Turing, the AI data startup based in Palo Alto, has announced that its revenue surged by 300% to reach $300 million in the past year, marking a significant milestone in the company’s growth. The firm, which helps AI labs like OpenAI, Google, Anthropic, and Meta improve their models, has also achieved profitability. Turing was last valued at $1.1 billion in 2021.
As AI models advance in complexity, the demand for human trainers with specialized expertise has skyrocketed. This surge has propelled the valuation of startups such as Turing’s competitor, Scale AI, which was valued at $14 billion last year.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
Turing’s business model focuses on matching AI labs with human experts in specific fields, streamlining the process of gathering and labeling data to train models. With access to a pool of over 4 million experts, including software developers and PhD scientists, Turing provides critical services to reduce the burden on AI labs to manage hundreds of trainers.
However, the cost of this service can be significant, with each complex data annotation potentially costing hundreds of dollars. Given that advanced AI models require millions of annotations, the price tag for training can quickly escalate. For example, Meta used over 10 million human annotations to train its Llama 3 models.
As AI labs reach what is known as the “data wall”—a plateau in model performance due to the lack of more internet-based training data—companies like Turing are playing an increasingly important role in helping overcome this obstacle. Turing’s CEO, Jonathan Siddharth, emphasized that these human data companies are essential for maintaining the growth trajectory of AI models.
“Companies like Turing are helping scale AI models to compensate for the data deficit we face,” Siddharth told Reuters.
