Turing Enterprises Inc., a provider of training data for artificial intelligence models, today announced that it has closed a late-stage $111 million funding round.

Khazanah Nasional Berhad, Malaysia’s sovereign wealth fund, led the Series E raise with participation from more than a half-dozen other backers. Turing is now valued at $2.2 billion.

Palo Alto, California-based Turing provides a service that enables AI providers to commission custom training datasets. At the start of a project, the company plans details such as the volume and complexity of data points an AI project requires. Turing then produces the records with help from a network of more than 4 million developers and other professionals.

The company supplies data for several types of AI projects.

Researchers can use Turing-provided datasets in projects that use RLHF, or reinforcement learning from human feedback. It’s an AI development method that involves having human experts guide the model training process. According to Turing, its datasets are also useful for direct preference optimization. This is the process of more closely aligning an existing neural network with customer preferences.

Turing promises to ease other AI development tasks as well. An offering called LLM Lab provides prepackaged software modules for common AI use cases. It includes, among others, AI agents that can generate coding advice and automate some customer support tasks.

Another offering, ALAN, promises to speed up the process of fine-tuning LLMs for specific tasks. Companies can also use it to measure how fine-tuning a model increases its performance.

Turing provides its data and AI development products alongside a broad suite of professional services. The company helps customers with tasks such as building new applications, moving existing workloads to the cloud and modernizing their interfaces. Turing can likewise manage ongoing application maintenance on enterprises’ behalf.

The company’s customer base includes OpenAI, Anthropic PBC, Google LLC and other major players in the generative AI market. Turing says that its annualized recurring revenue has tripled over the past year to $300 million. Moreover, the company is profitable.

“We’ve built a suite of products on top of the world’s largest talent cloud to improve AI models and build applications on top of them,” said founder and Chief Executive Officer Jonathan Siddharth (pictured).

Turing will invest the proceeds from its funding round in sales and marketing initiatives. Additionally, the company plans to expand the number of development tasks that its services can speed up.

The funding round comes less than a year after Scale AI Inc., another AI training startup, raised $1 billion at a $18.8 billion valuation. Like Turing, the company works with a network of experts to create training datasets. Those experts craft annotations, explanations of the information inside a dataset that can help AI models learn more efficiently. 

Photo: Turing