Aktoria Robotics is pitching a network of remote workers to fill the gap between autonomous promises and deployment reality, offering round-the-clock human oversight for robots that still can't quite handle the job alone.
San Francisco-based Aktoria calls itself "Human as API for robots," and the framing isn't subtle. According to a Y Combinator Launch post the company published recently, the startup pairs teleoperation software with a geographically distributed workforce trained to intervene when autonomous systems stumble. Think of it as outsourcing the humans who keep robots running when the AI fails.
The company emerged from Y Combinator's Summer 2026 batch and launched its service on August 7, 2026. Its target customers are hardware makers and logistics operators confronting a familiar problem: their machines work most of the time, but someone still needs to step in for edge cases, equipment failures, or tasks the algorithms haven't mastered yet.
Low latency, global workforce
Aktoria's platform routes robot feeds to operators scattered across India, Mexico, and the United States. According to the startup's website, one-way latencies range from 25 milliseconds for U.S.-based workers to 60 milliseconds in Mexico and 140 milliseconds in India.
Co-founder Shreenabh Agrawal wrote on LinkedIn in July that the company had 15+ teleoperators in India running robots at four U.S. sites around the clock, achieving 158-millisecond latency in practice. It's a business model built on the premise that full autonomy remains elusive despite industry hype. "Every 'autonomous' robot you've seen this year had a human behind it," co-founder Jaskaran Singh Walia noted in the launch announcement.
Beyond immediate intervention, Aktoria packages operator sessions into training datasets. The platform captures teleoperation episodes, stereo camera feeds paired with IMU sensor data, and demonstration recordings. The company labels these R1, R2, and R3 datasets, though it hasn't specified how customers use them or whether they've proven useful for model training.
Carnegie Mellon roots, Y Combinator backing
Both founders studied at Carnegie Mellon University. Walia earned a master's degree in machine learning there and previously interned at Microsoft Research, listing roughly 20 publications across robotics and AI venues. Agrawal worked in robotics at CMU after earlier study in physics and quantum technology at the Indian Institute of Science. His bio mentions recognition from India's Prime Minister awards program, though details remain sparse.
The two-person team secured Y Combinator backing, with Diana Hu listed as the YC partner for the batch. Aktoria's website displays logos for both Y Combinator and Accel, though the company has not publicly disclosed details of any Accel investment. The startup hasn't disclosed funding amounts or valuation.
Pricing promises, competitive landscape
Aktoria told prospective customers the service would cost "at least 50% less than the cost of in-house operators," according to the YC post. The company offered to waive onboarding fees for its next 10 customers in a July LinkedIn promotion. Specific pricing tiers, service-level agreements, and named pilot customers remain under wraps.
The launch comes as robotics companies navigate widening gaps between marketing claims and operational reality. Teleoperation platforms have multiplied over the past two years. Competitors include Formant, Guident, TeleOperator AI, Proxy Robotics, and Phi Robotics, all offering variations on remote operation and human-in-the-loop control. Teleo raised a $16.2 million Series A extension in February 2024 for supervised autonomy in heavy equipment.
Whether Aktoria's geographic arbitrage model proves durable depends on factors the startup hasn't yet addressed publicly: operator turnover rates, training costs, liability frameworks for remote intervention, and whether customers will trust critical operations to workers they never meet. The company invites robotics teams to book calls at [email protected], positioning the service as a way to deploy faster without building internal teleop staff.
For now, Aktoria is betting that the robotics industry's dirty secret—that humans still run the show behind many "autonomous" machines—represents a business opportunity rather than a problem to solve.
