Rumble's AI story is no longer just a stock-market slogan. By closing Northern Data and launching Quake AI, it now has real Nvidia hardware to sell, and that makes the bet harder to dismiss.
On June 18, 2026, Rumble stopped asking investors to value it as only a video platform. The company said it would become RUM Group Inc., split itself into two main businesses, and launch Quake AI, a cloud infrastructure unit built on the Northern Data assets it had just acquired in a $767 million all-stock deal. Barron's reported that RUM shares jumped about 16% after the announcement. For one trading day, at least, the market treated the move as more than a fresh coat of AI paint.
Frankly, that is the right first reaction. You don't have to like Rumble's politics, or its long-running pitch as a conservative alternative to YouTube, to see that 22,000 Nvidia H100 and H200 chips are not imaginary. Those GPUs are spread across nine data centers, according to Barron's, and MarketWatch reported that the deal gives RUM access to up to 250 megawatts of power capacity. In a market where cloud buyers still spend months hunting for reliable compute, that is the part of the story you should pay attention to.
The timing also matters. Rumble said it had closed the Northern Data acquisition on June 17, then announced the RUM Group structure and Quake AI the next day. That is not a quiet integration. It is a public argument about what the company wants to become: one business for the video platform, another for selling AI and cloud capacity.
The Together AI agreement is the cleaner test of whether customers believe the pitch. On June 3, Rumble announced a multi-year, $270 million deal to provide dedicated GPU cloud capacity using Nvidia's liquid-cooled HGX B300 systems, part of the Blackwell generation. Together AI is not a random buyer pulled into a headline. It runs infrastructure for developers building and serving AI models, so its job is to know the difference between useful capacity and a loud announcement.
That is where Rumble's story becomes interesting for you as a reader. The company is not merely saying AI will help its existing business. It is trying to make infrastructure a business in its own right. CEO Chris Pavlovski told MarketWatch that the company had already launched its cloud business to the public in 2024 and needed infrastructure for its own operations, while also seeing a chance to compete in the market. Rumble Cloud has counted Truth Social and the Miami Dolphins among its customers, according to the same report.
Still, the hard part starts after the stock pop. Owning GPUs is not the same as running a durable cloud business. Hyperscalers like Amazon Web Services, Microsoft Azure and Google Cloud have years of customer trust, security tooling, billing systems, developer support and enterprise sales muscle. RUM has a sharper headline today, but it still has to prove it can keep utilization high, deliver capacity reliably and win customers that have no interest in the video platform at all.
Also read: Hyundai takes full control of Boston Dynamics as SoftBank exits for $325 million • Texas just rewrote the rules for connecting AI data centers to its power grid • Elastic's $85 million bet on DeductiveAI is a signal that AI-native ops tooling is now acquisition currency
There is another reason to be careful. The market has rewarded strange AI pivots before. Barron's pointed to Allbirds, the sneaker company that said in April it would move into AI compute infrastructure and later sold off the Allbirds brand as it became Smartbird. That comparison is awkward for RUM, and Pavlovski pushed back on it in MarketWatch, saying the company was not simply jumping into AI because the sector was hot.
He has a fair point, but only up to a point. Rumble did not wake up yesterday and discover data centers. It has been building cloud services, and Northern Data gives it physical assets that most AI hopefuls can only rent. But investors should not confuse possession of scarce hardware with proof of a long-term advantage. The real question is whether Quake AI can turn those chips, data centers and power rights into contracted revenue after the first wave of excitement fades.
For now, RUM has earned a more serious hearing than the usual AI rebrand. The company has named the unit, closed the deal, brought over real Nvidia capacity and landed a $270 million customer commitment from Together AI. That is enough to make the pivot credible.
It is not enough to make it proven. The next test will not be the name RUM Group or the Quake AI launch. It will be whether customers keep buying capacity when the announcement cycle moves on.
