Autonomous Investment Choices?

Autonomous mobility promises to be one of the most profound disruptive technologies of the next decade. But what does a smart investment roadmap to navigate these opportunities look like?

Autonomous mobility technologies, with their limitless potential to deliver wholesale disruption of established businesses, represent extraordinary investment opportunities. However, consistent with most deep tech, the investment case is fraught with structural challenges including high investment quantum, long-tail investor yields and the danger of IP obsolescence during lengthy investment cycles.

But for investors who know and understand the challenges that come with the deep tech territory and are behind the soaring growth in private capital flows into the sector (estimated at 22% YoY since 2015 by Boston Consulting), autonomous transportation brings its own particular investment playbook — forget the eye-catching moonshot pitches and rather think of tech that makes a clear near-term viability case, is based on low-speed and the delivery of services.

This approach is investigated by the CEO of a very distinct autonomous mobility company, StreetDrone and Lukas Neckermann, a strategist, advisor, researcher, and author on the subject of Smart Cities and Smart Mobility. For his part, Neckermann contends that “Either by design, or by necessity, European-based autonomous tech developers have been challenged to be more creative and frugal in allocating their resources. From an investor point of view, this means, more is being done with less. Given that there is certainly also a need and public desire for home-grown autonomous tech, the European investor is well-placed to look at home, rather than abroad, for value.”


No Route to the Moon

From the Neckermann perspective, therefore, the first rule of thumb for investors surveying the autonomous mobility landscape is to ignore the industry’s poster boy, Waymo and the skewed impression of the market opportunity that the company singlehandedly generates. As a subsidiary of Alphabet Inc, Waymo and a number of its competitors are distinct in their pursuit of fully automated driverless vehicles with the capacity and capability of a human driver. Despite the over-amplified speculation of Tesla’s Elon Musk and others, many industry experts consider the full automation to be almost unachievable and a more complex technical challenge than putting man on the moon. And even if the technical challenge was resolvable, the investment cost and timeframe render these ambitions as the stuff of indulgent R&D rather than a worthy investment pursuit.

As Waymo’s CEO, John Krafcik commented last June, “Level 5 is a bit of a myth. Level 5 means you can drive anything anywhere in any weather conditions, like, you can drive from San Francisco to Santiago, Chile, any time of year, just press a button. This is probably never going to happen. Humans don’t even do this.”

To date, Waymo have raised $3 billion in its first external funding round and is understood to have consumed equivalent sums courtesy of the deep pockets of its parent company. To show for its efforts, the company operates a 500-strong fleet of robotaxis in the Metro Phoenix area in Arizona, Mountain View, California and a number of other locations in the US. The maths is crude, but the amortised cost per vehicle of a sum approaching $6m is the clearest indicator that the creation of a market for private ownership of self-driving cars even in the medium term is about as likely as mass space tourism.

The Waymo robotaxis operate at a peg below completely capable robot driving (otherwise termed ‘Level 4’), but still require a human safety driver in every vehicle until the tech is fully validated. One of the primary commercial motivations of removing costly human resource from the overhead of running passenger or delivery services is, therefore, as yet unrealisable. Massachusetts Institute of Technology and the FT’s Alphaville suggest that even an ‘automated hive of driverless taxis would actually be more expensive for a consumer to use than the old-world way of owning four wheels.’ It would seem as if this high profile enterprise is not, therefore, focused on meaningful commercial viability.

One of the consequences of a pool of large, well-financed pre-profit autonomous mobility tech companies on the West Coast of the US is the significant distortion of the autonomous mobility market. Capital is inefficiently allocated, expectations of commercial and technical potential are overstated and the cost of key resource, such as the supply of capable R&D engineers, is inflated because these substantial and influential companies are in pursuit of R&D objectives rather commercial imperatives.

As far away as the other side of the Atlantic, the ‘skew’ created by these autonomy behemoths is keenly felt. StreetDrone, a leading full-stack autonomous software & vehicle company in the UK, faces a continuous battle to re-set investor perspectives on more realistic opportunities than Waymo’s moonshot ambitions.


Services, Not Products

“Waymo has been responsible in part for creating a powerful notion that we’ll all have self-driving cars parked on our driveways in a few years,” says StreetDrone’s CEO, Mike Potts. “There is no precedent in the history of humanity for tech that costs upwards of five million dollars to be progressively commoditised to reach the point of affordability of an average car.” He continues, “The problem is fundamentally misconstruing the purpose of an enabling technology. Its purpose is not to refine a product such as a car that has no need of a driver, its purpose is to deliver services. The step that makes a ‘service’ (rather than a self-driving product) viable is the removal of the safety driver from any autonomous vehicle to drastically reduce its operating overheads. That progress will get an investor a whole leap closer to something with dependable yield potential.”

But the ambition to dispense with the costs of the safety drivers who currently oversee the operation of all robotaxis in downtown Phoenix is something Waymo holds in common with StreetDrone. So the first investment canon is clear — focus on service opportunities, not products. But how does a service become viable if a human driver is needed to oversee the safe operation of expensive technology?


Less Pieces in the Jigsaw Puzzle

The solution to making autonomous technology commercially viable is to deploy in near-field applications that can be realised and scaled soon, pushing driverless vehicles up the technology readiness index. Not only does a near-field approach eradicate many of the liabilities of deep-tech investment, but it provides a solution to scale as its dependence on capital is reduced. So what essential components of near-field solutions should investors be looking out for?

One of the primary drivers of cost and extending R&D timelines in autonomous technology is speed. As StreetDrone explain, the technology costs per 1km/h over a threshold of 32km/h sends the engineering curve into an exponential increase in complexity. So keeping autonomous solutions slow is critical. Low speed means far less tech and a quicker route to removing the safety driver from the loop. As soon as this principle of slow-speed autonomy is enshrined, suddenly the rest of the jigsaw puzzle — now made up of less pieces, begins to fall into focus.

Having determined that twenty is plenty, the application of autonomous solutions takes shape. People, on the whole, are not given to travelling slowly, (except perhaps in cities where smart driverless cars do have the potential to overcome the traffic management inefficiencies that cause congestion) which immediately pre-disposes commercial applications towards delivery services. Average last-mile delivery speeds are, in any case, already slow due to structural factors such as urban speed limits and the stop-start delays caused by drivers dropping to the doorstep.

The insight that slow-speed applications such as contactless pharmaceutical deliveries, grocery drops or automated provision of municipal services such as street sweeping and refuse collection are the way ahead is not necessarily revolutionary, but the watchword for investment value are use cases that can operate at low speed, can scale and are viable in a defined operational design domain, or ODD. The ODD defines the complete operating parameters for an autonomous application. More sensors, more hardware redundancy and more code may be required to ensure safe operation of autonomous applications in bad weather, or in complex environments, so the less jigsaw pieces the ODD requires, the closer commercial viability will be.


Let’s Play Leapfrog

StreetDrone’s approach presents an interesting investment alternative to the $$$b dollar R&D enterprises on the West Coast. At every turn, they pursue methods of reducing cost — for instance, adopting open-source as an approach to reducing software costs or subscription payments for services to defray large capital outlays for customers. All of these initiatives aid the commercial viability case and thus the route to scaling the use of driverless solutions. In itself, however, these may not be enough to drive an investable proposition.

So the final part of the StreetDrone playbook which provides a useful insight for investors is the partnership model the company is now adopting. This has the scope to provide clearly definable revenue streams and shift the technology risk from the autonomous service providers to the service users.

“We have been very successful in the R&D market, but this is very finite. Our approach for growth been therefore the pursuit of ambitious partners who need a driverless service that saves costs over manned delivery – and we provide the wheels for that ambition,” Potts says. “And don’t look in the obvious places. For instance, many grocery retailers and supermarkets in the UK have manned delivery services. But there are a substantial number who do not, and their business vulnerability has been exposed by the pandemic. An automated, slow speed, neighbourhood delivery service provides these retailers with an opportunity to not just catch up with the competition, but at a stroke, to leapfrog them by removing all of the overheads associated with manned deliveries.”


Smart Isn’t Sufficient

Smart engineering is essential – but not enough on its own to build a credible prospectus for investors considering driverless tech options. To create a robust investment opportunity, it pays dividends to steer away from over-ambitious moonshots especially when the technology ambitions are well ahead of the business case. Autonomy to solve simple problems slowly and limited ODDs to provide a service that is scalable in the near term is the price of entry for portfolio consideration. But to be a surefire opportunity, robust partnerships with multiples, retailers and municipal authorities in need of a strategic point of difference needs to be in evidence.

Are Electric Vehicles Really Better for the Environment?

Are Electric Vehicles Really Better for the Environment?

(First published in The Engineer, August 9, 2021)

If FCEVs and BEVs are to replace ICE vehicles (ICEVs), then it is important to establish that they are truly beneficial overall and reduce negative impacts in a broad range of contexts from ecosystems to human health to demand on natural resources. With future regulations expected to focus on a more holistic view of product life, HORIBA MIRA’s Advanced Energy Research Scientist, Adam Zucconi, considers the debate in more detail.

To best understand the overall impact of both classes of vehicle, life cycle analysis (LCA) is applied. This is a complex process of quantifying a product’s impact from the extraction and refinement of raw materials to the manufacture of components, the assembly process, the production and consumption of energy throughout the operational phase, and finally, the end-of-life phase, varying from disposal to reuse or recycling.

There are several drivers supporting the use of LCA. The European Commission has released a new ‘Circular Economy Action Plan’ that connects product design to end-of-life to promote a circular economy. For instance, it promotes a framework for battery manufacturing that demands sustainability and transparency requirements to span the carbon footprint of manufacturing, the ethical sourcing of raw materials and the reuse, repurposing and recycling of used batteries.

During the manufacturing stage FCEVs and BEVs produce up to two times the amount of greenhouse gases compared to ICEVs

LCA is a topic of great interest to HORIBA MIRA in order to provide guidance on the development of appropriate standards and to support the industry in designing vehicles that reduce negative environmental impacts and adhere to any new regulations.

HORIBA MIRA recently conducted a review of life cycle analyses utilising a wide variety of data, models, and databases to assess the impact of FCEVs and BEVs. The high level impact areas of interest were climate change, human health, ecosystems, and natural resources, which are quantified using metrics including greenhouse gas emissions (kg CO2 eq), human toxicity potential (kg dichlorobenzenes eq), acidification (kg SO2 eq), abiotic depletion (MJ), etc. Careful attention was also paid to the system boundary definitions and life cycle inventories, both of which have substantial impacts on the results.

LCA identified that during the manufacturing stage FCEVs and BEVs produce up to two times the amount of greenhouse gases compared to ICEVs. In FCEVs, this is predominantly caused by the platinum extraction and processing for the fuel cell stack, as well as the energy consumed in carbon fibre production in components such as hydrogen storage tanks. In BEVs, the battery cell manufacture has the most significant impact. The LCA also demonstrated that for both human health and resource depletion concerns, FCEVs and BEVs also perform worse than ICEVs. For EVs, if the manufacturing and assembly process uses low-carbon energy, the impact significantly decreases.

Crucially, the operational phase, which includes the distribution and consumption of fuel, is where FCEVs and BEVs have the greatest potential advantage over ICEVs. The lack of local exhaust emissions and significantly higher efficiencies compared to ICEVs are paramount to this. However, this advantage can only be achieved using low-carbon and/or renewable energy. The length of vehicle life is also crucial, as sufficient time is required to overcome the higher manufacturing impacts of electric vehicles.

In terms of climate consideration, FCEVs can deliver reductions up to 85% in greenhouse gases, while BEVs can achieve reductions up to 95% over comparable ICEVs. The main factors impacting these benefits are the hydrogen production method – where compression is favoured over liquefaction – and the use of low-carbon and/or renewable electricity e.g. nuclear, wind or hydro generation. One study suggests that if coal was substituted for clean electricity generation for BEVs, their total impact would be approximately four times worse than the equivalent ICEVs.

End-of-life for FCEVs and BEVs provides an opportunity for abatement of impacts through the reuse of materials, where the use of renewable and low carbon energy to manage these processes is key. FCEVs have good recyclability potential (around 90%), while figures vary – between 50-90% – for BEVs, with current technology at the lower end.

From HORBA MIRA’s analysis of available data, the answer to this important question of whether EVs are actually beneficial to the environment throughout their life cycle is ‘yes, providing certain conditions are met’.

These conditions include a sufficient proportion of renewable or low-carbon energy used throughout the entirety of the life cycle; for FCEVs, hydrogen should be compressed, as liquefaction is too inefficient; the vehicle lifetime needs to be sufficiently long to overcome the significant impact of the manufacturing process, which outweighs today’s ICE vehicles.

With the expectation that legislation will move towards a circular economy and a more holistic view of product life, and discussion still open about what those standards could be, HORIBA MIRA is in a position to guide and support the development of these standards, as well as provide consultancy to businesses creating products and services in line with LCA requirements. This could be for the entirety of a vehicle, or focused on specific components that have a particularly negative impact.

Platooning – Making Space to Drive?

The number of electric vehicles on UK roads is expected to hit 25 million by 2035; an increase that will significantly exacerbate congestion. While political solutions are being sought to contend with this predicted rise, platooning technology for passenger vehicles has the potential to improve motorway utilisation and make significant vehicle energy savings, which have been conservatively estimated at between four and ten per cent.

To keep pace with government carbon reduction plans and electric vehicle (EV) deployment targets, it is anticipated that the number of EVs on UK roads could increase from 100,000 today to three million by 2025, soaring to 25 million by 2035.

These figures have been outlined in a paper by the Tony Blair Institute for Global Change, which goes on to suggest that the lower taxes and reduced running costs compared to ICE vehicles will encourage more use of private cars and more road miles to be driven by motorists. Consequently, congestion is set to rapidly worsen, and the time wasted while stuck in traffic is at the same time predicted to rise by up to 50%.

The suggested solution is to reform taxes and shift to a road pricing methodology, yet Innovate UK’s Transport Vision 2050 report indicates that this could take as long as 2040 to introduce. Even with tax reform to discourage unnecessary road travel, population growth means that congestion will remain a significant issue for the foreseeable future, resulting in lost productivity and reduced quality of life.

As a result, other solutions are needed to help ease congestion, increase road utilisation and improve traffic efficiency, regardless of political changes. Platooning is one such a solution. While platooning has been on the horizon for some time, the technology typically focuses on commercial vehicle applications. Yet the development of such solutions for passenger cars has the potential to be ground-breaking, with the benefits significant and wide-reaching.  Platooning optimises the flow of traffic streams, allows shorter distances between vehicles while still ensuring safe operation, and prevents ‘perturbations’ – the phenomenon of waves of braking and acceleration that transmit through manually controlled, high density traffic. It enhances passenger comfort, increases vehicle efficiency, lowers emissions and improves utilisation of the existing road infrastructure, which in turn reduces the need for new road building to accommodate the anticipated rise in traffic levels.

Cooperative positioning between platooned cars, comprising advances in ranging and vehicle communications, along with algorithms developed to ensure data integrity, enabled by a variety of range-finding sensors, with the resulting data communicated between vehicles via low latency communication protocols. The development of algorithms to ensure data integrity – and required to authorise the positioning of platooning vehicles – will be hugely significant to such solutions.

While there are a great many factors that influence traffic density on any given road – from weather conditions to highway design, platooning on main arterial routes has the potential to increase capacity compared to manual driving, while improving vehicle efficiency and energy consumption, and simultaneously improving passenger comfort.