Serve Robotics (SERV) is having one of those days where you really can see both sides of the story. The delivery robot company got slapped with a bearish short report on Thursday, and also received an enthusiastic bullish initiation from a Wall Street analyst. Talk about mixed signals.

The Bear Case: Robots Gone Wrong

Edwin Dorsey, who writes The Bear Cave blog, published a short report calling Serve Robotics "a well-intentioned experiment with poor economics and a subpar solution for last-mile delivery." That's the polite version.

Dorsey points out that while Serve has partnerships with delivery giants like DoorDash and Uber Eats, and benefits from investor enthusiasm for anything robotics-related, the economics are rough. The company has been losing around $80 million on just $2 million in revenue over the last twelve months, according to the report.

But the really interesting stuff is about the actual robot experience. Dorsey says the robots travel on public sidewalks where people reportedly flip them over and steal the food from paying customers. He claims to have watched Serve robots in person and seen them block pedestrian traffic and frustrate people trying to get around them.

The blogger even tested the service himself and ran into problems ordering and receiving food.

When Robots Go Viral

Customer complaints online paint a picture of spilled drinks and longer wait times when Serve robots handle deliveries. There's also the annoying factor that customers have to meet the robot outside their building, while human delivery drivers typically bring orders inside to offices and apartments.

Dorsey's report includes social media evidence showing Serve robots tipped over, blocking a man on a mobility scooter, getting stuck in crosswalks, and in one memorable incident, a Serve robot getting hit by a Waymo self-driving car after running a red light at a crosswalk. It's like a blooper reel, but for autonomous delivery.

The Bull Case: Physical AI Pioneer

On the flip side, Oppenheimer analyst Colin Rusch initiated coverage of Serve Robotics with an Outperform rating and a $20 price target. That implies roughly 100% upside from current levels, which is a pretty bold call.

Rusch called Serve a "Physical AI pioneer" and highlighted the company's cost advantages and faster learning cycles compared to competitors. He also pointed to impressive third-quarter revenue growth of 210% year-over-year.

The timing isn't random. Serve now has more than 2,000 delivery robots operating across the U.S., with plans to expand into additional markets in 2026. That's real scale, even if the path to profitability remains unclear.

Market Reaction

Serve Robotics stock jumped 5.8% to $10.10 on Thursday, trading within its 52-week range of $4.66 to $24.35. Despite Thursday's gain, shares are still down 32% year-to-date in 2025.

So who's right? The short seller pointing to operational headaches and brutal unit economics, or the Wall Street analyst betting on rapid scaling and AI innovation? The answer probably depends on whether Serve can fix those pesky real-world problems before they run out of cash.