Amazon, Meta-Backed Scale AI Raises $1B, Boosting Value to $13.8B

Artificial intelligence startup Scale AI has raised $1 billion in a series F funding round to help businesses accelerate the development of AI models.

Founded in 2016, Scale labels dataused to train AI models. Organizations hire Scale to organize and improve the underlying data they plan to feed into their large language models.

Accel led the round, joined by existing investors including Nvidia, Y Combinator and Peter Thiel’s venture capital firm Founders Fund.

Meta, AMD, Amazon and the investment businesses of Cisco, Intel and ServiceNow were among the new investors that joined the round.

The latest funding brings Scale AI’s total capital raised to $1.6 billion and raises its valuation to $13.8 billion.

Scale AI CEO Alexandr Wang said the startup plans to use the funds to “move into the next phase” of its journey toward building a “data foundry for AI.”

“As we develop progressively more powerful AI, we must build frontier data which is always pushing the boundaries of AI capabilities towards complex reasoning, agents, multimodality and more,” Wang said. “Accelerating the abundance of frontier data will pave our road to artificial general intelligence (AGI).”

Scale AI has worked on several large-scale AI projects, including a $250 million deal with the Department of Defense to develop tools to evaluate AI and machine learning systems.

Related:AI News Roundup: OpenAI Picks Scale to Fine-tune GPT 3.5

The startup was also tasked by OpenAI to help fine-tune GPT-3.5, the original base model behind ChatGPT.

AI models are only as good as the data that goes into them. Scale AI staff use its platform to analyze data, creating labels that large language models can use to better understand the information they’re trained on.

The company analyzed and labeled relevant data which helped to improve the performance of OpenAI’s finetuned GPT 3.5 while also lowering the costs required to run the model.

Scale AI’s platform was also used by Meta to help perform similar fine-tuning tasks during the development of the Llama 2 large language model.

Its data tools and Enterprise Generative AI Platform have since been made available on Azure, providing Microsoft’s cloud customers tools to customize and fine-tune their large language models.

“Over the past eight years, Scale has powered nearly every major breakthrough field of AI,” Wang said. “Today, Scale supplies data to power nearly every leading AI model.”

The company recently expanded into Europe, opening an office in London as its base of operations.