Meta’s reportedly looking to expand its AI infrastructure even further, by investing in AI startup Scale AI, which specializes in data labeling to facilitate AI model expansion.
According to Bloomberg, Meta is in advanced talks to potentially invest up to $10 billion into the company, which Meta already works with as part of its expanded AI development.
Originally founded in 2016, Scale AI has become a key provider of qualified data, which can help to improve AI system training significantly.
Scale AI ensures that data sources are accurately labeled and annotated, essentially making sense of data input streams in order to improve the training process. According to research, this can significantly reduce AI training time, by feeding more valuable data sources to AI companies, reducing manual load on their part.
And as noted, Scale AI already works with several major AI projects, including OpenAI and Meta, to assist in their current AI training process. Meta previously invested $1 billion in Scale’s Series F funding round.
An expanded partnership could give Zuck and Co. a significant advantage, by ensuring that it has exclusive access to Scale’s evolving data classification tools, which could help to improve Meta’s AI models significantly.
This comes as Meta is also working to grow its data processing capacity, with the development of a 2 gigawatt data center, among various major infrastructure elements. Meta’s also currently holds around 350,000 Nvidia H100 chips, which power its AI projects (both OpenAI and xAI have around 300k H100s), while it’s also developing its own hardware for expanded AI development.
Essentially, Meta’s pushing hard to become the leader in the global AI race. And with expanded investment in key infrastructure tools like Scale AI, it’s well on the way to leading the market on potential AI development.
That could have major implications for Facebook, Instagram, and Meta’s broader AI tools, including in VR environments and in wearables, moving forward.
