Rumble Q2 Earnings Call Highlights

Rumble’s Northern Data acquisition reshapes the company around AI infrastructure. Its new Quake AI unit combines Rumble Cloud with approximately 22,000 NVIDIA H100 and H200 GPUs, which were operating at more than 85% utilization. Second-quarter revenue rose 61% to $40.4 million, while adjusted EBITDA loss narrowed to $16.6 million. Net loss widened to $80.3 million, primarily due to $28.3 million in acquisition-related costs and higher depreciation and amortization. RUM Group issued third-quarter revenue guidance of $87 million to $93 million, its first full-quarter forecast including Quake AI. Management expects to report separate results for Rumble Video and Quake AI going forward. Rumble NASDAQ: RUM reported second-quarter revenue of $40.4 million, up 61% from $25.1 million a year earlier, as the company completed its acquisition of Northern Data and reorganized under the new parent-company name RUM Group Inc. The closing of the Northern Data transaction marks the company’s entry into cloud and agentic AI infrastructure. The combined company will operate two business units: the Rumble video platform and Quake AI, which combines Rumble Cloud with Northern Data’s GPU fleet. Quake AI has roughly 22,000 NVIDIA H100 and H200 GPUs. The GPU estate was operating at more than 85% utilization, reflecting improvements in customer support, software and infrastructure-as-a-service execution. Quake AI is the company’s future financial engine, positioning RUM Group as an end-to-end AI infrastructure provider that combines Rumble’s existing compute, content-delivery network and streaming infrastructure with Northern Data’s AI compute operations. In June, the company signed a multiyear agreement with Together AI to deploy NVIDIA HGX B300 GPU capacity. The agreement establishes RUM Group as an independent provider of large-scale AI infrastructure outside the traditional hyperscaler ecosystem. The company is targeting the monetization of 250 megawatts of power capacity in 2027. Its portfolio includes a 180-megawatt powered site near Atlanta, where the substation has been built and transformers are on site, as well as a smaller Pittsburgh location and European sites with roughly 50 megawatts in Sweden and 20 megawatts in Norway. Monetizing the currently unmonetized 250 megawatts could represent a $3 billion-plus annual revenue run-rate opportunity, though that figure reflects the company’s view of the potential opportunity rather than guidance. Management expects demand for AI compute to remain ahead of available capacity over the next one to two years, citing growth in inference and agentic AI. The company is focused on AI compute as a service rather than a “powered shell” approach, in which a company provides a data-center facility and power while customers furnish their own computing equipment. RUM Group does not intend to enter customer contracts unless it has strong visibility into obtaining the hardware and capacity needed to fulfill them. The company recorded nearly $47 million in investing activities during the quarter, largely related to IT capital expenditures needed for its AI compute-as-a-service operations.