Teleo bills itself as a construction robot startup, but its mission is bigger than automating heavy equipment like excavators and tractors. Teleo's refurbished machinery now allows customers to operate their existing vehicles semi-autonomously. In the future, the startup believes that the data it collects will be a key factor in helping the robotics industry reach its “ChatGPT moment.”
It's not a desire to reach the same level of hype surrounding ChatGPT. Instead, Teleo CEO Vinay Shet wants to make a big innovative leap in the field of robotics by using a robotics company, the same one used to build ChatGPT. We believe there is an opportunity to collect huge amounts of data sets.
And investors seem eager to help the startup reach that milestone. TechCrunch has learned that Teleo recently raised $16.2 million in funding through two extensions to its Series A round in 2022. A $9.2 million expansion was completed in April, and an additional $7 million expansion was completed this week, according to recent filings and information from the company.
“The underlying model that led to the creation of ChatGPT relied heavily on data representing trillions of tokens that are effectively freely available on the internet, languages, videos, images, etc. It doesn’t exist,” Shet told TechCrunch. “In the world of robotics, the best dataset we know of is about 2.4 million tokens, but in the world of languages, datasets are trained using trillions of tokens.”
Teleo aims to fill that gap by logging data from the company's day-to-day operations, which will ultimately lead to generalized intelligence that can be used to train “true robot-based models.” “It will be the foundation,” Sheto said.
To build a data repertoire, Teleo must be deployed quickly and at scale across multiple industries. And the company's strategy for achieving that boils down to a semi-autonomous approach. Teleo can be retrofitted with any equipment with the necessary autonomous driving software and sensors, including cameras, lidar, and radar, to drive itself under limited conditions. A remote human operator can then step in and perform more complex tasks, such as unloading a dump truck, typically handling multiple vehicles at once.
“This combination allows us to solve the entire customer use case while [return on investment] We deliver to our customers and generate revenue through standalone products,” said Shet.
To maintain a diverse data set, Teleo has recently expanded beyond the construction industry, adding equipment such as wheel loaders, terminal tractors, and excavators to a wide range of industries including pulp and paper, logging, port logistics, agriculture, and munitions. We are introducing autonomous heavy equipment. removed. Teleo is also targeting industries such as airports, waste and recycling, logistics, and snow removal.
The hope is that the data Teleo collects, including input from human operators, video footage, and sensor feedback, will be used to fine-tune and specialize basic robotics models. Masu. This could eventually allow humans in the loop to be replaced or augmented by cloud-based AI agents that can learn how to control various machines in the same way humans do.
This long-term thinking is undoubtedly what attracted investors to the company. Teleo's recent expansion was led by UP.Partners, with participation from new investors Trousdale Ventures and Triatomic Capital, as well as returning investors F-Prime Capital and Trucks VC.
Teleo said the funding will be used to enhance the startup's AI capabilities, including increasing customer adoption, continuing to expand into new industries, and integrating language models at scale to increase operator efficiency. is.
“Over the next few years, you're going to see vertically integrated companies like ours actually deploy into the real world in a way that makes economic sense and grow economically based on that.” Schette said. “But along the way, they will collect enough data in the right format to reveal that ‘aha’ moment in a few years.”