Serve Robotics is now trading below book value.
With its market value dropping fast this year: → Jan 2026: ~$1.1B → Aug 2026: ~$490M, just before they announced they were breaking up with Uber → Nov 2026: ~$200M, below its $305M book value
So, what's going on?
Serve has built a large fleet, but revenue hasn't kept up. With two issues standing out: → Active robots: They grew 16% in Q3 to a record 917 daily active robots. Good progress, but Serve has a fleet of 2,000+ robots. → Revenue per robot: Fleet-services revenue grew only 12% in Q3 but at a slower rate than the active fleet growth. And at roughly $31 per active robot per day, revenue is still low relative to their costs.
In Q3, Serve brought in $2.8M in revenue against $10.4M in cost of revenue and a $59M net loss.
Three of their initiatives stood out: → Accelerate revenue from their delivery services: DoorDash and Grubhub are helping offset Uber's decline. Daily deliveries with these partners more than tripled from the end of Q2 to the end of Q3. → Better autonomy: helping reduce the cost of revenue by 14% and operating expenses by 6% QoQ. My guess is from decrease spend in remote supervisors. → Add new revenue streams: launched Physical AI Services, offering its connectivity and data tools to other robot makers.
My take: It’s a sensible plan, but I’m not sure it will grow revenue fast enough.
At their Q3 burn rate, their $198M in cash would last a little over a year. And with a ~$200M market cap, they don’t have much room to raise more capital. Over the next couple of quarters, these bets need to show up as real revenue growth and a lower burn. If they don’t, I expect deeper cost cuts.
They have a great team. I’ve worked with most of them, and I wouldn’t count them out.
But they need to lock in. The focus has to shift from demonstrating scale to proving the economics.