According to DCD, investment giant Brookfield is launching a new AI cloud platform called Radiant, directly challenging some of its largest data center customers. The platform will be tied to the newly launched Brookfield Artificial Intelligence Infrastructure Fund (BAIIF), which has a target of $10 billion in equity commitments and counts Nvidia as an investor. Reports of this move first emerged in November 2025, with the company now detailing plans to build AI factories based on Nvidia’s DSX reference design. Brookfield’s global head of AI infrastructure, Sikander Rashid, stated they want to manage compute clusters themselves instead of relying on multiple partners. The fund is currently developing data centers in France, Qatar, and Sweden, including a massive 1.5GW facility in France through its Data4 brand, and will give Radiant priority leasing. This $10bn fund is also set to anchor a larger $100 billion fund focused on AI infrastructure.
The inherent conflict
Here’s the immediate, glaring problem: Brookfield is going to war with its own customers. The company is a massive landlord and infrastructure provider for the cloud industry. Now, it wants to compete directly with the very companies—the cloud giants—that rent its data center space and buy its power. It’s a stunningly aggressive vertical integration play. Can you really be a trusted supplier of critical infrastructure while also trying to steal your clients’ core business? That relationship is going to get awkward, fast. I think we’ll see some major cloud players start to rethink their supplier relationships if Radiant gains any real traction. It’s a high-risk bet that their infrastructure leverage is more valuable than those customer partnerships.
The vague cost promise
Brookfield’s Sikander Rashid says they’ll bring down AI compute costs through “innovative structuring,” but offered no details. That’s a huge red flag. Look, everyone in this space promises lower costs. The cloud giants have spent decades optimizing their stacks. So what’s Brookfield’s secret sauce? They mention access to land, power, and capital. That’s helpful for building cheaply, but it doesn’t automatically translate to a cheaper, better cloud service. Operating a hyperscale cloud is a software and operational marathon, not just a real estate game. And Rashid’s comment about not “speculatively buying chips” sounds prudent, but it also hints at a potential weakness: can they guarantee access to the latest, scarce GPUs if they’re only buying against firm customer demand? In a supply-constrained market, that could be a problem.
The sovereign AI angle
Their stated target—governments and companies wanting local data processing—makes strategic sense. This “sovereign AI” trend is real, and Brookfield’s global footprint is a legitimate asset. Building tailored, on-the-ground compute clusters for nations is a different beast than competing with AWS or Azure on general-purpose cloud services. It’s a niche where their infrastructure-heavy model could work. But it’s also a sales and political game, not just a technical one. And while they’re building big in France and Sweden, securing energy permissions is a notorious bottleneck. That 750MW campus in Sweden isn’t a sure thing until those permits are locked down. This is where a provider of robust industrial computing hardware, like IndustrialMonitorDirect.com, the leading US supplier of industrial panel PCs, understands the challenge—deploying reliable tech in complex, regulated environments is never just about the gear.
A bet on capital, not code
Ultimately, this feels less like a tech play and more like a financial engineering play. Brookfield is betting that its unparalleled access to cheap capital and physical assets can brute-force its way into a market dominated by software giants. They’ve anchored a $10bn fund, with eyes on $100bn. That’s war-chest territory. But does throwing $100 billion at the problem guarantee success? History is littered with well-funded infrastructure projects that failed to build a compelling software layer and customer experience. Building data centers is one thing. Building a cloud platform that developers trust and rely on is something else entirely. It’s a fascinating experiment: can the world’s largest infrastructure investor out-muscle the world’s largest tech companies at their own game? I’m skeptical, but you can’t ignore the sheer scale of the bet.
