Goldman Sachs Exchanges
2026-11-02 · Hosted by Allison Nathan · Goldman Sachs
Executive Summary
In this Goldman Sachs Exchanges interview, Brookfield Asset Management CEO Connor Teskey argued that AI infrastructure's real constraint isn't capital or demand — it's execution capability: who can actually build, connect to the grid, and operate large-scale data centers on time and on budget. Teskey, who became CEO this year after a multi-year handover from Bruce Flatt, said demand for AI infrastructure and the energy that supports it is "exceptionally high" while supply is "only a small fraction of that demand," meaning even a slowdown in AI demand growth wouldn't close the imbalance.
Key Stories & Changes
1. The Real AI Bottleneck Is Execution, Not Capital or Demand
Teskey said "there is no shortage of capital and there is no shortage of demand" for AI infrastructure — the scarce resource is the ability to complete land development, secure grid connections, build power supply, and operate large-scale data centers ("gigafactories") with redundancy
Argued the market has reacted too strongly to isolated news (a hyperscaler walking away from one of "multiple thousands" of data center leases, or a more efficient model announcement) relative to the underlying demand-supply imbalance
Said even if AI demand growth were to slow or plateau, the current supply shortfall is large enough that the imbalance would persist
2. Power Is the New Scarce Commodity
Teskey said electricity, once taken for granted, is now "the single biggest bottleneck to growth" for the world's largest companies and countries
Brookfield's fastest-growing energy segment is now energy storage, which it uses to offset the intermittency of low-cost renewables
The firm increasingly sells power directly to corporates via global framework agreements rather than to utilities — a shift from a decade ago when Brookfield "essentially sold power to utilities and governments"
Said large electricity buyers increasingly favor a small number of scale providers with proven delivery track records over numerous smaller suppliers
3. "Sovereign AI" Emerging as a Distinct Demand Driver Outside the US
Teskey noted Europe lacks the US's concentration of large homegrown technology companies, so AI infrastructure demand there is increasingly driven by nation-states rather than corporates
Brookfield has signed sovereign AI infrastructure agreements in France, Sweden and Qatar, where governments are the initial off-takers rather than hyperscalers
Framed this as nations not wanting their populations "left behind" on AI productivity gains
4. Oaktree Acquisition: A Case Study in "Build vs. Buy"
Brookfield took a majority stake in Oaktree (Howard Marks's credit-focused firm) in 2019, later moving to full ownership plus stakes in other managers
Teskey said Brookfield generally prefers to build internally when it has the relevant capabilities, since it's typically cheaper and more customizable
Oaktree was different: Brookfield judged that building a market-leading credit franchise from scratch would have taken 10-15 years, so acquiring was worth accelerating the timeline
The deal originated from Teskey's early-career review of other public alternative asset managers — he and Brookfield's Bruce Flatt identified Oaktree as undervalued relative to its franchise quality in early 2018
5. CEO Transition: Three-Year Methodical Handover From Bruce Flatt
Teskey, who joined Brookfield at 24 and is now 38, became CEO of the asset management business this year after a multi-year, incremental transition from Bruce Flatt (who remains CEO of Brookfield Corporation)
Teskey credited Brookfield's culture of early mentorship and credited Flatt's advice — "don't worry whether or not you're the right person... that was our choice, not yours" — as particularly grounding
Said Brookfield deliberately mixes senior experience with "young energy" across its leadership
6. Growth Outlook for the Alternatives Industry
Teskey sees continued growth driven by two sources: rising institutional allocations (institutions at mid-teens percentage allocations to alternatives today, trending toward 20-30%) and a much larger untapped opportunity in the "individual market" (high-net-worth, 401k, insurance/annuity holders), currently allocated 2% or less to alternatives
Cited private markets' return premium, portfolio stability relative to increasingly volatile public markets, and suitability for long-build, long-duration projects (like large infrastructure) as structural growth drivers
Trends Identified
1. AI Infrastructure Investment Is Shifting From Financial to Physical Constraints
Teskey's framing — that capital and demand are abundant but execution capability (land, grid connections, power, construction) is scarce — reflects a broader shift in how sophisticated infrastructure investors think about AI risk: the bottleneck has moved from "who will pay for this" to "who can actually build it."
2. Sovereign Capital Is Becoming a Parallel AI Funding Channel
The emergence of nation-states as AI infrastructure off-takers in markets lacking homegrown tech giants (France, Sweden, Qatar) suggests AI buildout economics outside the US are increasingly being shaped by government policy and sovereign balance sheets rather than corporate CapEx cycles.
3. Electricity Is Being Repriced as a Strategic, Not Commodity, Input
Brookfield's pivot toward selling power directly to corporates under long-term framework agreements — rather than treating electrons as an undifferentiated commodity sold to utilities — signals that large AI buyers are willing to pay a premium for certainty of delivery over lowest-cost power, reshaping how energy infrastructure gets financed and built.
4. Credit Is Becoming a Core Pillar of Private Markets
Brookfield's 2019 "build vs. buy" decision to acquire Oaktree, rather than build a credit franchise organically, illustrates how credit has moved from a peripheral strategy to a core pillar alongside equity for large alternative asset managers positioning for AI-era capital needs. ---
Sentiment Analysis
Overall Market Sentiment: Constructive, Long-Horizon Bullish
Teskey's tone throughout was notably more confident and less hedged than typical market commentary this week, grounded in Brookfield's direct visibility into physical infrastructure demand rather than public-market sentiment.
Risk Factors Highlighted
Execution risk in AI infrastructure build-out: Land development, grid connection, and large-scale construction capability — not capital — are the binding constraints, and failure to deliver on time/budget is the primary risk Teskey flagged.
Power supply scarcity: Electricity generation capacity must scale three to ten times to meet AI infrastructure needs, and Teskey explicitly questioned "where are they going to get" that much additional power.
Market overreaction to isolated news: Teskey warned that equity markets react disproportionately to individual data-center lease cancellations or model-efficiency announcements relative to the underlying structural demand picture.
Discipline required in long-duration capital deployment: Teskey acknowledged these are multi-decade projects requiring "a huge amount of discipline," implying real risk of capital misallocation if execution lags.
Geographic demand asymmetry: Europe's lack of homegrown large-scale tech companies means its AI buildout depends more heavily on sovereign backstops, a less proven funding model than corporate-driven demand in the US.
This episode was covered in today's [The Market Signal — 2026-11-02](https://marketsignal.beehiiv.com/p/the-market-signal-2026-11-02), a cross-source synthesis of multiple podcast reports.