On the final day of 2025, Brookfield Asset Management, headquartered in Canada, made a bold announcement: it would independently venture into the cloud computing arena, directly leasing high-end AI chips housed in its data centers to developers via a newly minted platform.

This 'old economy' titan, traditionally entrenched in tangible assets such as power plants and real estate, has embarked on a transformative journey, repurposing its extensive infrastructure network into computing power services tailored for the AI era. This move positions it as a direct contender against cloud service behemoths like Amazon AWS and Microsoft Azure.

Fueling this ambitious endeavor is a grandiose $100 billion global AI infrastructure vision, bolstered by a robust $10 billion dedicated fund supported by industry giants, including NVIDIA.

Brookfield's foray not only reshapes the competitive landscape of the cloud computing market but also heralds new paradigms in slashing AI development costs and revolutionizing access to computing power.

Brookfield's strategic pivot is anchored in a core conviction: artificial intelligence presents a once-in-a-century opportunity, yet its large-scale proliferation is hampered by a dearth of physical infrastructure. The company projects that the global AI infrastructure will necessitate a staggering $7 trillion in investment over the next decade.

In November 2025, Brookfield Asset Management unveiled its inaugural Brookfield AI Infrastructure Fund, aiming to amass $10 billion in capital. The fund has already secured approximately $5 billion in initial pledges from Brookfield itself, chip behemoth NVIDIA, and the Kuwait Investment Authority, among other stakeholders. Through co-investments and financial leverage, the fund aspires to ultimately acquire and manage AI infrastructure assets valued at up to $100 billion.

Previously, akin to most investment firms, Brookfield's business model ceased at developing and holding the physical edifices of data centers, which were subsequently leased to third-party cloud service providers.

Its newly established cloud computing entity, Radiant, is set to construct an AI cloud based on NVIDIA's technological blueprint and will enjoy preferential access to data centers developed by its fund. Its most notable innovation lies in its strategy to directly lease AI chips (predominantly NVIDIA GPUs) in data centers to developers, offering more foundational computing power services.

Brookfield's confidence is underpinned by its sprawling ecosystem that spans energy and infrastructure. It already ranks as one of the world's largest owners of renewable energy assets and holds a controlling stake in nuclear equipment manufacturer Westinghouse Electric, which has inked an $80 billion pact to construct new nuclear reactors for the U.S. government, catering to AI's voracious power demands.

Furthermore, it has deployed on-site power generation solutions for data centers through a $5 billion investment in Bloom Energy's fuel cell initiatives.

Brookfield's cloud computing venture is not a reckless expansion but is imbued with a clear geographical and target customer focus. Its inaugural batch of seed projects for the AI infrastructure fund has swiftly landed in France, Qatar, and Sweden.

Europe's pressing demand for AI computing power and its technological sovereignty strategy offer Brookfield a pivotal market. The company has forged partnerships with France and Sweden to invest up to $30 billion in constructing AI infrastructure. In France, its portfolio company Data4 is erecting AI data centers, with power consumption projected to reach 1.5 million kilowatts by 2030.

Timotheus Höttges, CEO of Deutsche Telekom, has confirmed collaboration with Brookfield and submitted a joint bid to construct one of the EU's AI superfactories.

Sikander Rashid, head of Brookfield's global AI infrastructure business, unequivocally stated that its objective is to cater to government and enterprise clients who prefer to store data locally and ensure information security; this demand is particularly pronounced in Europe.

This model of directly addressing specific clients' data sovereignty and compliance needs enables it to sidestep a full-scale clash with public cloud giants and carve out advantages in niche markets.

Despite its lofty aspirations, the construction of ultra-large AI data centers confronts practical bottlenecks such as power infrastructure and land approvals. Brookfield is currently in negotiations with grid operators to secure power supply for the expansion of its data center west of Stockholm, Sweden.

Rashid underscored that its investments are not 'speculative' - implying it will not indiscriminately hoard chips in anticipation of clients - but rather aim to reduce costs through innovative business architecture designs and further propel the widespread adoption of AI as chip costs decline in the future.

References:

https://news.futunn.com/post/66841994/brookfield-a-private-equity-firm-announced-that-it-will-independently?level=1&data_ticket=1766116194545042

https://www.slrbs.com/cj/cjyw/2025-12-31/1821837.html