Now Loading

AI Data Startup, Turing triples Revenue to $300 Million

Turing

Palo Alto-based artificial intelligence data startup Turing announced that its revenue soared to $300 million last year, tripling from the previous year and reaching profitability. The company, which provides human trainers to AI labs, has seen rapid growth fueled by increasing demand for specialized data annotation services.

Turing, which counts OpenAI, Google, Anthropic, and Meta among its clients, was last valued at $1.1 billion in 2021. As AI models advance, they require more nuanced human expertise to improve their capabilities, a trend that has significantly boosted the valuations of AI data firms like Turing and its competitor Scale AI, which was valued at $14 billion last year.

“Companies like Turing are helping the scaling laws keep going to make up for the data deficit that we have,” said Turing CEO Jonathan Siddharth. He noted that as AI labs hit the “data wall”—a plateau in model performance due to a lack of sufficient internet training data—human trainers become indispensable in refining AI models.

Turing boasts access to a vast network of over 4 million human experts, including software developers and scientists with doctorate degrees, who help label data for AI models. This approach alleviates the burden on AI companies, allowing them to focus on model development rather than managing vast numbers of data annotators.

The costs associated with high-quality data labeling are substantial. “One complex annotation can cost hundreds of dollars,” Turing stated, emphasizing the extensive investment required to train cutting-edge AI systems. For instance, Meta utilized over 10 million human annotations while developing its Llama 3 models, according to Meta executive Joe Spisak.

Upcoming Conferences