North American asset management leader Brookfield Asset Management Ltd (BAM.US) has finalized the acquisition of cloud computing firm Ori Industries, wagering that government entities and global technology companies will require massive-scale access to core AI computing infrastructure—including AI chips and power resources—to prevail in the intensifying artificial intelligence race. The cloud startup, previously backed by Saudi Aramco's venture arm, has been integrated into Radiant, a newly established Brookfield entity designed to provide on-demand access to AI computing infrastructure. This includes offering cloud-level AI training and inference capabilities based on token usage or contractual agreements, as well as leasing entire AI server clusters alongside power systems and data centers.
Brookfield announced the transaction in a Tuesday statement without disclosing financial terms. Radiant’s core business model resembles an AI infrastructure revenue stream centered on “AI GPU/AI ASIC chip leasing plus locked-in computing contracts.” The company is positioned to deliver full-stack AI training and inference services, often referred to as AI factories, with the capability to serve regulated computing environments such as sovereign clouds. By transforming “GPU/AI chips” from a one-time heavy capital expenditure into an on-demand leasable service with long-term contracts, Brookfield is incorporating this initiative into a broader AI infrastructure fundraising framework. Starting with equity commitments, the firm plans to layer co-investment and debt leverage to scale the platform into a sizable investable and developable portfolio, targeting returns above those of its traditional flagship infrastructure funds to match the unprecedented early-stage volatility of AI industrialization.
From a supply and engineering perspective, the model’s essence is not merely selling GPUs but bundling critical elements required for computing delivery into an integrated solution: chips (computing power), data center space (including cooling and networking), power (grid connection and backup supply), and contractual service-level and compliance agreements. Brookfield is deepening its investments across both digital infrastructure and core power assets, explicitly targeting AI infrastructure as a full-value-chain investment theme. For instance, its energy collaboration with Bloom Energy addresses one of the most critical bottlenecks for AI data centers—power availability and connection speed—using more controllable energy solutions to enhance the reliability and scalability of “compute-as-a-service” delivery.
Brookfield anticipates that global government and corporate efforts to advance AI technology and next-generation large language models will necessitate substantial capital investment. The firm’s AI infrastructure fund is actively seeking investor commitments, aiming for returns that exceed those of its flagship infrastructure offerings. While many infrastructure investors have shied away from major “GPU-as-a-service” commitments due to concerns over potential obsolescence of AI chips, Brookfield has chosen to double down, having already placed significant bets on power utilities and AI data center development. The company aims to meet customer demand for cost reduction and avoidance of lengthy chip supply delays.
Radiant will also collaborate with corporations and government agencies seeking highly controlled computing environments, or “sovereign clouds,” where data cannot cross national borders. Brookfield is betting that clearly structured contracts will lock in rental revenue over an estimated five-year chip lifespan or optimized depreciation cycle, thereby insulating the business from losses. “We view this as an infrastructure leasing business,” said Sikander Rashid, Brookfield’s Head of AI Infrastructure, in an interview. “We aim to be among the first movers to unlock large-scale capital for this theme.” The asset manager’s contracts are designed so that Radiant’s clients—expected to be investment-grade—must pay in full even if they no longer require the chips. “We are not taking any technology risk in this business,” Rashid emphasized.
Brookfield estimates that achieving global AI proliferation will require $7 trillion in capital investment, with $3 trillion directed toward AI computing infrastructure. The recent rollout of Google’s Gemini3 AI application ecosystem has driven a surge in global AI computing demand, prompting Google to curtail free access to Gemini 3 Pro and Nano Banana Pro and impose temporary restrictions on Pro subscribers. Meanwhile, strong trade export data from South Korea reflecting explosive demand for SK Hynix and Samsung Electronics’ HBM memory systems and enterprise SSDs further supports Wall Street’s view that the AI boom remains in an early stage of infrastructure undersupply.
A recent Bank of America research report notes that the global AI arms race is still in its “early to middle stages.” Vanguard, one of the world’s largest asset managers, indicated in a study that the AI investment cycle may only be 30% to 40% complete relative to its eventual peak, though it also warned that risks of a pullback in big tech stocks are rising. According to Wall Street firms including Morgan Stanley, Citi, Loop Capital, and Wedbush, the global AI infrastructure investment wave—centered on computing hardware—is far from over and has only just begun. Driven by unprecedented demand for AI inference computing, this investment cycle could reach $3 trillion to $4 trillion by 2030.
Radiant represents one of the first major investments from Brookfield’s multi-billion-dollar AI infrastructure fund. The asset manager is seeking $10 billion in investor commitments and expects additional co-investment and debt financing to expand the total program to approximately $100 billion. The AI flagship fund targets returns higher than those of Brookfield’s flagship infrastructure fund, reflecting the inherently higher risk-reward profile of financing greenfield projects for the emerging AI industry.
Under CEO Connor Teskey’s leadership, Brookfield has joined other top asset managers such as Blackstone and BlackRock in the race to finance AI-driven data center power equipment, core hardware, and infrastructure components. Their substantial capital is reshaping the industry. Brookfield has already invested in data center operators that lease AI infrastructure to tech giants and acquired stakes in power utility companies benefiting from soaring electricity demand. The firm’s AI fund is investing up to $5 billion in Bloom Energy Corp. and plans to deploy the company’s fuel cells in several large AI data centers.
“AI chip leader” Nvidia (NVDA.US) maintains multiple business ties with Brookfield, underscoring the tech giant’s deepening integration with major financial players. Nvidia participated in the initial $5 billion equity raise for Brookfield’s AI fund and serves as a core AI chip supplier to Radiant. The arrangement is expected to include Nvidia’s upcoming Rubin-tier AI chips, which are designed for agentic AI systems requiring less human supervision.
Brookfield’s plan is to activate investments following proposals from its utility company clients. The asset manager intends to bundle chip access with data center space and power contracts. “We are trying to offer a complete, integrated AI computing infrastructure solution for various stakeholders,” Rashid stated. The AI fund also forms part of Brookfield’s strategy to deepen relationships with high-net-worth investors in the Middle East. Although Radiant will initially focus on Europe, the company plans to lease AI chips within a large AI data center campus in Qatar, a project Brookfield is developing in partnership with QIA subsidiary QAI. Ori’s investors, including Saudi Aramco’s Wa’ed Ventures, are expected to retain stakes in Radiant.
Comments