Road freight is one of the most traditional sectors of the global economy, and it too is undergoing a radical transformation that doesn’t get nearly enough attention — one that touches both propulsion technology and vehicle automation. Several companies are working on this, but Swedish firm Einride stands out for combining autonomous driving with electric heavy transport through an ecosystem that genuinely delivers, already working with clients such as Heineken and PepsiCo.
We spoke with Henrik Green, General Manager, Autonomous Technologies at Einride, who walked us through the developments and evolution of this technology.
Trucks without a Cab
Einride is first and foremost a developer and provider of electric and autonomous trucks, and the first thing that strikes you when you look at them is that they have no cab. In fact, in their fully autonomous versions, they travel with no driver on board whatsoever. This falls under what the industry classifies as level four autonomy — a distinction Green is keen to spell out precisely.
Image: Einride
“The key with level four is that it’s driverless — there is no safety driver on board,” he explains. “That means that we have a completely cab-less truck that has no cabin and no driver on board. And that’s really the ultimate goal for autonomous transportation.”
The difference from lower levels — two and three, where a driver is still present and ready to intervene — is substantial: removing the human factor from the operational chain allows for higher vehicle utilisation and significantly lower costs.
As for level five, full universal autonomy, Green downplays its practical relevance: it would mean a system capable of driving anywhere and in any condition, from the Alps in a blizzard to the backstreets of any city on earth. “That’s probably very far away — but that’s not the problem with level four. We achieved the goal that we are after.”
Einride processes millions of data points to manage its autonomous vehicles, but the approach to AI is anything but blind. Green describes a dual-path strategy that combines machine learning with rule-based deterministic algorithms.
“A machine learning algorithm provides the absolutely best human-like driving behavior in most of the scenarios — it is at the same time still challenging to prove exactly how it will behave because of its non-deterministic nature in any given condition.”
For this reason, classical deterministic algorithms act as guardrails, defining the boundaries within which the machine learning component is allowed to operate. This is not merely an internal technical choice, but a regulatory necessity. “When you need to go to the authorities to get a road permit for public road, it more or less requires you to deterministically prove that you’re safe in the conditions where you aim to operate.”
Alongside this architecture sits Saga AI, Einride’s proprietary transport planning system. Saga AI does not handle vehicle driving, but orchestrates the entire logistics chain: assigning routes, scheduling charging stops, distributing loads across vehicles, and optimising the deployment of human operators wherever they are still involved. It is a crucial element even with driver-operated electric trucks, because unlike a diesel vehicle — which can refuel in minutes anywhere — an electric vehicle requires precise planning of when and where to charge, to avoid wasting time and resources.
A reality, even in Europe
On the European front, Einride has obtained permits to operate on public roads in Sweden, Norway and Belgium, and is working with authorities in the Netherlands and Germany. Denmark has not yet issued equivalent permits. Green is nonetheless optimistic: “I don’t think there should be any rational reasons for that to fall behind. I’m expecting Denmark not to come a lot after Norway and Sweden.”
The European regulatory landscape, as we know, has its strengths and weaknesses. On one hand, Germany introduced a dedicated framework for autonomous vehicles as far back as 2021; on the other, fragmentation across individual member states forces companies to negotiate country by country, multiplying complexity and costs. Green would welcome a unified EU-level regulation, but acknowledges that — contrary to what many assume — the United States faces a similar challenge, with rules that vary from state to state in the absence of a federal standard.
On the competitive front, Europe lags behind the US and China primarily in terms of risk capital available for scaling. “Companies are growing faster and moving faster both in US and in China — and I would attribute that to the access of risk capital in those countries.” The Old Continent, however, can claim an industrial tradition in automotive — and heavy transport in particular — that is hard to match anywhere else in the world.
Responsibility
The question of legal liability is one of the most complex issues in the entire industry. Green answers without hedging: in the event of an incident during an autonomous operation, Einride is the responsible party, as it assembles the vehicle, manages its operation, and ensures its safety and maintenance.
This vertical integration — from truck design to service delivery to the end customer — is, according to Green, the only viable model at this stage. “If we as a player just say, here is an autonomous truck, you can do whatever you want with it — it’s very difficult to get a permit. It’s very difficult for somebody else who didn’t develop it to operate it and understand how it works and do that in a safe way.”
But there is also a question of employment responsibility — one that autonomous driving is frequently accused of undermining, namely the elimination of jobs. And yet, in the heavy transport sector, the shortage of drivers has been a serious problem for years, one that this technology could actually help resolve. Add to this the creation of new roles — remote operators, dispatchers, technicians — that in many cases offer more regular working conditions, more compatible with family life, compared to long-haul driving, which is widely seen as more gruelling and wearing.
All that said, interest in Einride and autonomous driving more broadly is growing steadily. Green notes that market sentiment has shifted radically over the past two to three years. The rise of generative AI — with tools like ChatGPT entering everyday life — has changed public perception of what technology can actually do, paving the way for greater acceptance of autonomy on the road as well.
