This time around, self-described “Tesla EVangelist” Steven Peeters is starting to question his decision to once again spend thousands on the carmaker's vaunted full self-driving (FSD) technology.

Drawn to the brand by Elon Musk's maverick style, the Belgian has long been fascinated by the idea of autonomous vehicles. Peeters ordered his current P100D right after Tesla announced in 2016 that all new cars would ship with the necessary hardware to enable self-driving once the software was developed. It has since become something of a hobby for him to track improvements in his car's intelligent driver assistance system and upload the results to his YouTube channel.

Peeters, the quality assurance chief at electric vehicle equipment provider Powerdale, has now been waiting for months to take delivery of a top-of-the-line Model S Plaid complete with FSD. For him, it will be the third Tesla in a row.

Yet while the first 1,000 or so American Tesla owners with perfect 100/100 safety scores have since October 11 been downloading the latest feature that steers the vehicle automatically in dense urban traffic, the rest of the world is at the back of the queue because of a mixture of regulatory hurdles outside Tesla's control and Musk's own strategic priorities.

When, and in what state, this technology eventually comes to Europe and beyond remains to be seen.

Shaking loyalty

If another EV competitor is able to beat Tesla to the European market with this technology, Peeters says he would reconsider his loyalty to the brand the next time he shops for a new ride. Simply offering auto-steering doesn't cut it either, as he bought FSD the first time around on the promise of being able to sleep in his car while it handles the driving for him.

“I paid €9,000 [about $10,400] for something that is basically worth half that because the rest of the promised features have not been delivered after almost five years of owning my car,” he told Fortune.

This is no small issue, either: FSD is considered fundamental to the equity story underpinning Tesla's lofty $800-billion-plus market capitalization. In January, Musk speculated FSD could one day generate profits of roughly $50 billion annually. Conservatively assign a multiple of 20 times earnings on that contribution alone and Tesla should easily break into the elite $1 trillion club alongside Apple, Google, Microsoft, and Amazon, he told investors.

Since the version 10.2 that went live last week only opened up Tesla's wider U.S. community to the previously closed "beta" test, customers must still supervise the system at all times. The carmaker explicitly warns them "it may do the wrong thing at the worst time" and drivers should "be prepared to act immediately." For that reason, it is prioritizing customers with the highest safety scores calculated on the basis of Tesla's own in-house risk models.

Ignore for one moment the point of the technology isn't to constantly watch a car steer itself like it was a teenager who just earned a driver’s permit, but to free up your valuable time to do other activities instead. That is not to be, at least not yet, with the current FSD beta setup.

What is noteworthy here is that Tesla does not even bear any legal liability should a crash occur while the experimental technology is engaged, hardly a sign of confidence by its lawyers. Instead, the customer must foot the bill both physically and financially.

Flipped relationship

The FSD beta flips the traditional vendor-customer relationship on its head. In a sense, Tesla owners have paid Musk for the privilege of helping him finish development of a feature they purchased, as their data is harvested and fed into Tesla's neural network to train the self-driving software. No other automaker has dared pocket so much money in exchange for a product it has yet to develop, nor rolled it out in an unfinished state.

Many Tesla fans readily acknowledge being citizen employees—some might even call themselves guinea pigs. Yet they are happy to volunteer if the reward is being the first on their block to test the limits of the new technology. Closed beta testers have also diligently advertised FSD for free, uploading online what are likely thousands upon thousands of hours they've collectively driven—subject to their contractual agreement with Tesla—ever since they were selected for early access last October.

Part of the reason why Europe has to wait along with the rest of the world for Musk's FSD is the uniquely favorable environment in the U.S., where risks of this kind are accepted by authorities. New innovations in automotive technology are in principle allowed so long as they are not expressly forbidden. Of course, this regulatory approach can backfire on companies in the form of lawsuits if they take potentially too great of a risk, given the fuzzy nature of what is known as "self-certification."

By comparison, many other jurisdictions, including the EU’s own Single Market, require approval by technical authorities (or "homologation") based on specific international standards set by a United Nations body in Geneva called the UNECE.

While carmakers welcome the legal certainty this provides them, progress depends on a painstaking process of building broad consensus among dozens of different countries, many of whom have no auto industry lobbying them to speed things up.

For this reason Musk prefers to focus on the U.S., where the pace of FSD rollout is largely in his hands and he can opt to accept or ignore recommendations from the National Traffic Safety Board at his leisure. On October 7, for example, he suggested for the first time to shareholders at an annual meeting that he is considering whether to implement a cutoff for customers that bought FSD but whose Tesla scores fell below 60 out of 100.

“Elon time”

With few options at their disposal, impatient European owners have often taken to Twitter to ask Musk directly when they will finally be able to access the FSD beta program like their American counterparts. Worse, some of Tesla's software that is available in Europe is less effective than those of homegrown rivals, such as in its ability to correctly respond to certain traffic signs like dynamic speed limits on motorways.

“They’ve always been focusing on the U.S. right now, while we’re not getting anything,” said Peeters, who even offered to serve as a Tesla test driver to help gather data for Europe if it would help move things along.

During early October’s County Fair at Tesla’s upcoming German assembly plant outside of Berlin, it didn't take long before Musk had to field questions from the crowd on the sore subject.

“Next year sometime is quite likely, so it’s looking good,” he told them on Saturday in his signature manner, which conveys general confidence without venturing too far into the realm of actual certainty. “FSD, which will be beyond beta I think, is looking highly likely to be in Europe next year.”

The problem is many Tesla fans now know to take these timetables with a grain of salt. Musk's prophecies are notoriously optimistic, calculated as they were on “Elon time.” While Tesla drivers trust that events will transpire just as he foresees them, they realize the plans could be fulfilled much, much later than expected. Only early October, for example, the Tesla CEO was forced to delay Cybertruck’s start of production by a full year, citing ongoing shortages in the supply chain. The Semi now won't come until 2023 at the earliest; Musk originally began taking orders for a scheduled 2019 delivery date.

No beta for Audi

His track record on FSD hasn't been any better. On Autonomy Day in April 2019, he famously predicted 1 million Teslas would be on the road driving without any human supervision by the end of the following year. In July 2020 he then claimed his team was on the cusp of solving Level 5 autonomy, the most advanced stage of progress that is so complicated many experts believe it could still be a decade away at the minimum. That didn't stop him from telling shareholders in January he was confident it would then come this year, only to ultimately concede this summer he "didn’t expect it to be so hard" to deliver on his promise, even though "the difficulty is obvious in retrospect."

Peeters thinks Musk's latest prediction of FSD coming to Europe next year is just another example of Elon time. “I very much doubt that will happen,” he said.

After all, the rollout is not entirely in his control. In Europe, it is the UNECE that determines the acceptable pace of deployment, not Musk, and it aims to boil all regulations down to mathematical formulas that precisely describe how a self-driving car should act in each and every situation. While this philosophy is increasingly being questioned by autonomous driving experts, it's unlikely to change dramatically any time soon, let alone next year.

“I have basically given up all hope of FSD coming to Europe in the next five years," Peeters said. "Unless you're planning to hold on to the car for a very long time, I don’t recommend anymore buying this feature until the regulators actually allow it."

That's why he wishes Tesla’s flexible new FSD subscription service would be introduced in Europe so he could switch to that before the Model S Plaid is delivered. Then he could pay a flat rate to test out the technology on a month-by-month basis only when major new features are added, and cancel at any time.

Tesla isn't the only carmaker facing Europe's regulatory rigidity. Another casualty of the UNECE’s glacial pace of work is Audi. The Volkswagen Group brand was initially prepared, in late 2017, to launch an automated pilot for highway traffic jams that would have been the world’s first unsupervised system. By the time regulators in Geneva finally agreed on how such a feature could be engineered to be sufficiently safe, it was already too late to redesign its flagship A8 luxury limousine to meet exact UNECE specifications.

In an interview in September, the brand's chief engineer said, however, he would not follow Tesla’s approach of beta testing self-driving technology just to bring it quicker to market, irrespective of whether owners signed off their rights to legal recourse. Established carmakers like his have reputations built over decades, while an upstart like Tesla that still needs to reach scale is more willing to take risks to narrow the sales gap.

“When it comes to issues of life and limb, I cannot imagine opening up a beta to our customers," Audi's Oliver Hoffmann told Fortune. "Not at all. Not even if they were willing."

Right now, Tesla buyers are in a tricky bind. On the one hand, there is no definitive sign that regulators outside the United States will adopt the same laissez-faire approach to self-driving technology and allow wide-scale customer tests of an unfinished product. On the other, Musk's customers fear he will continue to ratchet the cost of FSD ever higher over time, so people like Peeters sense the urge to lock a price in now lest they end up paying more later.

An easy solution would be allowing FSD customers to migrate their purchase to their next Tesla so long as it is not feature-complete. Currently it is paired to the vehicle, so if a customer wants to trade in for a new model, they will have to buy the optional extra all over again, likely at a higher cost. But Tesla has thus far shown no willingness to concede here, even though every car it builds comes with FSD hardware embedded—all that would be required is activating the feature via a software patch.

Now that Musk is opening up his extensive EV supercharger network to the competition in Europe, where rival owners won't need to invest in so much as a new cable thanks to standardized plugs, Peeters is considering shopping around should another brand navigate the regulatory self-driving hurdles in Europe more quickly.

"I’m not sure my next car after the Plaid will still be a Tesla," he said.