TL;DR — Key Takeaways
- Anthropic has reportedly signed a $13.7 billion, six-year cloud computing agreement with Rum Group, adding another major source of infrastructure capacity.
- Rum Group plans to provide Anthropic with computing capacity from a new data center in Georgia, though financing and construction remain key uncertainties.
- The deal adds to Anthropic’s aggressive infrastructure expansion as demand rises for products including Claude Code and Cowork.
Anthropic has reportedly entered a $13.7 billion, six-year cloud computing agreement with Rum Group, the social media and artificial intelligence (AI) infrastructure firm behind the conservative video platform Rumble.
The deal, reported Monday by The Information, underscores Anthropic’s rapidly expanding compute appetite following surging demand for its Claude Code and Cowork offerings.
Rum Group shares surged as much as 28% in premarket trading on the news. Rum Group had previously disclosed the multibillion-dollar contract in an August regulatory filing, though it left the identity of the customer unnamed at the time.
Formerly known as Rumble, the company rebranded as Rum Group in June after acquiring German neocloud provider Northern Data. The platform — backed by early investors Peter Thiel and Vice President J.D. Vance — hosts President Donald Trump’s Truth Social and maintains strong political ties to the conservative ecosystem.
Under terms of the agreement, Anthropic will lease computing capacity at a newly constructed data center in Georgia. However, key details surrounding the facility’s completion remain uncertain. Northern Data previously projected the site would open late this year, but Rum Group disclosed in August that it lacked the necessary financing to complete construction and acquire equipment. A source familiar with the situation indicated that Rum Group intends to rely on debt financing to complete the project.
The agreement marks the latest in a series of massive infrastructure plays by Anthropic.
Over the past year, the company has forged major computing deals with tech giants and specialized neoclouds, including Google, SpaceX, and Nscale. Excluding the new Rum Group partnership, Anthropic has committed to at least 14.8 gigawatts of compute capacity, representing an estimated $517 billion in infrastructure expenditure over the next decade.
The breakneck expansion comes as Anthropic CEO Dario Amodei advocates for stricter industry oversight. Amodei recently proposed placing independent evaluators directly inside leading AI developments to monitor risks, a move designed to temper the industry’s rush toward increasingly potent AI models.
However, industry analysts warn that self-regulated safety efforts may fall short.
Critics note that because AI firms directly invite, compensate, and retain redaction rights over these external observers, evaluators risk functioning as mere consultants rather than robust checks on power. Experts argue that effective governance will instead require unified accreditation standards and direct reporting lines to federal regulators.
