The Car is Dead, Long Live the Car

I was fortunate enough to live two years in Florence and wandered daily around the birthplace of the Renaissance. Leonardo da Vinci is often called the Renaissance man because he embodied so many of the core characteristics of that time when humanity found itself again after the Middle Ages. The world, and especially those of us interested in innovation and creativity, are still fascinated by all his early conceptualizations of inventions.

Many of these inventions were drawn in large notebooks, and today they're scattered around the world. Codex Atlanticus is one such notebook, located in Biblioteca Ambrosiana in Milan. One of the drawings in this notebook is a self-propelled cart from around 1478, often called "Leonardo's car." This mechanical construction, driven by springs, can be seen as an early theoretical vision of autonomous transport. It wasn't practical, but it represented an idea about machines that move without human intervention.

Both H.G. Wells and Asimov described futuristic worlds with advanced means of transport, including automated vehicles that move without direct human control. General Motors also showed its visionary side during an exhibition in New York called Futurama in 1939. The car company presented a vision of the future (1960) where cars drove on automated highways with radio-controlled systems, an early theoretical representation of autonomous vehicles.

The dream of self-driving cars has thus fascinated us for over 500 years. Have you and I been born into the period that realizes the dream? Are we that incredibly lucky? We think so.

Thomas Kuhn is one of our other disruptive heroes, whom we've described in several contexts. Kuhn's concepts of paradigm shifts and anomalies are a meaningful theory for understanding and analyzing powerful changes, or disruptions. An anomaly is something that the existing paradigm cannot solve and is therefore the underlying door opener for a new paradigm. For the transport sector, both costs, environmental impact, and traffic accidents are anomalies that human-driven gasoline cars have failed to solve.

The paradigm shift from privately owned gasoline cars to what we call "Transportation-as-a-Service" (TaaS) is a shift we're experiencing in this disruptive decade (2020-2030).

Introduction and Historical Backdrop

Aristotle, one of antiquity's greatest philosophers, placed great emphasis on categorization as a foundation for understanding the world. The significance of categorization for Aristotle lies in providing a framework for organizing thoughts and observations. By dividing phenomena into clear groups, he could analyze their properties and relationships systematically. We follow Aristotle's well-meaning advice and categorize our phenomenon.

To categorize self-driving units in a clear manner, you can group them based on use case, environment, or by degree of autonomy (SAE levels). We can also categorize based on how autonomous the units are (level 0-5, where 5 is fully autonomous) where partially autonomous (level 2-3) are cars, buses, trucks with driver assistance and fully autonomous (level 4-5) are Mars rovers, some cars and drones, last-mile robots.

A common denominator for all these self-driving agents is the varying complexity of the environment. It affects not only the ability to perform tasks precisely, but also safety. In our world, self-driving agents are driven by two factors: increased productivity and increased safety. The most important in an early phase is safety. Without self-driving agents having a high degree of safety and reducing accidents, society will not give its regulatory consent to the development.

The Disruption of Traffic Accidents?

Every single traffic accident is a tragedy. In Norway, the number of deaths in traffic has fallen from 500 in the 1970s to under 100 annually. There is broad political agreement through parliament's decision in 2001 about "Vision Zero" where no one should die in traffic by 2050.

At the time when the iPhone and smartphones were launched, there was a sudden increase in the number of accidents where pedestrians were the victims. This may be a spurious correlation, but we don't think so. We humans often sit on our mobile phones while our feet press the brakes and gas and we have one hand on the wheel. The eyes that should be on the road are on the screen.

These are hard, cold, and not least sad facts. In a world with self-driving cars, both of these numbers from inattentive accidents will fall toward 0. In a world with only self-driving cars, these numbers are 0.

The Fight Against the Seat Belt

When seat belt mandates were introduced in various countries, they met significant resistance and criticism, especially in the USA. Why?

Seat belts first became mandatory in the USA in 1984, when New York became the first state to introduce such a law. The introduction of these laws was part of a broader traffic safety campaign, but met significant public and political resistance.

Some quotes from the opposition to seat belt mandates:

• "Unless the Constitution has been replaced with actuary tables, the governing principles of this country are still based on basic rights of individual freedom" - New York Times reader, 1984
• "This is not supposed to be Russia where the government tells you what to do and when to do it" - Bronx resident
• "This is a pretty good lesson in mass hysteria created by a corporate-controlled media" - Michigan

A survey from 1984 showed that 65% of Americans were against mandatory seat belts and penalties for non-use, despite evidence that belts save lives. In the early 1980s, belt use was low, between 11% and 14% in the USA, but increased significantly after the laws were introduced. Despite all the resistance, the belt mandates stood. The numbers spoke for themselves. Deaths and serious injuries fell sharply. Today, over 90% of drivers in Western countries use belts. Statistics triumphed over resistance.

What Exactly is Full Self-Driving?

We ordinary people have read about people who have private drivers and have taken taxis ourselves. The concept that a car and driver can move you from A to B is well established. Now the driver is removed and replaced with a powerful AI computer, cameras (computer vision), and in some cases a bunch of sensors and radars. The consequence is that you can get your own private driver or significantly cheaper taxi services. For us, that's the core of the consequences of Full Self-Driving (FSD).

When Henry Ford assembled the Model T over a hundred years ago, the car was a simple machine that asked the driver to do everything. Today, Tesla (and some others) are turning the whole concept on its head. The company's plan in 2026 is to produce the "cybercab," a car without steering wheel, pedals, mirrors, back seats, or the panic sweat that comes with the first driving lesson. It's a stripped-down model of a "car" made for disruption with a price point between $25,000 and $30,000.

Elon has said that in full volume production, a "cybercab" will roll out every 5 seconds, that's several million "robots on four wheels" or popularly called cars per year. The business model and production capacity are ready for scaling in the coming years. The big question is whether the self-driving technology is ready. Tesla calls the hardware and software platform for self-driving Full Self-Driving (FSD).

Competitors like Waymo, Pony.ai, and Cruise have built their business model around LiDAR (Light Detection and Ranging). It's laser that continuously measures precise distances, but Tesla sticks to cameras and computing power. They removed radar in 2021 and ultrasonic sensors in 2023. The idea is that if humans can manage with vision and brain, AI should be able to do the same. In total, Tesla owners have soon collected over 4 billion miles with FSD or Autopilot activated. The numbers give faith that big data can replace expensive lasers.

FSD vs. LiDAR: Real Roads versus Legoland

If there's one thing you as an investor should take away about the difference between FSD and LiDAR, it's scalability. Think of FSD technology as a driver's license. With a driver's license from Norway, you can drive cars in the USA, EU, Australia, and Asia. A little local adaptation means we can relatively quickly drive on roads in all these places, even in the strange countries that drive on the wrong side of the road.

Waymo and other LiDAR systems only work where the maps are good and updated. A bit like when I was a proud little boy on a trip to Legoland and got a driver's license on the measured track among the sand dunes in Denmark. I could drive wherever I wanted as long as it was the track in Legoland.

Self-Driving and Autonomous Agents as a Disruptive Force

Self-driving enables new business models, such as Transportation-as-a-Service (TaaS) and data license revenues, and changes consumer behavior from ownership to experiences. But the road is bumpy. Safety, regulation, and ethics come into play.

The technology is a catalyst for changes in how businesses operate and how consumers behave. To make this easy to understand, let's use an analogy. Think of self-driving transport as a new type of "digital guide" on a medieval map. Before, you had horse and cart, which required a coachman (human effort). Now you have a magical cart that finds the way itself, and it can transport everything from goods to people, without stopping for rest.

New Business Models

Changing Consumer Behavior

When autonomous cars can pick you up anytime, why own a car? Younger generations already show less interest in owning cars and prefer flexible solutions. This shifts focus from product to service.

For consumers to use self-driving cars, they must trust the technology. Tesla has built trust by gradually introducing features like autopilot, so people become comfortable with letting go of the wheel. This changes our relationship with machines. We go from controlling to collaborating.

Think of a self-driving car as a personal butler. Before, you had to make dinner yourself (drive the car). Now the butler shows up with a finished dinner, tailored to your taste, exactly when you're hungry. You begin to trust the butler and value the experience more than owning the kitchen equipment.

Case: The Disruption of the German Auto Industry

The automotive giants like Volkswagen, BMW, and Daimler have long been a cornerstone of Germany's and European economy, as well as a global symbol of engineering excellence. Yet the industry has shown resistance to electrification and digitalization, such as in the development of software and self-driving cars. To explain why these large organizations have been and are being disrupted, we can use some theories we've picked up through analyses of innovation.

Technological Lock-in and Path Dependency

The resistance or inertia in investing in electrification and digitalization can be explained through the concepts of "technological lock-in" and "path dependency," both of which illustrate how historical choices and structures limit future opportunities.

Examples of Resistance

• Dieselgate: Volkswagen's attempt to manipulate emissions tests in 2015 showed desperation to maintain diesel's position, even in the face of stricter regulations
• Late EV Transition: While Tesla launched Model S in 2012, Volkswagen's first serious electric car, ID.3, came in 2019
• Software Struggles: Volkswagen's ID.3 launch in 2019 was delayed due to software integration problems, showing lack of expertise in this field

As Europeans, we hope the German auto industry can get out of the squeeze and maneuver through the next five years. Licensing of hardware and software may be one way forward.

European Self-Driving Initiatives

Volkswagen and Bosch have now announced some self-driving breakthroughs in the so-called "Automated Driving Alliance," launched in 2022. The ambition is to move their self-driving initiatives from expensive top models to regular family cars, and do it with European software.

The first milestone is estimated for mid-2026, when a new autopilot will be rolled out in mass-produced cars with access to hands-free driving on highways. The plan is to start at level 2 and gradually open for level 3 autonomy.

The alliance is building its own AI stack instead of leaning on American technology providers. The reason is the desire for digital sovereignty - control over models, code, and the upgrade path. At the same time, it's a practical path toward lower costs, because Volkswagen can extend the features to large volume models and not just the luxury segment.

Tests are already running with ID. Buzz prototypes on public roads in Europe, Japan, and the USA, and around a hundred extra test cars will be deployed to increase reliability. A huge data pool from around 45 million VW cars driving around the globe provides training data that few other European actors can match.

From Cars Sold to Miles Driven and Utilization Rate

One of Kuhn's main messages is that in the new paradigm, the industry's "tribal language" also shifts. Number of cars sold, how fast it goes, different colors, whether it can jump, acceleration, inventory, and gross margin on new models are thrown on the analytical scrap heap. 25 years ago, book turnover was how analysts and investors analyzed Amazon. There are probably some who still do, but it's not very relevant.

But what will be the new KPIs or key metrics for analyzing what we today call the auto industry? When the analytical focus shifts from number of cars sold to cars as part of "Transport as a Service" or TaaS, we must use new metrics that reflect service sales, usage patterns, and economic performance.

The Technology and New Ecosystem

FSD as the Interplay Between Cameras, Software, and Hardware

Imagine a skilled tennis player. The eyes follow the ball, the brain quickly calculates where it will land, the body responds with a precise shot. Tesla's Full Self-Driving (FSD) tries to do exactly the same thing, just for the road. The cameras are the eyes, the data chip is the brain, and the software is the tennis technique itself. None of them win points alone; development is about making the interplay flow.

Camera Technology Evolution

Hardware Evolution

• HW1 (2014): A modest Mobileye chip with barely 0.2 TOPS
• HW2 and 2.5 (2016-2017): Nvidia's Drive PX2 card provides up to 4 TOPS
• HW3 (2019): Tesla's own silicon, 144 TOPS, designed only for FSD
• HW4 (2023): Approximately 400 TOPS, enabling one enormous "end-to-end" model
• HW5 (2026/27): Rumored 4000 TOPS

Robotaxi: The Business Model Revolution

Self-driving technology transforms a car manufacturing company into a mobility company. Tesla launched its robotaxi service in Austin, Texas in June 2025. FSD is probably at least level 3 on the autonomy scale, technologically. What's needed to reach level 4? Then the person sitting in the passenger seat probably needs to be thrown out. If they succeed with that, Tesla will quickly move to level 5 because they then don't need to drive within a specific area.

Ethics and the Trolley Problem

Who is responsible if a self-driving car makes a wrong decision in an accident? Statistics are on the side of autonomy. But accidents will happen. This is a moral and legal minefield. On the positive side for self-driving are lower accident rates (human errors account for 90% of car accidents), reduced environmental impact (optimal driving routes), and increased accessibility for people who cannot drive themselves, such as the elderly or disabled.

Rollout Strategy for FSD and Robotaxi

The deployment strategy follows a careful progression:

• June 2025: Robotaxi established in Austin, likely rapid establishment in San Francisco, Los Angeles, and others afterward
• Focus on warm-weather cities with good infrastructure initially
• Gradual expansion to more challenging environments
• Building public trust through transparent safety data

Can the Market Increase by 25X?

In principle, Tesla can sell a car to a private customer or fleet owner at a loss on the car itself, as long as the customer buys a subscription to self-driving software (FSD). Tesla can also choose to produce cars they own themselves and use the cars in the robotaxi network. We can expect several different combinations and regional and national differences in the coming years.

Tesla has previously indicated a goal of producing 2 million cybercabs in 2026, equivalent to the annual production they've had on current models in 2024 and likely in 2025. Price, volume, and business model tied to these cars are unknown, but some numbers have been suggested. High gross margin and high scalability make the calculations interesting.

Infrastructure and Changing City Landscapes

These disruptive changes will shape urban infrastructure for many decades to come, and the change comes through more charging stations, fewer gas stations, fewer parking spaces, lower emissions, and greater traffic safety. At some point, the point-to-point transport of self-driving cars will challenge larger parts of public transport through cheap mobility between A and B.

Case Study: Oslo

What happens to a city like Oslo when self-driving cars become the norm? Consider the parking situation: Oslo has approximately 100,000 public parking spaces. If car ownership drops by 90% due to robotaxis, what happens to all this space?

The Three-Dimensional Infrastructure

Some of the freed space from privately owned cars will be freed up to be part of another infrastructure that will be built. There exists an infrastructure that is already built, maintenance-free, 3-dimensional (over, beside, and under), and follows the air line in efficiency. It exists above our heads. In a few years, city planners will also create drone corridors and landing spots for drones (former parking spaces).

Transport Revolution Beyond Cars

Last Mile Delivery

The "last mile" - getting packages from distribution centers to customers' doors - is the most expensive and complex part of delivery logistics. Autonomous delivery robots and drones are revolutionizing this space:

• Sidewalk robots like those from Serve Robotics navigating urban environments
• Amazon's Scout robots delivering packages in suburban neighborhoods
• Drone delivery for urgent medical supplies and remote areas
• Autonomous cargo bikes for dense urban centers

Self-Driving Trucks

Long-haul trucking is particularly suited for automation. Highway driving is more predictable than city streets, and the economic incentives are enormous:

• Trucks can operate 24/7 without rest requirements
• Fuel efficiency improves through optimal routing and platooning
• Addresses the chronic driver shortage in many countries
• Companies like TuSimple, Embark, and Tesla Semi leading the charge

Tractor 2.0: Agricultural Automation

Self-driving tractors and agricultural robots are transforming farming. John Deere, CNH Industrial, and startups are developing autonomous systems that can plow, plant, spray, and harvest with precision that exceeds human capabilities. This addresses labor shortages and improves yield through precise application of seeds, water, and fertilizers.

eVTOL: The Third Dimension of Transportation

Electric Vertical Takeoff and Landing (eVTOL) aircraft represent the next frontier in urban mobility. These aren't just electric helicopters - they're a fundamentally new form of transportation.

The Wright Brothers and eVTOL

Just as the Wright brothers transformed transportation in 1903, eVTOL pioneers are creating a new era of aviation. The difference? This time it's electric, autonomous, and designed for everyday urban transport rather than long-distance travel.

The Ecosystem

Leading Players and Their Designs

• Joby Aviation: Six-rotor design, 150+ mile range, backed by Toyota and Uber
• Lilium: Unique jet design with 36 electric ducted fans, 155-mile range
• Archer Aviation: Twelve-rotor design, partnership with United Airlines
• Volocopter: Multi-rotor design focused on short urban routes
• Beta Technologies: Fixed-wing design for longer range cargo and passenger transport

The Goal: Autonomous eVTOL

While initial operations will have pilots, the ultimate goal is fully autonomous flight. This dramatically improves economics by eliminating pilot costs and enabling 24/7 operations. The technology for autonomous flight already exists in military drones - the challenge is certification and public acceptance.

Real-World eVTOL Routes

Manhattan to JFK Airport

One of the most compelling use cases for eVTOL is airport connections. The Manhattan to JFK route currently takes 60-90 minutes by car in traffic. An eVTOL could make the journey in 8 minutes. At $200-300 per passenger (compared to $70-150 for a taxi), the time savings justify the premium for business travelers.

Norway's Potential

Norway's geography makes it ideal for eVTOL deployment:

• Bergen to Oslo: 1 hour by eVTOL vs. 7 hours driving
• Island connections avoiding ferry schedules
• Emergency medical transport to remote areas
• Oil platform personnel transport
• Tourism routes over fjords and mountains

Drones: The Workhorses of the Sky

Healthcare Applications

Drones are already saving lives by delivering medical supplies:

• Blood delivery in Rwanda by Zipline - reducing delivery time from hours to minutes
• Vaccine distribution to remote areas
• Emergency defibrillator delivery for cardiac arrests
• Organ transport between hospitals
• Lab sample collection from rural clinics

Commercial Use Categories

Regulatory Evolution

Regulations are evolving rapidly. The USA's FAA has established Part 107 for commercial drone operations and is developing rules for beyond-visual-line-of-sight (BVLOS) operations. Europe's EASA has created a comprehensive framework with different categories based on risk. Norway has been progressive, allowing BVLOS operations for specific use cases like powerline inspection.

Disruptive Investment Strategies for Autonomous Transport

The autonomous transport revolution presents enormous investment opportunities across multiple sectors. Understanding where value will be created and captured is crucial for positioning portfolios.

Investment Categories

Investment Thesis

Timeline and Milestones

• 2025-2026: Robotaxi services in 10+ major US cities
• 2026-2027: First commercial autonomous truck routes
• 2027-2028: eVTOL passenger services begin
• 2028-2030: Mass adoption of TaaS model
• 2030+: Fully autonomous transport becomes norm

6.0 eVTOL - The Third Dimension of Autonomous Transport

6.1 Wright Brothers and eVTOL

Human history and our mythical ideas contain many notions of flying like the birds above our heads. Despite many attempts, this remained a utopian dream until the Wright brothers defied the prevailing paradigm that humanity's feet should stay on the ground. Perhaps aviation faces a new chapter as electric vertical takeoff and landing begins its aerial journeys within the next year. Is this also a paradigm shift?

Before the Wright brothers took off in the early 1900s, the prevailing paradigm within transport and engineering was that human flight was impossible or limited to vessels driven by, for example, balloons lighter than oxygen. The Wright brothers challenged this by developing a manually powered, controlled aircraft, which radically changed how people viewed flying.

There were two innovations that changed the entire game. First was a three-axis control system that provided better steering and stability, solving the critical problem of controlling an aircraft in the air. Second was training - not large heavy neural networks in a data center, but wind tunnels. By using wind tunnels to test wing designs, the brothers introduced a scientific approach to aerodynamics, moving away from trial-and-error methods.

eVTOL technology developed for urban air mobility has similarities to the Wright brothers' innovations in its potential to disrupt established forms of transport. eVTOL aircraft use electric propulsion and vertical takeoff and landing, distinguishing them from traditional planes and helicopters. If the aircraft and services are successfully launched in the USA, this could represent a new paradigm in aviation and urban transport.

6.2 The Ecosystem

Behind every eVTOL concept lies a tight interplay between design processes, physical hardware, smart software, and the finished product that passengers ultimately board. The journey starts in the computer. Engineers let CFD programs "blow" virtual wind around prototypes, searching for minimal resistance and maximum lift, print parts in 3D plastic and stress-test them in wind tunnels before building a full-scale machine.

All theory falls apart if the machine doesn't function optimally. Therefore, much of the development budget is allocated to light, strong materials (e.g., carbon instead of aluminum), electric motors with multiple independent rotors for backup, and batteries packed as tightly as in a modern electric car, but with stricter requirements for fire safety and fast charging. Solid-state batteries are still expensive, but the industry follows Moore's and Wright's laws: more performance per kilogram and price drops for each doubling of production.

Where classic aviation is built around a human captain, eVTOLs are born digital. The autopilot connects control effectors, sensors, and increasingly advanced machine learning that learns local wind gusts between skyscrapers. At the same time, a cloud-based traffic system is needed that tells where other eVTOLs, drones, and helicopters are right now. Here, 5G networks, satellite links, and edge data processing form the digital air map that allows hundreds of airborne units to cross the city without meeting too closely.

Joby and Archer have room for 2 to 6 passengers and 50-300 kilometers range. DHL and EHang are betting on cargo versions that drop seats in favor of cargo volume. Regardless of variant, they need somewhere to land, so architects and energy companies are designing vertiports with wireless charging in the floor and automatic battery swapping. Actors invest early in infrastructure. A bit like when the gas station came before everyone had a car.

6.2.1 Vertiports - The New Infrastructure

For eVTOL aircraft to become commercially viable at scale, they need dedicated landing sites called vertiports. Think of it as multiple helicopter landing pads but with the infrastructure (buildings, check-in counters, chargers, etc.) compressed onto a rooftop or corner of a building. The platform itself, called the "Touch-down and Lift-off Area" (TOLA), is built to withstand both the weight and vibrations from multiple eVTOL aircraft landing and taking off minutes apart.

Around it you'll find fast chargers. Some manufacturers aim for 10-15 minute pit stops before the next departure. A small waiting room handles mobile check-in, quick security screening, maybe a cup of coffee. Nothing is superfluous here, as every extra square meter costs dearly in cities.

Behind the scenes, software works. The vertiport is connected to an unmanned traffic management system that monitors airspace in real-time and gives each aircraft a precise approach corridor, much like tower control does at a major airport, only autonomously. Ground sensors help the autopilot with precise landing even in rain or crosswind, and an energy management system distributes power between aircraft that need to depart quickly and batteries that should receive slower, gentler charging to last longer.

During rush hour, an aircraft can depart every couple of minutes, and any delay propagates quickly. Operators therefore need good communication between landing, charging, boarding/deboarding, brief technical inspection, and departure. The rhythm is key to making ticket prices low enough that eVTOL becomes a real alternative to ground taxis.

6.3 Is It Just Electric Helicopters?

Not exactly. It may look that way at first glance, but the eVTOL aircraft that Joby and Archer are developing differ significantly from classic helicopters. They're built around lightweight carbon fiber frames and electric propulsion systems that distribute thrust across many small rotors. This provides less mechanical complexity than a helicopter's large main rotor and tail rotor, and it opens up software-controlled approaches closer to the drone world than traditional aviation.

Today, eVTOL uses lithium-ion batteries storing around 250 Wh per kilogram. That's enough for city hopping between 100-200 kilometers, and manufacturers already point to 2030 cells exceeding 400 Wh per kilogram that will extend the radius to 3-4 times today's capacity. A helicopter solves the range problem with fuel but pays in terms of noise, safety, and CO₂. A modern turbine helicopter emits approximately 0.5 tons of carbon per hour.

The difference is also audible. Helicopters typically measure between 85 and 90 decibels at 100 meters, about like a chainsaw. Joby reports 60-65 decibels at takeoff and down to 50 in cruise, which is lower than a normal conversation. The economics point the same way. Joby and Archer suggest a price tag around $3 million per aircraft, about 1/3 of a medium-sized helicopter model with comparable passenger capacity. Operating costs fall even more dramatically because electric motors have fewer components and because electricity costs significantly less than Jet A-1 per kilometer.

6.4 Leading Players and Their Hardware Design

Joby has chosen a tilted rotor setup with 6 propellers that tilt from vertical to horizontal position when the aircraft transitions to cruise. The solution provides one continuous construction. The rotors lift the machine straight up from a roof or backyard, before pointing forward and flying through the air at around 300 km/h. Underneath sit lithium-ion battery packs storing capacity for about 240 kilometers range with 4 passengers plus pilot. Six independent electric motors build in backup exceeding requirements from both FAA and EASA. This means one motor can fail without the aircraft losing control.

There are several different paths to space. Archer's "Midnight" has 12 fixed rotors mounted on a wing. Six of them lift at takeoff, while the other six push forward constantly. The rotors don't tilt; they just change work tasks when the aircraft tips its nose down to accelerate. The principle halves the mechanical complexity compared to Joby and provides a possible advantage in production time and certification work. The battery packs are also smaller.

Both companies have much in common. The fuselages are cast in carbon composite to cut weight, and both pack the aircraft full of lidar, radar, and cameras that monitor airspace and provide the data foundation for the autonomous functions coming gradually. Both bet on fast-charged lithium-ion cells that refill in about fifteen minutes. And both have brought in automotive industry scale expertise. Joby gets support from Toyota, while Archer builds factories with Stellantis.

6.4.1 Vertical Integration Strategies

Vertical integration is when a company takes control over multiple stages of the production process, like owning both suppliers and sales channels. It can be compared to a chef who grows their own vegetables, makes the food, and runs the restaurant. This is to ensure quality and cost control. For eVTOL manufacturers, this often means owning the production of components like batteries, motors, and structural parts, as well as operating their own air taxi services.

Joby is perhaps the most vertically integrated company in the eVTOL industry. They develop, design, and produce most of their components internally, including batteries, motors, and structural parts. This includes their factory in Marina, California, where they've established a 55,000 square meter production unit, with plans to expand to 500,000 square meters. By owning production, Joby can optimize costs and reduce dependence on external suppliers, which is crucial in a new industry where supply chains are still developing.

Archer follows a more traditional approach by relying on Tier 1 suppliers for many components. They focus on design and assembly while outsourcing component production to established aviation suppliers and Stellantis. By outsourcing production, Archer can focus on design and operations, potentially reducing initial investments. Collaboration with established suppliers ensures high-quality components and leverages their specialized knowledge. However, dependence on suppliers can limit their ability to control costs, quality, and timelines.

6.4.2 Batteries - The Critical Component

Researchers and manufacturers still expect solid progress in battery development in the coming years. Small improvements in today's lithium-ion chemistries will lift the numbers to 350 or 400 watt-hours per kilogram before the decade is over, and prototypes with solid electrolyte can push the level even higher. When solid state reaches 500 or 600 watt-hours per kilogram, the result could be up to 50% longer trips without extra weight. Archer flies shorter routes initially but can roughly double the distance from 100 to 200 kilometers, and Joby from 250 to 500 kilometers.

Who delivers the cells is strategically important. Joby assembles the battery pack themselves and has so far bought cells from Panasonic, among others. In-house production means Joby can switch to a better cell type when ready, for example a solid-state variant, without changing the entire architecture. Archer has chosen a simpler model where Taiwanese company Molicel delivers finished cells that go straight into the pack. This provides faster startup and lower development costs but binds them more tightly to one manufacturer.

When the FAA eventually grants type certificates, the battery is defined as a critical part of the aircraft on par with rotors or control systems. A change of cell supplier after approval therefore triggers a new round of tests for fire safety, charge control, and thermal stability. The process costs time, typically half a year or more, and requires access to test aircraft and engineers who could otherwise work on operations.

6.5 The Goal: Autonomous eVTOL Aircraft

Transforming Joby and Archer's models from manned to fully autonomous requires first and foremost that cabins are ripped out and replaced with more cargo space. Both Joby and Archer must be equipped with more and better sensors than today's versions. Cameras already installed will be joined by radar, LiDAR, and infrared eyes providing full visibility in darkness, fog, and sandstorms. All data feeds into dual, encrypted autopilots that can fly the machine entirely without human help.

The software builds on the flight control algorithms that already handle transitions from vertical to horizontal flight, but the learning loop expands with pattern recognition that understands traffic patterns in urban airspace or changes routes when air defense radar flares up on the battlefield. The system also tracks cargo weight, center of gravity, and any temperature sensitivity so medical samples arrive intact.

When all this is in place, civilian drones can glide into Norwegian logistics chains and take the route from Oslo's Østland terminal to Drammen in fifteen minutes or deliver blood bags from Rikshospitalet to Ullevål three times faster than an ambulance in afternoon rush hour. The cost per kilogram per hundred kilometers is set to fall below two dollars, well below helicopter prices even before the fuel bill is added. In Northern Norway, the same platform can replace short-haul aircraft that today transport mail and medicines out to the coastal strip.

6.6 Production Capacity and Timeline

Joby and Archer have chosen two quite different paths to mass production. The former keeps as much of the value chain close to its chest as possible. The first 5 aircraft have already rolled out from the line in Santa Cruz, and the same line will build additional prototypes entering a test program during the year. In the longer term, the volume effort moves to Dayton, Ohio, where Joby has bought a factory of just under 4,000 square meters, plus an option on a land area the size of a factory ten times larger. Toyota assists with expertise and capital.

Archer's new Georgia factory of almost 40,000 square meters is already fitted out, and Stellantis will run the line the same way they produce hundreds of thousands of car models annually. The ambition is to leave 2025 with two new aircraft per month, to over 650 total by 2030. The building's capacity is dimensioned for 650 machines annually, and Archer has included options for expansions.

When production models are in the air, both companies meet a physical limit that cannot be negotiated: battery life. Today's lithium-ion packs can handle perhaps 1,500 cycles before capacity becomes too low for scheduled traffic. That corresponds to 3 to 5 years in daily operation. Batteries can be replaced, but each change costs six figures in dollars. The carbon fiber airframe itself and the electric motors behind the rotors are designed for about 12 years of service, around 25,000 flight hours, so battery replacements will likely occur 2-3 times during the lifecycle.

Investment Implications for eVTOL

Looking through Clayton Christensen's glasses, eVTOL is typically disruptive. It starts as a niche but offers something entirely new where cars and helicopters struggle: sustainable air transport adapted for cities that also doesn't create noise pollution. As technology curves make machines cheaper, quieter, and safer, the niche will grow into a market that can eat into both the ground taxi industry and the shortest regional air routes.

Both Joby Aviation and Archer Aviation plan a soft launch in the Middle East in Q4 2025. For a US launch, they still need some stamps from American authorities to get permission to fly with paying passengers. Landing sites must also be built and ecosystems launched. But following the eVTOL industry and these two companies will hardly be boring in the coming years.

The disruptive potential is enormous. In the urban market, where noise, emissions, and price are the most important decision factors, eVTOL will likely eat large pieces of the helicopter market before 2030. We can see eVTOL routes between airports and city centers, and between cities and nearby towns long before electric helicopters venture out to offshore platforms. Offshore transport, military operations, and rescue helicopters in the wilderness will remain the helicopter's most important areas for a long time to come.

The helicopter industry won't be eradicated, but it will lose the short, lucrative routes that have subsidized parts of the business. And when the revenue base shrinks, they're pressed to streamline and innovate faster than they otherwise would have. That's exactly the hallmark of a genuine disruptive wave. It appears from the side, fills a need that established players overlooked, and forces everyone to fly in new ways.

Conclusion: The Autonomous Revolution

Self-driving agents are one of the four horsemen, as we call them, that will ride into the labor market and replace humans' hands, feet, and cognitive processes' contribution to economic growth. The paradigm shift from privately owned gasoline cars to Transportation-as-a-Service represents one of the most significant transformations in modern history.

The technology is ready or nearly ready. The business models are proven. The economic incentives are enormous. What remains are regulatory approvals, public acceptance, and infrastructure adaptation. These are solvable problems, and they're being solved right now.

For investors, this transformation presents both enormous opportunities and significant risks. Companies that successfully navigate the transition will capture trillions in value. Those that resist or ignore it will join the scrap heap of history, alongside the horse-and-buggy manufacturers who dismissed the automobile.

So let's see if we're right in some of our assumptions. It's exciting regardless. We have limited ambitions to write ourselves into the Nobel Prize for Economics or other academic exercises, nor to be part of the new lexicon of the Truth Directorate. We manage stocks, not truths. We only share our perspectives at the time we write them, and if you feel like buying something, we hope it's a gift for someone close to you.

With that starting point, these are our perspectives on self-driving agents. The dream that has fascinated humanity for 500 years is finally becoming reality. And we're fortunate enough to witness and invest in this transformation.