AI: The catalyst for systemic transformation

So far since the start of the Industrial Revolution, the core economic incentive when making physical goods and products has been intrinsically built on the maximization of the number of units sold to customers, as well as the minimization (optimization, alternatively) of their lifespans.

In other words; more products equal higher economic benefit for manufacturers, their employees, and entire countries. Short product lifespans encourage customers to make frequent repeat purchases. The economic incentive is based on units and linearity.

This article explores a paradigm shift towards an economic model based on access and circularity, enabled by the convergence of disruptive innovations, particularly AI. Through the combination of autonomous vehicles, humanoid robots, remanufacturing, and renewable materials – AI could become the catalyst for systemic transformation.

 Key takeaways from article:

• This article is written for SEB's latest Green Bond Report, highlighting the prospective shift from a linear economic model to a circular model, driven by AI and other disruptive innovations.

• Success for digital access-based models like SaaS indicates potential for similar physical world models, focusing on usership.

• AI and smart sensors can transform products into service-providing robots, promoting longer lifespans and high-quality goods.

• Humanoid robots could enable localized manufacturing and address demographic changes.

• AI-driven material discovery and autonomous mobility can lower costs and enable the sharing economy.

• The article is based on the report Regenerate the Economic Machine (link).

Since the turn of the Millennium, we’ve seen previously unquestionable physical products convert to being fully digital. Information morphed into ones and zeros. Newspapers, CDs, and manual work turned into software streamed from anywhere as information could be accessed and shared from anywhere. The internet and cloud services opened the capability of selling these informational products as something customers had access to from anywhere, at any given time. The access-based business model proliferated in the digital world when ones and zeros reached a state of near-zero marginal cost and became “renewable”.

The Software-as-a-Service (SaaS) model now dominates most informational sectors worldwide. The largest digital tech companies on global stock exchanges have emerged or evolved as providers of access to their underlying products. They adopted this model because customers prefer access-based services over purchasing units or licenses.

Customers like to “press Play” and eliminate problems. And furthermore, the access-based business model raised recurring revenues and customer loyalty quite substantially – consequently raising market caps across the digital board. For providers, this model resulted in higher earnings, whether through profits or increased valuations.

In essence, the core digital manufacturing incentive changed, and then the global market changed.

Play button represents the usership model. Source: Christopher Lyrhem, SEB, Regenerate the Economic Machine.

THE LINEAR GAME REIGNS SUPREME

In the physical realm, the access-based business model is still in a minuscule state. We make units of products for customers to buy and consume. The reason for this is quite simple. Material input into physical activity is inherently not easily accessible from anywhere, not close to being near-zero marginal cost, and certainly not renewable. Products are made to be purchased and consumed by their owners.

Globally, this linear unit-economics model is the strategy that management and boards strive to perfect. It’s a game of units, not a game of service (or access). And in this game, both manufacturers and their retailers need to make products at the location where it’s most cost-efficient, and then the products need to be distributed swiftly to customers all over the world.

Offshoring has exploded since the start of the 1990s, and offshoring means that the feasibility to refurbish and remake physical products has diminished. Supply chains have gotten much longer in distance and in turn also the ability to standardize parts that go into products so that they are easily replaceable and remade.

My main point is that it is currently (largely) economically irrational not to engage in the linear unit-economics model, from most perspectives.

Products must be affordable, or manufacturers risk losing market share. This necessitates factories in low-wage countries. This systemic reality makes the global economy rigid and resistant to change. The linear model is inherently rigid. And in order to change this rigidness, there needs to be something economically better than the unit-economics model.

The incentive for something new needs to arise for global manufacturing to transform.

THE COLLECTIVE POWER OF DISRUPTIVE INNOVATIONS

All throughout history we’ve seen disruptive innovations transform marketplaces because they offered the “job to be done” (the customer value) in fundamentally novel ways. For example, the car was a completely different and superior alternative to the horse. The telegraph outperformed physical message delivery. Computers were better than manual systems. And the list goes on. The list of something new replacing the old is longer than anyone had imagined when the Industrial Revolution took off some 300 years ago.

And the list will continue to flourish as the disruptive innovations of today are by far the most potent than any time throughout history, while also being able to converge into each other. The first, second, and third Industrial Revolutions formed into distinctive iterations of what gives their names – through the convergence of disruptive innovations. When new innovations come together, they support an aggregate change on a global scale.

But in spite of the constant increase of the spirit of human innovation – the linear manufacturing model has not changed.

This persistence is due to recent disruptive innovations making the linear model more efficient. For example, e-commerce greatly accelerated offshoring. Efficiency has made the linear model more rigid.

Enter Artificial Intelligence, the most impactful technology offering a next era of human evolution.

As a general-purpose technology, AI could enable a circular economy and fundamentally rewire global manufacturing in the coming decades. If this turns out to be true, the impacts on most sectors will be profound and is hence vital to explore. This prompted me to write the report "Regenerate the Economic Machine", outlining future scenarios for global transformation. I encourage your to read it.

Historically, when aggregate change proliferates – the change is not sprung from a vacuum, instead it is the result of several disruptive innovations building and enabling it. Things always build on top of each other.

Currently, we are witnessing the start of a domino effect, beginning with cost-efficient smart sensors widely installed in physical products, paired with AI capabilities - which together enables the proliferation of usership business models.

Smart sensors can now track physical activity at an order of magnitudes higher level than only 10 years back. This capability has led more companies to experiment with the access-based "Product-as-a-Service" model. While some companies have long adopted this model, the vast majority have not. This model can be placed into the umbrella term “usership” (instead of ownership of physical products).

Check out the chart below, for my thesis of how the coming 10-20 years could see a domino effect , ultimately leading to a peak and reversal of carbon emissions on a global basis.

Source: Christopher Lyrhem, SEB, Regenerate the Economic Machine

ALL PRODUCTS TO BECOME ROBOTS

And furthermore, only over the past couple of years, we’ve seen an exponential rise of AI’s ability to execute human-like behavior, and even significantly surpass it. AI can now generate text, images, video, audio, and even code. The pace of improvement is unlike anything we’ve ever seen and in the coming years we will witness AI to take on the human senses so that it is able to not only execute tasks within the digital realm, but also out in the physical.

When paired with smart sensors, AI can transform any product into a utility-generating robot. Products will perform any tasks we tell it to perform. Or in other words – we’re starting to see the Play-button arrive in the physical world. Pressing Play on products to paint our walls, clean our houses, weld components together, and even drive us where we want to go.

When products gain the ability to perform tasks autonomously, it unleashes the ability to sell the product as a service, instead of only selling it as a unit. From ownership of units, to usership of access. We pay for the utility to be performed, not the unit itself. And when manufacturing companies start selling physical products as a service (on a proliferated global scale), managements and boards across sectors, countries, and continents – could (prospectively) raise the question of making the lifespans of the products longer.

Source: Christopher Lyrhem (Regenerate the Economic Machine), SEB

Because when you place a product into a marketplace and it earns revenue multiple times, not only one time – you want that product to last as long as possible. The economic rationale could hence change from minimized (or at least low) durability over time, to maximized durability.

And such outcome starts with designing a product for durability a long time before it’s made – which in turn will require an increase in quality of material input.

When a product is subscribed to by its’ user – the maker wants that product to last as long as possible. Let’s take the car as an example. As cars become autonomous in the next few years (as indicated by Tesla, Amazon, and Alphabet), customers will be able to summon them with a click of a button. No matter where they are, it will arrive at your feet and deliver you securely to wherever you want to go.

This will by most likelihood diminish the prevalence of personal car ownership.

This will by most likelihood diminish the prevalence of personal car ownership. We don’t need to own a car unless we want to. Some of us will instead subscribe to autonomous mobility services and share cars with other people. For the manufacturer of such autonomous cars, the economic rational will be to produce the cars with high durability – and in long-term even design it for remanufacturing (remaking it to its’ birth state).

This could be the incentive change we need in order to curb global carbon emissions.

Because people change when incentives change. People rarely change due to new directives and regulation change. And change is most often sprung from inventions.

Now, let's say the usership business model actually proliferates in society, wouldn't that require a new way to measure commerce? The answer is most likely a distinct yes.

Below is an illustration of how usership could become the third dimension in societal commerce, adding to 1) physical sales, and 2) e-commerce.

The chart is not a projection of how the future might fare, it's more of a provocative visual how the composition of commerce could change if usership really takes off.

Maybe we'll reach an equilibrium where physical sales, e-commerce, and usership are three equal parts?

Source: Christopher Lyrhem, SEB, Regenerate the Economic Machine.

HUMANOID ROBOTS: REDEFINING THE WORKFORCE

Furthermore and simultaneously, if we draw a theoretical timeline into the coming decades, the impacts AI could have on manufacturing and consumption is not only circled around the business model, but also in terms of how we actually make products.

Just like all products will be able to become robots to various capacities, the equipment and machines within factories that make these products – will also become dexterous beings able to take on a lot more tasks than today.

So far in time, we’ve seen a massive increase in industrial robots that are programmable to certain tasks. We have programmed them to perform specific tasks. But they are bulky mechanical animals that cannot evolve.

But there is a new breed coming, in the form of Humanoid robots, which are dexterous and can be upgraded so that they understand what to do even if the problem is novel for them.

Humanoid robots have seen a massive increase in investments over the past few years and the number of contending manufacturers is now well above 30 – predominantly from the US and China.

Source: Figure AI

This application segment of AI is still a quite underappreciated segment, because it’s not only about the technological progress itself – it’s also about macroeconomics and demographics.

Firstly, when Humanoid robots are equally as good as humans on low-skilled manufacturing jobs, the cost advantage low-wage regions hold over high-wage regions could diminish. Perhaps even eliminated. Because Humanoid robots will be able to do a quite high share of what human labor can do, and will be able to do it without sleeping, breaks, vacation, and even insurance – to a much lower cost. This could hence improve the economic rational of having a factory close to the point of consumption. Close to the customers that use the products. This is far from the way it is today.

Secondly, declining birth rates and an aging population in both developed and developing countries can result in fewer manufacturing workers per unit of economic output. Prospectively at least. This creates a need to use AI and robotics to maintain productivity and economic growth in a competitive global economy. Thus, the integration of robots into the workforce is not just about technological progress but also a response to demographic challenges.

All products in society can become robots, the workers in our factories that make these products – will become dexterous and highly capable of taking on advanced tasks.

All products in society can become robots, the workers in our factories that make these products – will become dexterous and highly capable of taking on advanced tasks.

But there is one major piece left in order to create a global society where manufacturing is local, and remaking has become the name of the game. Specifically, the development of materials that can be perpetually renewed and reused.

ARTIFICIAL MATERIAL DISCOVERY

Since Google DeepMind released their findings of over almost 400,000 potentially novel materials, discovered by their AI-models, the theoretical endgame scenario for novel materials replacing traditional materials – has shaped up as a potentially transformative factor in increasing the current 7% global circularity level. AI could revolutionize the core structure of the global manufacturing footprint. Discovering cost-efficient, renewable materials would reduce the need to purchase virgin materials from specific regions.

Historically, we’ve always replaced traditionally used materials with novel materials. Plastics, advanced metal alloys, and cement, are each an example of this. Renewable materials that can be continually remade would greatly enhance the economic feasibility of local manufacturing.

Companies that sell physical products would have the ability to service their customers through their balance sheet (or third-parties), where the material is placed and serviced onto customers through the balance sheet – as opposed to the current model where depletive materials flows through both the balance sheets and income statements. This could be a fundamentally new way to structure economic activity and financial models.

AUTONOMY ON WHEELS – THE DRIVER FOR THE SHARING ECONOMY

So far in this brief summary of how AI could change the game of manufacturing from the core – I’ve outlined three distinct theses; 1) AI could enable the usership business model and change the economic incentive, 2) Humanoid robots could enable manufacturing close to the point of consumption, and 3) AI could discover renewable materials.

But there is also a fourth impact of AI looming on the horizon. This impact is distinct from the previous three and may be more relevant in the next 3-5 years, compared to the 5-10 year time frame for the others. Namely, autonomous mobility as a catalyst factor for the sharing economy (usership in other words) Because, when autonomous cars proliferate in society, they will be able to move products with very low marginal-cost, as well robots and materials. Seamless movement of physical objects in an interconnected system could scale the first three impacts.

Let’s take a tangible example. Today the sharing economy is a minuscule part of economic activity. We have a last mile-delivery infrastructure with the word delivery as a central component. We deliver products whereafter we let go of the responsibility of enabling a sharing economy, as well as refurbishing and remaking.

Source: Christopher Lyrhem (Regenerate the Economic Machine), SEB

But once autonomous vehicles arrive, the 80% share of the logistical cost that is human driving – can be diminished substantially. And the availability of logistics could increase exponentially. For example, whenever an owner of a product (private or corporate) wants to re-sell or rent out a product, she will be able to do so by the click of a button whereby and autonomous vehicle arrive and takes it to the next buyer or renter.

Autonomous mobility could create a next mile-delivery infrastructure.

Not only the last mile – but the next.

Such infrastructure could serve as e a general-purpose platform for any manufacturer or reseller of physical products. This enables them to sell products and earn a profit when the customer resells it – or rent them out. Currently, this poses a significant challenge for companies adopting a Product-as-a-Service model.

Because when customers unsubscribe (stops paying for usership), it is quite the hassle to move the product at hand to another customer. Especially making profits from it. Autonomous vehicles can hence become a proponent for the sharing economy – which helps change the core economic incentive, from volumes to access.

Furthermore, in our SEB Subscription Survey 2023 (from Regenerate the Economic Machine), we found that the prospect of the car becoming a pioneering product category for usership – is very high. In the chart below, we have compiled the binary yes-and-no responses on the question: “would you consider subscribing to the following products”?

The data clearly concludes that products that solve a problem for us are filled with high share of electronics to a relatively high-ticket price – are more prone to partly convert from ownership to usership. Meanwhile, products that design our lives (bodies and homes) are not prone to convert.

Source: SEB Subscription Survey 2023 (from Regenerate the Economic Machine)

FINAL THOUGHTS – EXPECT THE UNEXPECTED

In economic terms, the physical world is an order of magnitude bigger than the digital world (5 times to be exact). AI's application value will likely reflect the current global GDP composition (85% physical, 15% digital), making AI in the physical world highly relevant.

Moreover, AI's ultimate impact in the physical world will be to disrupt the linear manufacturing model, replacing it with a circular model, rather than merely improving efficiency.

Efficiency innovations has so far, and will most likely continue to, not curb global carbon equivalents emissions, especially as the global middle class is expected to grow another billion in the current decade (from 4 today to above 5 billion) and prospectively upwards of 3 billion more people by the year 2050 (to almost 7 billion) if current trajectory prevails. Meanwhile, the global population is expected to increase from 8 today to 10 billion during this time period.

Therefore, disruptive innovations will likely be more effective than perfecting the current economic architecture and model. Or in other words: efficacy (the ability to produce a desired or intended result) is more important then efficiency.

The avoidance of consumption is critical and the usership economy is a vital factor (especially the sharing subsegment) – and AI is a key to unleash the incentive for this.

My main argument in this article is that AI is essential for disrupting the current linear and wasteful economic model, a view that differs from many economists.

This conclusion is based on extensive research on disruptive innovations, customer surveys showing a preference for usership, manufacturers' interest in usership models, and the urgent need for a global circular economy.

And for this to proliferate, we need a business model that is more lucrative than the current model. It’s all about creating the right incentives.

Companies need an economic upside in order to forcefully transform their factories, supply chains, logistical infrastructure, sales organization, and the way they create customer value. AI is a core proponent for such transformation.

In the next 10 years, the physical and digital worlds will be able to fuse together when the pairing of Smart sensors and AI proliferate in society. Physical products will autonomously solve customer problems more effectively, adopting successful business models from the digital realm. The streaming services of prior physical products will enter the bigger physical realm and make streaming a prominent choice of business models.

We all desire the simplicity of pressing a Play button to eliminate problems and focus on what truly matters.

And when adding Humanoid robots, autonomous factories, and novel renewable materials to the mix – the outlook for reshoring and a shift toward a globally scaled circular economic incentive, is more than intriguing.

Expect AI to lead to the unexpected in the manufacturing world...

Written by Christopher Lyrhem,

This article is included in SEB's latest Green Bond Report (link), and is based on the report Regenerate the Economic Machine (link).