Connor Teskey, chief executive of Brookfield Asset Management, told Semafor that the scale of the AI infrastructure buildout is large enough for multiple investors to participate rather than forcing exclusivity. Brookfield estimates the AI buildout will require about $7 trillion over the next decade, a volume that is prompting competitors to cooperate.
Brookfield has a joint venture with Nvidia to acquire up to $100 billion of AI infrastructure assets. It is also part of a chipmaker‑arranged consortium that aims to marshal another $500 billion alongside Apollo, Blackstone and KKR — firms that are among Brookfield’s fiercest competitors in other areas.
Teskey argued that data centers, the hardware within them and the power plants that feed them play a similar economic role to bridges or ports: unglamorous but fundamental infrastructure. The main difference today, he said, is that the scale is much larger and the pace much faster.
Energy as the primary bottleneck
In a Q&A with Tim McDonnell and Liz Hoffman, Teskey was emphatic that power remains the leading bottleneck for digital infrastructure: "By a very large margin." He expects the imbalance and capital requirements for energy to persist through the end of this decade and beyond.
Teskey identified the fastest‑growing energy technologies as fuel cells (fast to deploy), onshore renewables (the cheapest form of bulk electricity), batteries and energy storage (which stabilize increasingly volatile grids), and nuclear (which provides large‑scale baseload and energy security). Regarding gas, he noted that while it will be important, new gas turbines are not a quick fix: "try and order a new gas turbine today, it’s not coming in a decade that has a ‘two’ on the front of it." That constraint helps explain stronger near‑term growth in solar and batteries.
Returns, competition and deal‑sharing
Teskey described AI infrastructure following a traditional financing trajectory: early investment by independent capital providers, then bank syndication of financing and entry by other capital sources such as private credit and insurance. The distinguishing features now are the much larger scale and the accelerated timing.
He stressed that Brookfield does not pursue every opportunity it sees: "we’re not doing every opportunity we see. We are doing the tiniest fraction." Even so, with a selective approach the firm can deploy significant amounts of capital.
On why competitors now sit at the same table, Teskey said the megatrends at play — one of the biggest energy buildouts in history and the reshaping of global trade toward resilience and local production — require far more capital than governments and public markets have traditionally supplied. "There’s enough to go around for all of us," he said.
He also cautioned that there will be dispersion in performance across investors: some will succeed, others will underperform. One mitigating factor is that many of the contracts being signed are 20‑ to 25‑year deals with high‑quality counterparties, which should deliver an initial return on invested capital regardless of demand two or three decades out.
Is AI really infrastructure?
When asked whether AI projects are fundamentally infrastructure or merely described as such, Teskey argued that as economies digitize, data must be transported, processed and stored like any other commodity. From that perspective, data centers and related systems are the pipelines and toll roads of the digital era.
Community pushback and supply‑demand balance
Addressing public resistance to data centers, Teskey warned against narratives that portray AI infrastructure demand or its supporting energy and supply chains as fragile. He said that even modest changes in model efficiency would not eliminate the current supply‑demand imbalance, and that existing supply capacity is far from sufficient for projected needs.
Conclusion
Brookfield’s partnerships — including the Nvidia joint venture and participation in a large chipmaker‑led consortium — illustrate how the magnitude of AI infrastructure needs is encouraging cooperation among traditional competitors. Energy supply will remain a central constraint, and Brookfield favors quick‑deploying and cost‑effective technologies such as renewables, batteries and fuel cells, while viewing gas as limited by deployment timelines.
