, geometric, level later in the manufacturing, maintenance, inventory, and logistics stages of product development. Take the following two part names for example:

Can you tell whether the above two parts are exactly the same or not based solely on those part names? Of course not. And neither can a computer. Let me make it a little bit easier:

Certainly the fact that they are completely different file types (“.par” vs. “.sldprt”) they must be different parts, right? Not so fast. Here is a screenshot of the 

 two parts next to each other inside of Physna:

Those file extensions were different simply because the files were created by 

. But we can see that when we look at the actual parts side by side they are in fact the same, in spite of both their file names 

Let’s now turn the parts slightly and view them from a different angle:

 that I can truly see that the two parts are in fact different. There is no search string in the world that could clearly identify this difference. We have even seen customers where they have had multiple part names be the same, even though the actual contents were different. It is only via the embedded geometric differences between the two parts that these can be seen as truly being different.

From a business value perspective, believe it or not the challenge is many times not even about parts being 

 that some of the biggest inefficiencies can creep in. Our customer conversations have uncovered literally billions of dollars of potential cost takeout if things like duplicate parts and excess inventory can be resolved. And they can. So even if ABC and XYZ 

 exactly the same in our example above, many organizations have to keep stock of both because they are both technically in use according to their CAD, PLM, and ERP systems. But if they are geometrically identical (along with material and other properties of course) these parts could be consolidated to help reduce inventories, which have immense downstream impact for a manufacturing organization.

This is where we come in. Physna has developed the first true 3D geometric search platform that makes it possible for you to search and action all of your 3D data in brand new ways - unlocking new business value directly from the products themselves.

 mechanisms in place besides text searching within large manufacturing platforms to aid in finding parts (e.g. dimensional characteristics, bounding box volumes, shapes, etc.). The good news is that Physna allows for this as well. But many of these constructs simply do not maximize the productivity and output that true geometric search can provide:

Customers tell us that at least 15% or more of their designers’ time is spent searching for existing models for use in their product designs. This number makes sense if you take into consideration our examples above. So if you’re a designer or an engineer, Physna allows you to spend less time searching for models and more time creating great products and experiences.

If you’re a supply chain or procurement manager, locating the right parts at the right price and minimizing duplication and excess inventory is key. Physna helps reduce inventory levels and provide strategic purchasing leverage by providing alternatives you may not have realized you had, reducing risk and enhancing supply chain resiliency in the process.

If you’re working in maintenance, locating the right part quickly is critical to reducing downtime and keeping your lines up and running. Physna’s 2D-to-3D photo search capability can reduce the time finding replacement parts by up to half or more by taking pictures of physical parts and leveraging the underlying 3D model to quickly return stock locations for replacement parts.

The term “digital transformation” has been used across many industries and many verticals for many years. From marketing and advertising, to application development and cloud computing, this term has been used anywhere where high-levels of digital innovation have led to significant business impact. We feel that 3D is one of the last bastions of true digital transformation.

Imagine taking any product anywhere in the product development lifecycle -- from concept and design to manufacturing and maintenance to logistics and the supply chain -- and at any stage be able to search, compare, and substitute parts of that product with other available parts across the entire supply chain based solely on the unique geometric properties of the products, assemblies, and parts you are looking for. This is just one example of what is possible once we are able to harness the value of 3D data. It is now the gold standard for everything from physical products, video games, and animated characters; to virtual reality, augmented reality, and even the metaverse. In the upcoming months we will be releasing a series of videos, blog posts, and other materials unpacking these and other examples; and I hope you will join us on this exciting journey of 3D digital transformation.

2D vs. 3D data​: Why they matter to your business

2021 had a common theme for many: COVID made it difficult to get together in-person, and remote work became the norm for millions. The limitations of virtual collaboration made it even more clear that the gap between the physical and digital needed to be bridged. Technology is recreating and revolutionizing how people do everything from participating in everyday meetings, to 3D printing for everything from gadgets to entire houses, to manufacturing some of the most complex machinery and vehicles in the world. We’re in a new age of a virtually-enhanced reality where 3D technology enables people to organize, analyze, build and experience things in entirely new ways. 

Facebook, now Meta, is making a significant bet on an immersive virtual-physical Metaverse, with Microsoft, Apple, Alphabet, and other prominent companies investing heavily in augmented reality. Unlocking the true potential of 3D, and in doing so, touching everything – supply chain, procurement, 3D printing, gaming, mechanical engineering, and so much more across our physical and digital worlds — requires a robust technical foundation. Transformative technology such as spatial reasoning, geometric search, and deep learning can materially improve business’ operations and people's lives.

That’s where Physna comes in. Physna is bridging these physical and digital worlds. We’re building the technology that enables a future in which technology is immersive, spatially aware, and physically representative.

2021 was an exciting year for Phynsa. We entered 2021 with a new $20M round of funding led by Sequoia with participation from Drive Capital, and raised an additional $56M later in the year led by Tiger Global with participation from Sequoia, GV, and Next Play Ventures. With this new funding, we’ve been able to expand our leadership, grow our team and continue to invest in our patented core technology that turns 3D models into data and simple universal code, used by some of the largest manufacturers in the world and by creators on our consumer 3D search and collaboration platform, 

Physna’s enterprise solution saw the addition of API support for a variety of features, such as 2D<>3D search (the ability to identify a precise 3D/CAD part in your inventory using an image), automated part classification, and GDPR compliance. 

 – a version and revision control automation tool for 3D designers and engineers (imagine a “GitHub” for 3D) – launched in beta. Thangs also added AR part previews for iOS and Android with no app required. Both Physna Enterprise and Thangs 

. We also updated our brand identity to reflect the dynamism and dimension of the technology we build.

The past year has been one of tremendous momentum across both our enterprise and consumer products:

Physna Enterprise was adopted by major manufacturers in automotive, appliances and beyond to promote part reuse, reduce procurement waste, improve supply chain, and increase accuracy of part identification in the field.


Thangs grew its user base by over 500 percent in 2021, and increased its searchable model volume from roughly 1 million to well over 7 million.

The size of our organization nearly tripled in 2021 and we added exceptional talent to our leadership team, including 

 to lead comms and marketing. People have always been the most important part of Physna and we continued to show our support for the team and their families by improving our benefits to include fully-covered health insurance for our employees and their families, 401K contributions, monthly professional development and health and wellness stipends, up to 16 weeks of parental leave, fertility, adoption, and surrogacy financial support , and 3 weeks of miscarriage bereavement loss.

2021 was an exciting year for Physna, but we’re even more bullish on 2022. Physna has some of the best and most relevant technology for our changing world, rapidly expanding awareness and adoption, and most importantly, a world-class team with a simple goal: Enable the next generation of immersive, 3D technology as the universal platform for 3D search and collaboration.

A year of unstoppable momentum for 3D technology

recently had the pleasure of sitting down

 with digital twin visionary Dr. Michael Grieves to discuss the current and future state of digital twins. Dr. Grieves is Executive Director and Chief Scientist of the Digital Twin Institute and helped pioneer the very concept almost 20 years ago. Our discussion was eye opening — we are still in the very early stages of realizing the full potential of digital twins, and it can be difficult to separate the actual value from the hype. At Physna, we see great benefit in being able to search the world of 3D with 3D as part of a broader digital twin strategy, but for our first Physna Guide we wanted to focus on some of the fundamentals.

Manufacturers today are under pressure to deliver value faster and more efficiently, even when faced with continuing disruptions. Many are looking to adopt new digital tools that provide real-time visibility and actionable insights to achieve more with fewer resources and less uncertainty. At its core, digital twin promises to deliver these benefits and radically change how businesses operate in an increasingly digital-first world.

Digital twin promises to reshape the way companies approach planning, production, and change management, moving work into the virtual realm before taking action in the physical world. Instead of static blueprints of physical objects, digital twin provides interactive simulations and administrative tools that leverage real-time information flows to help users manage systems and facilities strategically. For example, clean energy companies that are concerned about the efficiency of wind turbines can create a digital twin that virtually models wind speed, revolutions per minute, energy created in kilowatts, and environmental variables.  That data can then be leveraged in the design of new wind turbines to increase efficiency by emulating physical changes to the virtual wind turbine. This allows engineers to burn “bits instead of atoms” as Dr. Grieves states, as atoms (i.e. physical materials, assets, products, etc.) are way more expensive to utilize than bits (i.e. computers).

Digital twin can help manufacturers as early as the initial product design and testing phases. In automotive manufacturing, designers can sketch prototypes of cars and use a digital twin to make essential design improvements and run tests to simulate driving scenarios through data sourced from Internet of Things (IoT) sensors. This tech empowers product designers to identify improvements in driving performance and safety before physically assembling vehicles. 

And digital twin isn’t limited to just manufacturers — in Building Information Modeling (BIM), architects and construction professionals can create a digital twin that runs scenarios for construction planning and facilities management, massively reducing construction costs and shortening project timelines. 

A digital twin can also streamline supply chain management to minimize bottlenecks significantly. By tracking KPIs like packaging performance and distribution routes, digital twin technologies can simplify and solve these logistical challenges before they occur through data analysis.

In understanding this broader scope, you can start to see the impact this concept can have as you stitch these previously isolated focus areas together. The key benefit of digital twin comes from its ability to empower users to manage and analyze data at a granular level, gleaning valuable insights about development, production, and logistics in real-time. This continuous access to data feeds means that a digital twin can dynamically adapt to and simultaneously reflect changes to its physical counterparts. In manufacturing, this could mean emulating supply chain impact when selecting part alternatives when designing a new product.  Whereas in medicine, a digital twin could provide valuable health data in real-time for health monitoring, such as by aggregating data from wearables to develop interactive patient records that adjust to health status changes.

Almost every industry is looking to incorporate digital twins to improve future business outcomes, with the digital twin market projected to be worth $48.2 billion by 2026. Yet most organizations today are only at the early stages of considering and implementing digital twin. To realize the long-term benefits, organizations must first identify the most achievable applications today and take steps to ensure that business processes along with IT and data capabilities are properly aligned for future integrations.

Key steps toward building a digital twin 

Given the many possible use cases and visionary language used to describe digital twin, it can be challenging to know where to start. Yet, like many other digital transformation initiatives, adopting digital twin begins with defining the purpose. This endeavor entails identifying which product or process could benefit from a digital twin. The initiative should only move forward if there is clear value to be unlocked. 


Begin by detailing how relevant procedures, people, applications, information, and assets interact, often in a diagram. Digital twin initiative leaders should then link these processes with data needs along with cost and time considerations to establish KPIs for return on investment. 

This is a crucial step for making the most out of a digital twin. When goal setting, you should identify the parts of your business that have room for improvement and then coordinate these objectives with data input from the digital twin. While digital twin should strive to integrate data from the entire product value chain, initial implementations should focus on a single component (like an assembly line part) or an essential business process. No ocean boiling here!

3) Choose digital twin software provider(s)

Next, find software providers that support product modeling and digital twin technology so that your company can create a digital version of the product or workflow. This is a tricky step that requires deep analysis to ensure the data being captured is relevant and thorough enough to meet your objectives. It may also require custom development, so keep that in mind when doing analysis and setting budgets. In the end, you need to ensure that you are modeling all necessary components and harvesting all relevant data to emulate your real world environment — from the initial design phase to distribution and even maintenance. The more thorough you are here, the more impactful your digital twin initiative will be.

Physna Blog

Protecting Designers’ IP in the Age of 3D Modeling

Creating Equitable 3D Generative AI

Physna 101: Identify

Physna 101: Match

Physna 101: Part Match

Physna 101: Compare

Physna 101: Classify

Physna 101: Component Finder

Physna 101: Search

Physna 101: Ingest

How to develop resilient supply chains: Part I

2D vs. 3D data: Why they matter to your business

A year of unstoppable momentum for 3D technology

Physna Guide: The expanding role of digital twins in manufacturing (and beyond)

How to develop resilient supply chains: Part II

The challenge of 3D data integrity for aerospace

Physna Blog

Protecting Designers’ IP in the Age of 3D Modeling

Creating Equitable 3D Generative AI

Physna 101: Identify

Physna 101: Match

Physna 101: Part Match

Physna 101: Compare

Physna 101: Classify

Physna 101: Component Finder

Physna 101: Search

Physna 101: Ingest

How to develop resilient supply chains: Part I

2D vs. 3D data​: Why they matter to your business

A year of unstoppable momentum for 3D technology

Physna Guide: The expanding role of digital twins in manufacturing (and beyond)

How to develop resilient supply chains: Part II

The challenge of 3D data integrity for aerospace

2D vs. 3D data: Why they matter to your business