The construction industry is one of the largest sectors on the planet. Yet even with an annual value estimated at $14 trillion, it has so far yet to experience digital disruption.

The potential is huge, too, and billions of dollars were invested in construction and property technology (contech and proptech) start-ups over the past five years, with the mission of sparking a revolution in how we create the built environment.

Yet what has been perceived as a field brimming with potential is still to produce the desired fruit. In this article, we will delve into the key challenges facing AEC start-ups and how these might be overcome.

In recent years, I have witnessed first hand the growth of the construction tech industry, through my own involvement in running a start-up, taking on advisory roles for government organisations and corporates, and working in the construction material commerce space. That’s given me an in-depth view of the huge potential here, and how far we are from filling the current gap. I present my views with hope that it can help AEC entrepreneurs and investors find some common ground on which to build.

Find this article plus many more in the May / June 2025 Edition of AEC Magazine

Subscribe FREE here

Why the delay?

So what makes construction so tricky and so seemingly immune to digital disruption as compared to other sectors? As I see it, construction tech start-ups face a number of key challenges, including:

Dinosaurs versus unicorns: Construction is a game that comes with a very heavy ticket price to play. Not only does it require a strong financial backbone and stamina, it also takes a lot of experience – one that cannot be bought, even with lavish funding. As opposed to the IT sector, where new start-ups can spring up to become market-leading mega companies with valuations measured in billions of dollars in just a few years, the building sector is dominated by veterans. Even in the construction software space, the market is led by well-established players offering tech stacks that are often 20 to 30 years old. This would seem like a natural breeding ground for a new age to emerge, where the Canva, Figma or Wiz of construction might flourish, but it is yet to happen. Let’s begin the analysis by seeing how the money is spread.

Construction start-ups are from Venus, venture capitalists (VCs) are from Mars: Traditionally, most VCs come from a software-first mindset, in which success is defined in terms of recurring revenue, user acquisition, and scalable SaaS models. They focus on large user numbers, lean, quick growth and owning categories. In contrast, construction start-ups are rooted in a world of physical infrastructure, where value is created through complex, capital-heavy projects with long lead times, tight regulations, and high execution risk. Even if they are software-oriented, the mindset is completely different, as we will explore later in this article. VCs expect exponential growth curves and monthly churn reports, while construction founders grapple with bidding cycles, procurement hurdles and delivery timelines measured in years, not sprints. This disconnect can make it difficult for construction innovators to translate their impact into metrics that VCs can understand. And that looks set to remain the case, at least until a new generation of investors emerges who are prepared to go deep into the screw level. Some construction tech oriented VCs have emerged over the last years such as Foundaental, Brick and Mortar, Building Ventures, Blackhorn Ventures, Zacua and Kompas. Some leading construction companies also make corporate venture capital (or CVC) available to promising start-ups. These include Cemex, Saint-Gobain and Vinci. However, we still need the ‘heavy lifters’ of the VC world to help scale the field.

Construction tech versus tech for construction: One of the crucial misconceptions of this market lies in the failure to distinguish between these two verticals. The term ‘construction tech’ refers to innovations in how we build things, focusing on construction processes, materials and methodologies. The term ‘tech for construction’, by contrast, is used to refer to pure technological applications, such as management software, Internet of Things (IoT) technologies, AR/VR and others. In other words, these are technologies that can be applied to construction, but are not specific to the field. But VCs tend to invest in the second category, the pure tech models, seeking the next Uber, Figma, Airbnb, Facebook or Tesla of construction. Their investments aim for disruptive, high-growth technology, but are often disconnected from the ground – or in our case, the building site. Fields like manufacturing, materials, and on-site work have always been less lucrative to tech investors, but this is where the greatest potential for sector-wise change lies.

Warning: hurdles ahead

So what are the biggest hurdles faced by construction tech start-ups? In my experience, the most prevalent barriers include:

1. The ‘butterfly effect’, where risk is greater than gain: In construction, time is money and projects are inherently risky. Each project is unique, involving numerous variables and stakeholders. For a start-up, developing a new solution might be straightforward and worth the risk of trying, but that’s only because they are not paying the price of their mistakes. A minor deviation from the plan can lead to significant delays and cost overruns in completely unexpected areas. The potential risks associated with untested technology are often too great for construction companies to bear. Therefore, start-ups must provide watertight solutions with a clear silo of their work scope that mitigate the risk.

2. Recognition that a one-trick pony is not a unicorn: Despite being a $14 trillion industry, the construction sector is a relatively small market for point solutions. These are specific, narrow applications designed to address particular problems. The number of practicing AEC professionals is estimated at between 3 million and 4 million, which translates to a limited customer base for specialised software solutions. It’s important to note that the AEC software market is valued at around $6 billion, controlled by centralised players. This discrepancy highlights the limited scope and scalability of many construction tech solutions. While a point solution might solve a specific problem efficiently, its applicability is often too narrow to attract widespread adoption to build a unicorn.

3. A monopolised software ecosystem: Due to the heavy nature of the industry, software start-ups face a difficult situation in which they can never replace the existing platforms, but only integrate. Players such as Autodesk, Nemetschek, Bentley Systems are so heavily rooted in the industry that start-ups must adapt their solution to mature infrastructure and formats that are sometimes not at the forefront. Closed formats create a very muddy puddle in which to play, while IFC simply does not provide enough of a format. This rough landing makes it difficult for new entrants to board. Start-ups must fight an uphill battle before even entering the ring. Creating non-intrusive on-boardings that don’t require switching and implementing a long learning curve can drastically improve chances of success

4. The difference between a service and SaaS: The days of software training and specialists are over. In the instant AI age, expecting users to retrain with complex new tools as in the past is unrealistic. With tech solutions popping up like mushrooms after the rain, users lack the skills to go deep and this often creates problems for anyone making a serious construction application. As a result, construction tech companies frequently need to provide comprehensive services alongside their product to tailor fit it and follow through. This service-intensive model can be very heavy and costly not just for the customer, but also for the start-up. Given the ‘plug-and-play’ expectations of the market, start-ups can easily find themselves in a swamp of unprofitable projects that far exceed their initial offering and remain unscalable. It is key to build products that are self-maintained and exercised by the client.

5. Makers developing for makers: A significant challenge in achieving product/ market fit arises when creators design solutions to solve their own problems, rather than addressing broader industry needs. This insular approach often results in products that resonate very much with a niche audience, but fail to gain traction in the wider market. For instance, an engineer might develop a tool to streamline a specific aspect of their workflow, without considering how many professionals face the same issue, or if the solution can be scaled. It may be so that 90% of their colleagues will find it useful, yet there are not enough of them to build a large business, or they can’t afford to pay the required prices. This type of company may be quick to build traction, but that can be very misleading.

6. Lengthy time-to-market: Construction projects often span two to five years. This extended timeline is fundamentally at odds with the rapid growth expectations applied to start-ups and their investors. Start-ups typically need to demonstrate exponential growth and achieve scalability quickly. However, the slow pace of the construction industry means that it can take years for a start-up to complete a real proof of concept (POC), regardless of how much funding it has raised. If a project spans three years, and a new tech is onboarded for trial, it could take a few years before the technology becomes a watertight solution and only then scan it start to scale. This protracted timeline hinders the ability to achieve the rapid growth necessary to reach unicorn status, deterring investors who seek faster returns.

Turning disadvantage to advantage

Having described the hurdles, let’s remember that along with barriers comes potential. The biggest advantage of construction is scale and stability. If it works, it’s huge! Yes, it takes a bit more time, but that time is won back as a competition barrier.

The market has proven that the building sector is one of the safest and most reliable markets for those with a competitive advantage. Start-ups offering true innovation can win big and become unicorns if they have the patience and foresight to understand their path. Seeking quick returns and hyper growth from the get-go is likely to lead to disappointment.

For me, the AEC unicorn formula is:

Onboarding ease * users * profit = Potential

This simple equation provides a quick analysis tool to predict the likely success of startups. It is simply a grading system based on three crucial metrics. Combine these together in the formula, and you will find a useful grading system to evaluate success.

The Unicorn formula for AEC

Onboarding (Grade factor 1-10)
Ease of adoption • Required support • Risks associated

Potential users
Total amount of users or projects. Take 10% of that.

Profitability
Estimated profit per user/project.

Ease of onboarding: This is perhaps the most crucial determiner of a go/no-go situation. As previously mentioned before, there is no time in today’s world for deep training and long integration times. Smooth onboarding and usage is the key to winning customers. This is calculated as a factor between 1 and 10, where 10 indicates the optimum ease of use.

Users: These are users/projects that show potential to become customers. The goal is to identify a realistic number of customers who are willing to pay for the tech product or service and assume 10% of that number will be won over.

Profit: This is a term that isn’t sufficiently talked about in a start-up world where everyone is chasing scale, but construction is a cut-throat business and if your product is not competitive or is just a ‘nice-to-have’ rather than a ‘must-have’, it will not survive. Profit margins should be watertight and based on competitive advantage.

Conclusion

As we step into the age of AI and automation, it is predicted that construction will transform itself, and with it the nature of its tools and the daily lives of its workers.

This is an exciting time in history and an opportunity to get on board. AI is not only being used in the tools that we make – it’s also being used to make them. That’s lowering barriers to entry like never before. For this reason, we can expect to see a big increase in digital solutions from developers of all sizes, opening doors that for years were thought of as closed.

I hope this article helps you position yourself in this brave new world, either as a user, an entrepreneur or an investor looking to take part in the digital transformation of construction. I am certain that if all six points are addressed, success is guaranteed. Best of luck to us all!

Tal Friedman will be presenting at NXT BLD on 11 June – ‘From AI visions to fabricated realities’

The post Why are there no unicorns in construction? appeared first on AEC Magazine.

Fabricated integrated modelling, or FIM, offers an innovative way to use digital technology to achieve a new economy of scale fit for the built world environment. A major issue that it aims to solve is the modular question. In short, as long as construction projects are conducted as a series of ‘bubbles’, thanks to fragmented data and supply chains, the same problems arise again and again.

Offsite construction has long been touted as the future of the construction industry, since way back in the late 1960s. It promises faster construction times, reduced waste and increased efficiency. However, despite the apparent advantages, offsite construction factories have never lived up to these promises.

But in recent years, there has nevertheless been a resurgence of interest in this area, as part of a wider construction technology movement. This has led to significant investments in new factories that claim to do things differently. In fact, it’s fair to say that the modular construction world has enjoyed something of an epiphany, backed by large investments from venture capital firms looking to profit from a revival of the concept. These investments have been spurred on by a view of construction as the next big opportunity for digital disruption. With the development of IT and automation tools, the timing seemed right for change, and governments were equally enthusiastic, pitching in to create incentives. It appeared for a while that a long-awaited transformation was about to happen.

So what went wrong? Why are we witnessing a slew of modular companies either going bankrupt or filing for chapter 11 bankruptcy protection? What happened to the built environment revolution that was apparently on the cards – and how might FIM get it back on track?

The rise and fall of Katerra was just the beginning. Soon, more and more investors joined the goldrush, lured by the potential rewards on offer for the company that finally succeeds in creating the ‘building factory of the future’. Unfortunately, these approaches have collapsed time and time again, leaving industry professionals still wrestling with the still unsolved modular question. How can something seemingly so logical on paper continue to be thwarted by reality? And if manual construction is outdated, slow and inefficient, how come it repeatedly beats offsite construction to the punch?

Find this article plus many more in the May / June 2023 Edition of AEC Magazine

Subscribe FREE here

Answering the modular question

In the hunt for a solution, a good place to start is by thoroughly understanding the problem. In this case, companies face a number of issues, typically summed up by the following:

To get scale, you need scale

Perhaps the biggest problem in the field of any automation, especially as heavy as construction, is the need for scale. Effective assembly lines only function well under standardised and repetitive conditions. Therefore, in order for any one project to be cost-effective, you need to have many projects in the pipeline.

Building twice

Building offsite essentially means constructing a building twice – once in the factory, and then again on site. That involves funding a large, spacious factory, equipped with high-cost assembly lines, even before the first commissioned project arrives. The upfront costs are huge and the return on investment is unclear. This leaves us once again tackling a scalability requirement.

The chicken-and-egg of distance

Transportation costs are another major challenge. It is extremely expensive and time consuming to transport modular buildings, especially volumetric ones, from factory to construction site. But the closer your organisation is to the target destination for building, the more you will pay in labour, factory rent and transport.

The customisation challenge

Not all buildings are created equal. Since most buildings are not designed for fabrication, most projects will be given permits based on spatial layouts rather than fabrication models. That means a complete rework to match the standardisation of offsite construction is needed, taking up a lot of time and extra cost.

The FIM solution

FIM presents a generative approach that can help solve the modular dilemma and increase the scalability and profitability of modular construction factories through smart planning automation. As artificial intelligence (AI) concepts unimaginable in the past unravel themselves on a daily basis, using FIM principles can unlock much of the long-awaited potential of modular construction.

FIM offers to standardise the construction process using a smart kit of parts that can be adapted to (almost) any layout. Using smart network effects, building plans optimised for smart supply chains can support mass customisation. Rather than limiting designs to fixed layouts, the approach seeks to standardise only what needs to be standard, while also permitting freedom of design where needed. Rather than fragmented point solutions for design, as in classic BIM, FIM promises to create a unified platform that supports the interests and activities of all stakeholders and where design and fabrication are linked at all times.

FIM can help architects and engineers create designs that are optimised for manufacturing from the beginning. Using generative AI to match design intent and real-world constraints means that designs can easily be translated into modular construction components in a seamless manner, without having direct expertise when it comes to all the manufacturing data. That allows them to focus their efforts on what really matters – the quality of design, rather than an endless game of Ping Pong with regulators, fabricators and other stakeholders.

Using generative AI to match design intent and real-world constraints means that designs can easily be translated into modular construction components in a seamless manner

Combining the power of AI with the creativity of human designers, FIM starts by defining the design requirements and constraints of a project. These requirements are then entered into the platform, which uses generative AI to create a range of design options that meet the requirements. Designers can then select the options that best meet their needs and refine them further. This process allows designers to explore a wide range of design options quickly and efficiently, reducing the time and costs involved.

FIM also incorporates DfMA principles into the design process, ensuring that designs are optimised for manufacturing from the very beginning. This means that they can be easily translated into modular construction components, reducing production costs and increasing efficiency.

Modular factories, meanwhile, can increase their scalability and profitability by supporting multiple projects that share similar attributes, and then optimising those for their facilities. Additionally, the speed and efficiency of the design process mean that modular construction factories can take on more tenders and approvals, further increasing scalability.

FIM versus BIM

There are six main verticals where FIM seeks to move BIM from a fragmented process supported by multiple point solutions to a uniform platform. It’s not about creating sketch pads and calculators, but about creating an informed design process fed by real-world data shared by all stakeholders.

1. Real-time data analysis: FIM provides real-time analysis of structural and thermal performance during the design process. This allows architects and engineers to make informed decisions and optimise designs for better performance and efficiency.

2. Efficient design iterations: With FIM, design iterations can be done quickly and easily. Changes to the design can be made in real-time, and the impact of those changes on performance can be immediately evaluated. This leads to more efficient design iterations and faster project completion.

3. Simplified collaboration: FIM simplifies collaboration between project stakeholders, including architects, engineers, contractors, and manufacturers. Because FIM is based on a shared data model, everyone has access to the same information and can work together more efficiently.

4. Improved accuracy: FIM uses a data-driven approach that relies on accurate and comprehensive information. This leads to more accurate modelling and better performance predictions.

5. Integration with manufacturing processes: FIM can be integrated with manufacturing processes, which allows manufacturers to optimise their production processes and reduce waste. This integration also allows for more accurate cost estimates and better project planning.

6. Sustainability: FIM can be used to evaluate the sustainability of a building design, including factors such as energy use, material selection, and waste reduction. This allows architects and engineers to design more sustainable buildings and reduce the environmental impact of construction projects.

The post Can FIM solve the ‘modular question’? appeared first on AEC Magazine.

It is a common notion in the AEC industry that building is a linear process. It begins with conceptual design, moves into compliance, then analysis, and finally, fabrication. That process can take months, and sometimes years, every single time.

But FIM (Fabrication Integrated Modelling) offers to reverse the equation, allowing us to get fabrication right from the start. Looking at the construction software industry roadmap in past years, it is evident that the design/engineering process is perceived as a ‘solved’ issue by the mainstream CAD/BIM companies that dominate the market. In the eyes of executives at these companies, the next stop on the journey to conquering the entire construction value chain is project management and procurement.

In the age of FIM, the one who controls the initial design/ engineering process controls the rest of the value chain, too

But I would like to propose an alternative perspective. To my mind, procurement and project management are not a ‘continuation’ of the design process, but a derivative of it. Instead of running forward, perhaps we need to go back to basics, and think about the building process more holistically. In the age of FIM, the one who controls the initial design/ engineering process controls the rest of the value chain, too.

Though BIM’s original concept was to simplify the design process and cut resources, current methodologies demand special BIM managers and experts and more planning time and costs. It is no wonder, then, that only an estimated 20% of projects use even the basic capabilities of BIM. A much smaller fraction is actually implementing detailed, high-level modelling, as seen in BIM Level 3.

In today’s construction world, space layout planning is not enough. Design teams (architects, engineers, consultants) often reside in ‘masstopia’ or ‘renderland’, their thinking dominated by virtual environments and documents nested as far away from a real-life work site as possible. As a result of this fragmentation, the level of control that design teams exert on the fabrication supply chain is diminishing.

In an age where the biggest challenge the construction industry faces is adopting advanced industrialised methods, smart materials and sustainable supply chains, why do these considerations come last, when it’s already too late to make changes?

Fabrication first

Getting started with ‘FIM thinking’ requires a mindshift at every stage of the process. Below, I outline the current situation and then the necessary shift to be made at each of those stages.

Design phase

Current situation: Architecture starts with a sketch and moves towards detailed design, regulation and building permits. Fabrication, sustainability and supply chain considerations come at a later stage.

FIM thinking: Building details and regulation compliance are integrated as part of the design process using AI. This ensures that everything we design is ready for construction from the get-go.

Supply chain and procurement

Current situation: After planning is approved and permits are given, the design and plans are ‘frozen’ in place and the search for methods, procurement and detailed fabrication drawings begins. This makes it impossible to adjust to any method or system that requires design optimisation, ruling out most industrialised methods by definition.

FIM thinking: Supply chains and bestpractice methods are an integrated part of initial design and decision-making. FIM methodology creates best-fit results using big data gathered from supply chains to help planners make the right choices from the start.

Cost analysis

Current situation: Costs are assessed as general numbers and after a design is finished, it goes to tendering. This creates many unknowns and a tendency to design for the lowest common denominator in order to reduce risk.

FIM thinking: Costs are embedded in the planning data and connected to a realworld supply chain, turning your BIM model into a shopping cart.

One offs vs. network effect

Current situation: A building is designed once and lives in a vacuum. It is a one-off in every sense of the word and lacks interaction with external supply chains, knowledge bases or other projects in its surroundings.

FIM thinking: Building plans live in a smart network. They learn from one another and share a unified supply chain. Not only do they all have access to procurement data, but they can also be optimised, to achieve economies of scale and reduce costs.

A digital revolution

Construction, the largest industry on the planet, is going through a digital revolution. If we want to see more sustainable, affordable and efficient cities, a radical approach is needed to break the current bottlenecks.

FIM offers such an approach — but it also demands that we push back the boundaries of our comfort zone in order to get there. This means letting go of old hierarchies, business models and workflows, leading the way to true human/ machine integration.

In my work with Foldstruct, these principles are being applied on a daily basis in a constant battle to expand architectural horizons with artificial intelligence (AI).

Main image: A midjourney, AI-inspired vision of the future for Fabrication Integrated Modelling (FIM)

The post The pillars of Fabrication Integrated Modelling (FIM) appeared first on AEC Magazine.

The next generation of buildings will have to be designed and built differently.

When connecting the dots, it is no wonder that 40% of CO2 is produced by the building industry alone. “Traditional” mainstream construction has proven to be unsustainable, inefficient, and unaffordable. It’s time for a change.

Towards a new industrial revolution

As Einstein allegedly said: “The definition of insanity is to do the same thing over and over again and expect different results.” Going by this example, the results we see today are inevitable.

Perhaps, these two facts tell the story better than any dystopian diagram:

1. A city the size of Paris is being built every week.
2. Only 1% of buildings actually meet sustainability standards and goals laid out by world government, according to a recent study by World Construction Forum. To make things worse, these numbers are not decreasing but exponentially rising due to labour shortage and material supply chain issues.

So, what needs to happen? Just look at the manufacturing world before and after the industrial revolution. The prices of commodity products like cars, furniture, clothing and more have decreased by x10 when compared to 100 years ago and their CO2 emissions due to manufacture have dropped due to standardisation.

Just like in all other manufacturing fields, the solution for tomorrow’s buildings is the trinity: industrialisation, scale, repetition.

But how can you industrialise a building? A building is not a shoe or a car. It has embedded attributes which require customisation.

Not all buildings are created equal. But they are darn similar

Welcome to the age of AI and “buildings as products.” Let’s start with the basics:

Why is industrialised construction sustainable? Industrialising construction consists of a few benefits vital to decarbonisation:

• Controlled manufacturing: Navigating CO2 emissions through automated EPD approved assembly lines.
• Reducing material waste through manufacturing standardisation.
• Increasing work efficiency through automated production around the clock.

Ultimately, this not only helps reduce CO2 on an individual basis, but, perhaps most importantly, creates a scalability effect able to reduce costs.

The equation is simple: produce more, produce repetitively in a controlled environment, and save CO2, cost, and time… the holy trinity.

The missing link

Simple, right? With all this said, we must remember that prefab methods have been around since the 70s and are notorious for creating repetitive socialist blocks all over the world. This, however, is rapidly changing thanks to mass customisation manufacturing abilities and, most importantly, AI-aided design tools that can help us design for machines.

Contrary to DfMA, FIM doesn’t only talk about the outcome, but, more importantly, about the process. The ability to unlock the knowledge for the industry without requiring it to change, and empowering the masses

To build smart, you need to plan smart, and that’s where AI comes into play.

Contrary to shoes or furniture, buildings cannot be 100% replicated, nor should they be.

Fabrication Integrated Modelling (FIM)

In today’s world, BIM is essentially a geometry canvas. Any data on top of that is a bonus, an expensive bonus delivered by expensive consultants (we already mentioned only 1% of buildings being sustainable and there is a reason).

Standard BIM solutions don’t provide real insight about what we should be designing and how it should be designed. This is the main challenge of my work with Foldstruct — empowering the planning process with embedded knowledge that makes all the difference with a term I call Fabrication Integrated Modelling (FIM).

Contrary to DfMA, FIM doesn’t only talk about the outcome, but, more importantly, about the process. The ability to unlock the knowledge for the industry without requiring it to change, and empowering the masses.

Sustainability=data

To reach net zero buildings, we need data! Data about materials, about life cycle assessment, and energy usage and about cost.

The data pillars of sustainability There are multiple ways to analyse a building’s sustainability. From multi physics analysis to supply chain footprint. However, all these can be nested in the following metrics.

• Embodied carbon – The amount of CO2 emitted in the production/manufacturing process.
• Operational carbon – The amount of carbon emitted during the building’s operational lifecycle, with each one having its own sub-metrics.

A building is more than the sum of its parts

Unlike common belief, there is no standardisation or one system that fits all. In fact, industrialised and modular methods include hundreds of different methods, materials systems, and manufacturers, each one having its advantages and disadvantages and demanding design optimisation.

Just to name a few: prefab 3D capsules, flat packed panelised systems, casted prefab element assemblies, dry connection systems and more. Now add to that different structural material systems such as concrete, wood or steel, and you can see how the number of variations grow exponentially.

The new age of planning

What all these do have in common, is that they all share the need for design optimisation to be integrated in projects.

With so many options and variations to choose from, it is no wonder that standard planning tools and methodologies cannot begin to provide solutions for embedding them.

Repeating success with AI

A recent study conducted on two best practices projects in London has shown a 35% CO2 reduction when compared to standard construction methods. So, can we duplicate this to the rest of the market? Yes and no.

As mentioned, every method has its strength and weakness, and is usually intended for a certain type.

For instance, 3D prefab ‘capsules’ work wonderfully when containing repetitive enclosed units such as small hotel rooms, student housing or sanitation room pods. However, they are much less efficient where custom sizes are needed, or for larger spaces that cannot be enclosed in a capsule.

Panelised systems are wonderful for large scale projects and provide an optimal solution for flat packed shipping. However, they require standardisation and large scale production.

What’s next?

This is exactly where AI comes into play. I believe we will soon see a merge of manufacturing integrated in design stages and create buildings as living products. With the great advancements in AI being created on a daily basis, it is only a matter of time before we can design optimal buildings at the click of a few buttons.

Tal Friedman is an architect and construction-tech entrepreneur active in automated algorithm-based design-to-fabrication. His work explores new possibilities for transforming the built environment through innovative use of materials and creating new typologies for architecture and structural purposes. Tal has also presented at NXT BLD.

The post From BIM to FIM (Fabrication Integrated Modelling) appeared first on AEC Magazine.

In nineteenth century philosophy, the term Zeitgeist was used by German philosophers to describe “the spirit of the age”. An invisible force dominating the characteristics of a given epoch in world history. Today, AI is promising to bring with it a new Zeitgeist that will change our stream of thought as we enter a new era. However, to unleash its power, we first must first give up some of our core beliefs.

Living in the empiric age

With the rise of computers and algorithms, we have reached an era which I would describe as “the empiric age”. To define it in short: beautiful, romantic, and sacred-out. Functional, optimised, efficient-in. It is an age where rational and functional reasoning is the prism through which we view reality and is the justifier of our decision making. But where does this meet architecture or art? Where does one draw the line between pure rationalism and desirability?

Architecture has always stood at the pinnacle of human ambitions, literally engraving mankind’s epoch achievements in stone. The Greek philosophers sought out for a catharsis – a purification and purgation of emotions through dramatic art and extravagant architecture. In the Middle Ages, the role of architecture was to empower the divine feelings of holiness. The Renaissance put humanism and the beauty of nature in the spotlight and the Futurists of the 20th century sought to worship the machine and the coming of the industrial age.

From a bird’s eye history perspective, over the years we have seen a gradual move from subjective streams of thought based on locality and culture into an objective and universal truth. This was ever so evident in architecture with the move to modernism, which blurred, not to say erased, the boundaries of locality and created a “one-truth-fits-all” approach. The death of the ornament and rise of the function.

AI to the rescue

Recent developments in AI have created a completely new standard for what we perceive as “computational” design. Not only can software provide us with a digital design canvas such as Photoshop, CAD sketching or 3D modellers, but, for the first time in history, it can design for us.

Text to image generators such as Dall-E and Midjourney are spreading like wildfire and demonstrating no less than a revolution in how we transform our thoughts into visual reality. However, they also reveal some of the weak spots of our abilities and make us rethink the “human advantage”.

Since the public release of such tools, and due to their ease of use, we are being bombarded with image creations in all fields possible – from children’s books to automotive design and architectural visions. What used to require skilled artists is now just a few letters away. So here is the real question: what is the value of these works?

Where reality stops, art begins – from “creation” to “generation”

A similar question has been posed with the invention of the camera, which pulled the rug under the feet of realist painters. This camera required artists to stop copying reality, as the camera could do a far better job than they could and forced them to give new interpretations to reality which the camera could not. In other words, the camera could only display what was there, but not “invent” reality. The same was deemed true about the computers of our age – “they can calculate, but they can’t think” was the common notion. The new age of AI and machine learning is changing all that. But just like any race car, without a responsible driver it can become a dangerous weapon.

One would expect AI tools to unleash creativity and reveal new and unimaginable layers of artistic value. However, the thing striking me is that most work being published resembles a mockup of reality which could have been produced manually. In fact, they are starting to become remarkably similar to one another, to the point where one can detect them just by their style, especially in the case of Midjourney. It is many times an uncanny and uncomfortable version of reality. This is happening for two reasons:

1.  The machine learning process learns from past examples, and therefore creates a mix and match of what has already been defined. It bases its assumptions on things which we have already perceived. Whether there is another option remains a philosophical question.

2. Innovation comes in small chunks. Can we grasp quick changes without getting used to one innovation at a time? This is especially evident in art, architecture, and fashion. We no longer need to wait for styles or trends to “sink in”. We can generate thousands of iterations at the click of a button and accelerate design evolution. Or can we?

Building a new world with AI

Construction is set to become one of the greatest benefiters from the AI revolution. Not only will we be able to release ourselves from the technicalities of drafting and long regulation process, but we will also be able to create better architecture. How? By simply sitting back and letting AI do the work.

If we truly grasp the notion of machine learning, we will let algorithms select not only the “what” but also the “how”. What do I mean by that? In today’s perception, it is us that define the questions and tasks and AI that shoots out answers. As an example, a common AI manoeuvre is “style learning” – learning a style and replicating it in different variations. A Zaha Hadid cupcake, A Frank Gehry wedding dress or, perhaps, a dog resembling a Le Corbusier building (don’t try that at home)? We create the visions and let AI interpret the rest. But if we let AI itself define its goals and only request “amazing architecture” on a plot of land? What if we give up complete authorship and just request things that are determined as “awesome”?

Teaching AI to recognise what makes us tick and raises emotions in us is, perhaps, the next step in human-machine interaction. Removing the A from AI means moving to an age of natural intelligence and giving up control of basic notions of what it means to create. It means exposing ourselves to machine learning and letting algorithms ‘learn us’ rather than learn objects or styles. By doing this, we will be able to discover the undiscoverable. We must learn to understand machines so they can understand us.

Tal Friedman is an architect and construction-tech entrepreneur active in automated algorithm-based design-to-fabrication.

The post Removing the ‘A’ from AI appeared first on AEC Magazine.

Despite the current “green hype” from regulators and industry, 99% of buildings are still being built to the lowest common sustainability denominators.

Einstein allegedly said that the definition of insanity is to do the same thing over and over again and expect different results, and the same is true in construction.

If we keep planning the same buildings, we will end up with the same results.

Sustainability cannot happen on a local basis. A “green” house, school or even neighbourhood which do not prove a cost effective and scalable model are many times a part of the problem, also known as “greenwash”.

In order to effect true change, we must see the industry as a whole and provide models that can work for the mainstream instead of one-offs.

The architect does not stand alone. Gone are the days where architectural design can be dispatched from its means of manufacturing and footprint analysis. New technologies now allow us, like never before, to create unified methodologies that blur the boundaries between disciplines. But to grasp the potential, we first have to understand the problem.

So how big is this problem?

You’ve all heard this one before: the construction industry is responsible for 40% of global CO2 emissions, making it the most polluting industry on earth.

In fact, reducing emissions for this sector alone can achieve the global benchmarks for preventing global climate change.

Add to that a geo-political turmoil caused by external energy source dependency, and you have the most burning issue of this century. The coin of the future is, therefore, not the Dollar or the Bitcoin, but the Kilowatt.

Bridging the gap

As demonstrated in my previous articles, the planning world has not changed fundamentally for thousands of years, relying mainly on pure geometric representations, very far from their actual manufacturing details and eco footprint. Yet, it is clear that industrial manufacturing holds the key to improving performance, reducing waste and using smart materials – the main factors affecting a building’s sustainability.

The biggest challenge is in matching the restraints of these in early design stages.

Today, “real world” data leading to CO2 calculations can only be extracted at the last 20% of the planning stage, after planning is pretty much complete, making it practically too late to change and optimise.

So how come in a world of BIM, digitation, and open knowledge, we are so far behind?

Compare today‘s AEC planning firms with those of 50 years ago, and you see something remarkable. Planning costs, times and complexity have gone up due to extra BIM experts, consultants – and rising software license fees (more on that in Martyn Day’s article “Prisoner of Vendor), yet the overall detail level has not seen significant changes for the mainstream. In other words, we have digitized the same problem.

The solution: optimise early on

Design for manufacturing (DfMA), the holy grail for the construction world, is often overlooked by AEC firms due to complexities. However, it is now becoming more feasible than ever.

According to the latest reports by bodies like McKinsey, Deloitte etc. adopting industrialisation and automation for the construction world can not only reduce CO2, but also cut costs by 15-25% when applied at scale. Yet, this requires all stakeholders to “play by the rules” in all stages.

Looking at other industries like furniture, automotive and aerospace that have managed to revolutionise their manufacturing, we can see the large role that designing to the screw level holds. The equation is simple: the more data you have, the better you can optimise.

It is, however, unrealistic to expect an architect or engineer to know how to write machine code, analyse CO2 footprints and be acquainted with all the latest building technologies. So how do we bring all this together?

A new state of mind

Using AI, a radical mindset shift is now available. No longer should architecture remain isolated as a ‘soft’ discipline, ending its role upon design completion. New architecture is deeply rooted in its environmental and social impact.

Planners must take full responsibility of what they design above its local usage.

On the other hand, manufacturers, contractor and regulators must provide the infrastructure and data for its adaptations.

The amazing possibilities of data analysis for AEC that can be achieved through digitation are immense and growing by the day. Just to name a few: daylight analysis, CO2 calculation for materials, thermal insulation with Finite Element Analysis (FEA), Computational Fluid Dynamics (CFD)-based wind calculations, Life Cycle Analysis (LCA), manufacturing supply chain optimisation – all these, when combined together, can help us create greener, smarter and more cost effective buildings.

From my personal experience with Foldstruct, working with corporates in the manufacturing field, it is all about data collaboration. A true open book approach from all sides can digitally dissect the project in ways that are not seen to the naked eye.

In a recent project, we were able to not only reduce CO2, but also cost, by optimising the design according to a specific system. The main challenge was to reverse engineer something that has already been designed. However, using AI and parametric logic, it was possible to make slight adjustments that were non-intrusive to the original design, yet proved to be of great benefit in terms of performance. It is indeed an ongoing journey to digitize the construction industry, but the path has been set with more and more corporates joining the game.

As more analysis tools spring up and regulations demand more sustainable buildings, we will see a unification of all disciplines in a “building as product” approach. It is a time where AI, robotics and smart materials are no longer buzz words, but the future of the AEC industry.

About the author

Tal Friedman is an architect and construction-tech entrepreneur active in automated algorithm-based design-to-fabrication. His work explores new possibilities for transforming the built environment through innovative use of materials and creating new typologies for architecture and structural purposes. Tal has also presented at NXT BLD.

The post Reducing CO2 by AI design and robotics appeared first on AEC Magazine.

Digital optimisation promises to revolutionise the world of con struction via ‘Industry 4.0’ mechanisms. But current BIM solutions, being pure geometry platforms, have little to contribute in terms of solving today’s problems. In fact, they may actually be part of the problem.

A controversial statement, maybe – but let’s look at the impact of optimisation and automation on another sector: manufacturing. The Industrial Revolution and the invention of the assembly line triggered a giant leap in manufacturing productivity, reducing the relative cost of a car by a 10x multiple over the course of just over a century.

Yet, over the same time period, construction costs have consistently risen, because on-site automation is nowhere to be seen. In the planning world, the focus has been on optimising the production of documents, rather than the performance metrics of the final building. In other words, we have industrialised planners, not buildings.

Focus on the right problems

So what does this mean for the promise of BIM? Is BIM even focused on the right problems?

These questions remind me of a story I heard about a large toothpaste company, looking to increase its market share and assembling a crack team of scientists, designers and marketers to come up with the world’s best toothpaste. After a lengthy R&D process, their decision was unanimous: increase the nozzle diameter on a tube of toothpaste by 2mm.

It seems that the AEC world is not so different from the world of dental product R&D. Despite rising software prices and the move to subscription-based software-as-a-service platforms, not much has really changed over the years since the introduction of BIM in the early days of Revit.

Little of the innovation we see actively seeks to address the sector’s inherent problems. Software suppliers scramble to make the move to the cloud and M&A activity flourishes in the sector, with deals springing up like mushrooms after the rain – but that’s because the holy grail for these technology vendors is to control the complete supply chain via unified platforms. The big question for customers, however, is this: What are we getting in return?

That’s not to say, of course, that BIM has created no value at all. On the contrary, it has supported some incredible projects, delivered with very high levels of detailing – but these typically involve extremely high budgets and designated BIM experts, and by no means reflect the average project.

In fact, despite BIM’s promise to support more design freedom and bring down planning times, it still takes an average of two to three years to design a multifamily project, regardless of the planning method involved. This may explain why only one in five offices have adopted BIM to its full extent. At the majority of firms, employees swap between four or five platforms — SketchUp, AutoCAD, Rhino, 3ds Max, Solidworks, Revit and so on — with each one providing just one small piece of the puzzle.

Five changes needed

So what needs to happen for BIM to truly deliver value and more closely meet the sector’s needs?

First, we need real-world data integration. As mentioned previously, BIM architectural models are still pure geometry, detached from real-world data. This means a design will undergo endless iteration loops, each involving different stakeholders and consultants who add their input and enforce changes.

Second, we need the integration of manufacturing data. If we are unable to estimate costs in the planning stages and understand the implications of design changes, we are designing for the lowest denominator. Simply put, we cannot speak of robotic automation and design for bricklayers.

Third, we need a lower barrier to entry for BIM. Due to its complexity, BIM managers have become the norm, meaning increased workforce numbers and costs. This often involves asking the client to increase the planning budget, too.

Fourth comes increased flexibility. Designing in template-based environments leads to high rigidity and an inability to customise without deep technical manoeuvres. Software must support more customisation while remaining in the scope of feasibility.

Fifth and finally, an easier learning curve is a must-have. It is said that for an average office to make a complete BIM transformation takes two to three years, during which the team will undergo a painful period of relearning. This risks losing control of projects and experiencing a degree of trial and error that few offices can afford.

The missing link here? Design with purpose. As buildings become smarter and more technological, it is clear they will also need smarter planning tools. The race to arms is on!

The software of tomorrow will be much more than ‘just’ software. It will be deeply nested within the value chain of a project. For this reason, the AEC field is hotter than ever, characterised by a race to dominate the sector in a 360-degree approach.

From the ground up

But the solution to the problem — as always — comes from the ground up. With mounting international efforts and pressure to reduce both CO2 emissions and building costs, it is no longer an option to ‘choose not to choose’. Developers and contractors have to show empirical evidence of gains in order to get ahead of the game. Projects that fail to comply with these standards will simply be thrown off the wagon.

The age of artificial intelligence, or AI, promises to put the intelligence into BIM, adding valuable information to models that will help optimise buildings and shorten design loops. For this to happen, the boundaries must blur between designer, builder and regulator, with all three groups working from a unified data hub. Writing about these topics naturally raises many questions on actual implementation plans and next steps. I cannot, of course, speak for the whole industry, but in my work with Foldstruct, we are working to implement those principles and use AI in a unified platform that calls all stakeholders to get involved, regardless of format or pedigree.

Technologies that bring value should take the lead. Those that don’t will simply have to try harder, regardless of market share.

It is time to bring the power back to planners and free them to spend more time designing and less time drafting. The age of closed circles and formats is over. Welcome to the age of optimisation!

About the author

Tal Friedman is an architect and construction-tech entrepreneur active in automated algorithm-based design-to-fabrication. His work explores new possibilities for transforming the built environment through innovative use of materials and creating new typologies for architecture and structural purposes. Tal has also presented at NXT BLD.

Read more of his articles here

The post Power to the planners (and why BIM doesn’t really exist) appeared first on AEC Magazine.

Advanced AI modelling techniques, together with robotic fabrication, promise to create the much-needed revolution in what we design and build. But these steps require us to disrupt the nature of architectural practice. So, what really needs to happen for the industry to succeed and are we really ready for such a change?

Einstein once said the definition of insanity is to do thre same thing over and over again and expect different results. Yet here we are, designing the same buildings, over and over again and expecting to decarbonise, save cost/time and improve architectural design — surprisingly, to no avail.

Responsible for 40% of CO₂ emissions and growing socio demographic problems as results of living standard gaps, it is time to admit modern architecture has failed us.

It has failed to use technology to guide it to its goals due to fear of disrupting its own centralised power structure. In fact, the most common question raised when speaking about design automation is the concern for the well being of the architect rather than the well being of society and the built world.

However, new AI technologies at hand can empower architecture to get back on track and democratise construction. It’s time to go back to the drawing board and replan planning as we know it.

The rise and fall of the architect

The architect, who has taken the historic role as guardian of architecture, has let its guard down and allowed a new player to dominate the realm, replacing the soft and often vague term of “architecture” with the hard and empiric term of the “construction industry.’

This is not new, of course. The architect has changed its role many times throughout history: from an on site omni present design-builder, to a “behind the desk” paper draftsperson and all the way to a computational specialist living in a virtual universe. What has always remained at the heart of the discipline is the desire to create a worthy living environment.

However, the shift from a physical on-site methodology to a world of data and theoretical geometry has disconnected the discipline from the ground to the point where the modern architect is merely a nester/drafter of repetitive shelf products that compose buildings.

As a matter of fact, it is fabrication restraints dictated by producers and contractors that limit 99% of the buildings to repetitive boxes from early design stages. For the average architect, it is a given that any deviation from the norm will create an exponential price increase due to the added engineering and customised production. ‘Starchitecture’, on the other hand, is a game reserved only for those privileged enough to have unlimited budgets.

In short, we have become strong on data, clouds and documentation, but weak on design flexibility and well being — aka architecture.

So yes, we have managed to build mega cities higher, quicker, and more industrialised. Yet, the industry fails to meet its own success criteria, deepening the problem year after year. Just like in the story of the tower of Babylon, our success is exactly what is leading us to our failure. So why should we be optimistic?

Build like a robot, sing like a human

Automation tools and robots have created the fourth industrial revolution in fields like automotive and aerospace and are now more feasible than ever before for mainstream construction. It is estimated that since moving to the age of advanced automation, the price of a car has relatively dropped by about x3, yet construction is getting more and more expensive due to growing labour costs. Robotic tools like 6-axis robotic arms from KUKA or life-like robots from companies like Boston Dynamics and Festo have led to a new age of mass-customisation.

But it wasn’t just the robots themselves that changed the game. More importantly, it was the ability to control them using digital DfMA (Design for Manufacturing and Assembly) methodologies. Now, these powerful tools are at the fingertips of the construction sector, we no longer have to think in terms of moulds and stamps. Instead, we can begin to design in free forms and new morphologies.

So how do you teach your architect or engineer next door to design for a machine, let alone a fabricator who is using traditional tools? After all, it took us thousands of years to come up with methods that match the human labour restraints. The answer: a singularity in the design process that connects all stakeholders to the same model, from initial design, all the way to fabrication and takes into account time and cost of building from preliminary stages. We now have the means to do so, if we only dare to step outside the box.

Make architecture great again

The magic moment in architecture is the awe of seeing a design come to life and being surprised by the outcome you designed yourself. Undiscovered spaces form hidden gems. Reflections and lights tell more than meets the eye.

With AI, we can achieve new levels of self-forming designs that will have to match our design goals as a principle, but have a free hand on how to get there. Technologies like GAN (generative adversarial network) and ML (machine learning) have the power to create unimaginable forms and functions if we only dare to open that door.

The building of tomorrow will have a life of its own and be based on an adaptable DNA, varying on needs and project goals, rather than design ownership of a single entity. The architect in that sense will have to give up a lot of the authorities it had as an individual designer, but in return gain access to a new world of opportunities. The boundaries must blur between an architect, engineer and manufacturer, and turn the building into a living product. We must guide, select and choose, rather than dictate. In the work of Foldstruct, the aim is to democratise the planning process and mix between endless data combinations in a platform where architecture is a metric and not a title.

So what will the future of architecture hold in an age of automation? Will it kill architecture as we know it, or return it to its old glory with a new twist?

The post The death of architecture (and the rise of a new one) appeared first on AEC Magazine.

In the AEC world, the digital age has brought with it a complete shift in how we design. In just 30 years we’ve gone from paper to screen, from pencil to mouse, from 2D to 3D and from local to cloud. In fact, it is fair to say that AEC has become a branch of IT and that the planner has become a user.

BIM, parametric design and automation allow planners to sketch out their dream projects and even generate new ones using Artificial Intelligence (AI) at the click of a button. Yes – the digital world tolerates anything we can imagine and it’s literally at our fingertips.

So how does all that look down here on Earth? Has digitising the AEC industry lived up to its promise and created the desired effect in our built environment? Unfortunately, today’s mainstream construction is a far cry from reaching goals like zero energy, architectural well-being and affordability. Just ask governments and regulators worldwide, who continue to pour billions of dollars into funding programs to help bridge these gaps every year.

Looking at the urban landscape, aside from seemingly budget-less ‘starchitecture’ springing up from time to time on magazine covers, in the last 50 years 99% of buildings have barely seen any change in how they are actually built. Indeed, the planning is all digital, but the product has remained all the same. So, when we speak about building faster, massively and globally, we must first stop and ask: are we so happy with the buildings we are seeing around us that we want to make more and more of them? It’s not enough to talk about IFC standards and cloud collaboration.

With millions of new buildings and renovations worldwide expected in coming years, it’s time to talk about the real potential of digital construction – the one that will actually change how and what we build!

Robots that think like humans instead of humans that think like robots.

It is commonly agreed that robotics and industrial automation hold the key to achieving the goals of the construction industry – those being: time and cost reduction, sustainability and architectural design. It all starts with a mind shift into a new era of designing, not just for humans, but also for the robots that build for them.

For this to happen, automation must be embedded within the planning process from initial stages and allow us to design especially for robotic fabrication. Rather than thinking from the total design and “componentising” shelf-products later on, we must take a reverse approach and design from the inside out and the screw level. Understanding the possibilities and restraints of robotic manufacturing will allow a new era in BIM which goes far beyond geometry documentation.

Looking at the automotive world, robotic industrialisation has brought down the price of a car to one tenth of its relative price back in the 1930s. All of this while improving safety, performance and becoming affordable for all. This was done not only by creating automated assembly lines but, most importantly, by optimising and linking CAD files with the means of manufacturing and machine code. Robotic companies like Kuka and others have reached precision levels and speed that allow drastically lowering costs and improving productivity.

In construction, however, relative building costs are on a constant rise while productivity goes down as it gets harder and harder to find skilled on-site workers in Western countries. Can we adapt those same mechanisms that worked for the automotive world for construction? Yes and no!

As opposed to the automotive world where a car is manufactured tens or even hundreds of thousands of times, the building world is all about customisation and adaptation. In fact, no two buildings are completely alike.

From SaaS to DaaS

Maybe we need to think differently and consider moving from Software as a Service to Design as a Service? Despite the big promise of BIM, today’s AEC world is suffering from an over complication of software. Often requiring a BIM specialist just to run, it is no surprise that only around 20% of firms harness full usage of BIM at all scales while the rest are picking up the bits with ‘hybrid’ solutions.

According to internal studies at Foldstruct, there are on average 15 different software tools that planning teams will use throughout the project lifecycle. With very limited interoperability and hundreds of design loops, it is not hard to see why it so difficult to build ‘outside the box’.

We have identified an exponential rise in costs in almost all sectors of AEC due to this ‘domino effect’. But perhaps the most troubling thing about the current workflow is that it does not help us talk to the factory floor in design stages and leaves way too many blind spots and unknowns. It is a mere graphic representation, not very different in essence from ink on paper.

3D mass modelling, BIM, structural analysis, shop drawings, electricity, HVAC, rendering, graphics, tendering and many more topics are simultaneously addressed, each requiring their own specialists for operation.

Construction 4.0

The days of software as a ‘blank page’ platform are soon to be gone. The fact is that over 5 million architects and engineers worldwide are looking for tools to help them design and build better buildings.

They are not impressed with vaulted cloud packages, smart subscription models and closed APIs. To put it simply, they are looking for the shortest path to get from an initial concept to an optimised building. We have to break free of the old supply chains and think of the leanest possible process to see the building as a unified product.

There are many ways of doing this and they all require an open-ended approach and multidisciplinary collaboration. At Foldstruct, we try to combine design, performance optimisation and digital fabrication in a one-stop shop. Using smart algorithms to capture the DNA of the fabrication system and building the designated tools with which designers can plan without them having to go into detail.

Tomorrow’s BIM is not a blank page but rather an adaptable, dynamic template. It requires opening the platform and inviting any software, hardware or service along that can add value to the process.

I believe that the boundaries will quickly blur between software producers, system suppliers and fabricators and that business models in AEC will become value-based and not subscription-based.

I envisage a world of expert systems which create optimised, fabrication-ready facades or AI which ‘lightweights’ buildings to make savings on foundation work, perhaps a system which can deliver a design in different materials making structural modifications with each option. If we can capture and codify production knowledge, these services could be used within the design of a building or in post rationalisation.

Given that the most expensive material in construction is knowledge, we are living in exciting times where that knowledge can be replicated by algorithms to spread across industries at the click of a button. It is now in our hands as an industry to collaborate and bring construction to the next level.

About the author 

Tal Friedman is an architect and construction-tech entrepreneur active in automated algorithm-based design-to-fabrication.

His work explores new possibilities for transforming the built environment through innovative use of materials and creating new typologies for architecture and structural purposes.

His company Foldstruct uses advanced algorithms and robotics to transform the way things are designed and made, giving new life to industry standard materials and leading the construction tech revolution.

The post Towards a parametric future appeared first on AEC Magazine.