Kerrisdale Capital dropped the hammer on Aurora Innovation (AUR) Tuesday, unveiling a short position and declaring the autonomous trucking company “promises investors a business model its technology cannot deliver.”

At the heart of Kerrisdale’s thesis is the belief that Aurora’s much-hyped hub-and-spoke model doesn’t stack up.

The setup — where trucks run between centralized hubs before fanning out to multiple destinations — looks neat on paper but, the firm argues, leaves little profit left on the table.

“In the end, the market for hub-and-spoke autonomous trucking is small, the required investment is huge, and the profit pool will have to be shared with truck and component manufacturers, competitors, and service providers such as terminal operators,” Kerrisdale wrote.

The firm also took aim at drayage, which are short-haul trips needed at either end of the hub-and-spoke journey.

Kerrisdale says it’s “expensive and time-consuming” and will make autonomous trucking slower, pricier, and less reliable than traditional point-to-point manned trucking for hauls under 1,500 miles.

For its part, Aurora continues to pitch a vision of driverless trucks crisscrossing the country. But to get there, Kerrisdale notes, its tech would need to master “hundreds of thousands of suburban roads.”

That’s something today’s AI-powered systems simply can’t do, according to the firm.

Then there are costs. While Aurora insists its driverless rigs will be priced on par with manned ones, Kerrisdale’s checks with OEM partners suggest they’ll be at least 50% more expensive.

Aurora stock dropped 7.1% on the report.

Doubts on scale

Aurora launched the first commercial self-driving trucking service on public U.S. roads in May, running deliveries between Dallas and Houston with no human behind the wheel. But Kerrisdale was unimpressed.

It contrasted Aurora’s “big accomplishment” — a 200-mile Texas route — with Waymo running robotaxis across six regions, spanning 600 square miles and 15,000 miles of urban roads.

By comparison, Kerrisdale joked, Aurora looks like it’s crawling.

“There are 50,000 interstate highway miles across America and another 20,000 of state-level highways and freeways,” the firm wrote.

“If it’s going to take a year for every incremental 450 miles, driverless trucking is unlikely to make a dent across America anytime in our investing horizon.”

Aurora’s Dallas–Houston run initially sent the stock up 13.5%. But the rally didn’t last. Days later, Uber — then Aurora’s largest shareholder — announced a $1 billion Exchangeable Senior Notes offering that could convert into Aurora stock.

That raised fears Uber was cashing out. CEO Dara Khosrowshahi later reassured investors that Uber would remain “a holder of Aurora shares for the foreseeable future,” according to Bloomberg.

Whether that backing is enough to stabilize the story is another question.

Kerrisdale expects the company “will never become a viable commercial operation” and warns investors should “expect a decade of continuous dilution before arriving at a dead end.”