Image credits: Ghost autonomy
Ghost Autonomy, a startup that develops self-driving software for automaker partners, has shut down, TechCrunch has learned.
The startup, which has raised about $220 million, announced on its website that it will cease global operations and dissolve the company as of Wednesday. The company employed approximately 100 people and had operations in Mountain View, Dallas, and Sydney.
“We are proud of the significant innovations and advancements the Ghost team has made in line with our mission to deliver software-defined consumer autonomy,” the company’s website states. “Given the current funding environment and the long-term investments required to develop and commercialize autonomous driving, the path to long-term profitability was uncertain. We are exploring the possibilities of
The shutdown comes just five months after the company partnered with OpenAI through the OpenAI Startup Fund and gained early access to OpenAI systems and Azure resources from Microsoft. Ghost also received a $5 million investment from OpenAI. The company just completed a $55 million down round last year, which included early investors Keith Lavoie of Founders Fund and Mike Speiser of Sutter Hill Ventures.
At the time, Ghost co-founder and CEO John Hayes was exploring the application of multimodal large-scale language models (LLMs) – AI models that can understand text as well as images – to self-driving cars. The company was touting its plans to consider. He argued that LLM offers a new way to understand the “long tail” and adds rationale to complex scenes where current models fall short. Experts were skeptical of this approach.
Like many startups trying to commercialize self-driving car technology, Ghost has changed its approach over the years. The startup, initially called Ghost Locomotion, was founded in 2017. The company made its public debut two years later with a total investment of $63.7 million from Founders Fund’s Rabois, Khosla Ventures’ Vinod Khosla, Sutter Hill Ventures’ Speiser and others. The company plans to develop a kit that allows private passenger cars to drive autonomously on expressways. The company said it would make the technology available in 2020.
After that deadline passed, Ghost raised an additional $100 million in 2021 with plans to focus on collision avoidance technology. The Series D funding round was led by his Sutter Hill Ventures, with participation from Founders Fund and Coatue. Hayes told TechCrunch in 2021 that the company hasn’t completely closed the door on consumer kit models, but is looking at general-purpose collision avoidance technology to get to market faster.
His premise was that self-driving systems do not need to recognize and classify objects before avoiding them. Instead, the company was tracking the movement of clusters of pixels within a scene. Most other autonomous systems start by identifying an object and then use image localization to determine its size, distance, and other relevant characteristics. This strategy is used because different objects can behave differently even if they are the same size.
Reached by email Wednesday, Hayes said the company is perfecting its product for highway use and moving in urban environments through what he called “last-mile delivery.”
“Ultimately, we were unable to finance the years needed to bring the product to market,” he wrote.