Serve Robotics SERV has introduced Maggie, a conversational, AI-powered robot showcased at NVIDIA GTC 2026, marking a significant step beyond autonomous delivery into interactive, human-centric robotics. Unlike traditional delivery bots, Maggie is designed to engage with humans in real time, serving as a showcase for the intersection of AI, connectivity and real-world robotics.
Serve Robotics is leveraging T-Mobile $TMUS’s 5G Advanced and edge computing as the backbone of this shift, enabling ultra-low latency and real-time responses for more natural interactions. This partnership reflects a broader move toward embedding AI directly into physical environments, where high-performance connectivity is critical for instant decision-making.
The launch of Maggie builds on a period of aggressive operational scaling for Serve Robotics. The company concluded 2025 with 2,000 robots deployed across 20 cities, a footprint that drove fourth-quarter revenues to $0.9 million — representing a 400% year-over-year surge. This growth was underpinned by a 99.8% delivery success rate and a substantial increase in daily active robots, which rose from an average of 52 units in 2024 to 273 units by the end of 2025.
Serve Robotics aims to strengthen its “flywheel” by using real-world interactions to continuously improve its AI systems. Each sidewalk interaction adds unique, hard-to-replicate data, creating a compounding advantage as the fleet scales. At the same time, these conversational capabilities open new monetization opportunities in branding and advertising, a segment that grew 50% year over year in the fourth quarter of 2025 and could eventually account for up to 50% of total fleet revenues.
However, execution remains key. While Maggie proves the potential of conversational robotics, scaling it across thousands of units and turning it into meaningful revenues is still uncertain, especially as SERV remains in an investment-heavy phase with negative EBITDA and high R&D spending.