Anthropic is reportedly looking to strengthen its grasp on AI-fueling technical infrastructure. The AI company is reportedly seeking to lease and manage data centres across the U.S. to get a hold onto a capacity of one gigawatt (GW). With this, Anthropic aims to power its research and development of advanced AI technology.
Anthropic is presently discussing a structured financial framework with long-time backer Google, The Information reported citing sources familiar with the ongoing discussions.
If the deal goes through, Google would serve as a financial guranteer to back Anthropic’s hefty lease payments. The search engine giant co-designs server chips which could be deployed inside these new data centre facilities.
For now, neither Google nor Anthropic have confirmed the reports.
The infrastructure push comes as the startup actively gears up to go public in the U.S. this year. Earlier this month, Anthropic submitted a confidential Form S-1 filing with the SEC for a preliminary review of its IPO plans.
For now, the company’s ultimate IPO funding goal remains undisclosed. It did, however, report a staggering annual revenue that skyrocketed from $10 billion to $47 billion over the past year.
With its IPO plans simmering in the background, the company has unveiled the next iteration of its proprietary large language model (LLM) — the Claude Fable 5 this week. After calling for a halt in the creation of new AI models for a while, Anthropic’s revamped Claude model has been launched with advanced vision and coding capabilities.
With AI showing signs of rapid advancement and legacy big tech players like Meta and Google going bullish on the technology, the race to get control over large capacity data centres among AI firms seems to be getting intense.
Just this week, Meta announced a partnership with India’s Reliance under which billionaire Mukesh Ambani’s company would build a massive AI data centre in the Jamnagar city of India’s Gujarat.
