AI training data provider Scale AI closes $1 billion round at $13.8 billion valuation

Scale AI Inc., a startup that helps OpenAI and Nvidia Corp. train artificial intelligence models, was valued at $13.8 billion in a $1 billion funding round announced today.

Axel was the lead investor. Nvidia contributed through its venture capital arm, along with Intel Corp. and Advanced Micro Devices Inc. The chipmaker was joined by more than a dozen backers, including Amazon.com Inc., Meta Platforms Inc., Cisco Investments and ServiceNow Ventures.

Advanced AI models are trained on large datasets and can be prohibitively time-consuming for machine learning teams to assemble on their own. Scale AI offers Generative AI Data Engine, a cloud service that promises to speed up processes. This allows a developer to specify the training data needed for her AI project and a team of experts to assemble the necessary files on her behalf.

One task the platform promises to speed up is data labeling, the process of enriching individual files in a training dataset with contextual details, such as text snippets that describe what a particular file contains. Such context helps AI models more fully understand the information they are trained on, and therefore can output higher quality responses.

Scale AI's platform can be used to annotate many types of data. It supports text, images, and audio, as well as more specialized files such as sensor logs from his LIDAR hardware in self-driving cars. The graphical interface allows machine learning teams to provide instructions to the experts annotating the data on how to perform the process.

“Our vision is one of rich data with the production means to continue to scale Frontier LLM by orders of magnitude more,” said Alexandr Wang, founder and CEO of Scale AI. Masu. “We shouldn't be constrained by data to reach GPT-10.”

Scale AI also promises to ease many related tasks for AI developers. We provide services that can be used to test newly developed neural networks for vulnerabilities, hallucinations, and other issues. Scale AI says it uses a combination of automated and manual approaches to detect weaknesses in its customers' AI applications.

The company has developed a chatbot called Donovan for public sector organizations. Can answer questions about sensitive files such as intelligence reports. According to Scale AI, chatbots not only assist developers with their coding tasks, but can also be used as translation tools.

Internally, Donovan leverages large-scale language models from commercial AI providers such as OpenAI and the open source ecosystem. According to Scale AI, the chatbot configures these LLMs to include quotes in prompt responses. Customers can optionally fine-tune their models by providing them with annotated information using Scale AI's flagship training data platform.

Scale AI said it will use the $1 billion round announced today to make training data more accessible to machine learning teams. In addition, a portion of the capital will be dedicated to initiatives focused on assessing AI risks.

image: Unsplash