The Infrastructure Investor Podcast

Episode: The AI Revolution: How Infrastructure is Adapting

Date: January 28, 2026

Host: Ben Paton, PEI Group
Guests:

• Stuart Upson, Co-President, Brookfield’s Infrastructure Group
• Udhay Mathaiallagun, Managing Partner & CEO, Brookfield Global Data Center Practice

Episode Overview

This episode of The Infrastructure Investor Podcast, part of the Private Markets 2030 miniseries, explores the transformative impact of artificial intelligence (AI) on infrastructure investing. Host Ben Paton interviews Brookfield executives Stuart Upson and Udhay Mathaiallagun on how the boom in AI is driving demand for advanced digital infrastructure—especially data centers and power generation—and what this means for investors, fund managers, governments, and tech giants worldwide. The conversation delves into global trends, the evolving partnership landscape, the importance of disciplined investment approaches, and the risks and opportunities presented by rapid AI development.

Key Discussion Points and Insights

1. The Early Stage of the AI Revolution

• Opening Perspective
  • Stuart Upson (01:24):
    “We’re actually only at the very early stages of AI technology, AI products, the developments and how that’s going to evolve over time... you’ve got, as far as we can tell, endless demand in terms of those improvements in use cases and improvements in the technology.”
  • AI demand is projected to grow exponentially, catalyzing improvements in chip efficiency, power generation, and infrastructure systems.

2. The Data Center Boom and Increasing Complexity

• Technical Requirements of AI Data Centers
  • Ben Paton (02:15):
    AI is driving the need for more powerful chips (GPUs), making modern data center operations far more complex.
  • Udhay Mathaiallagun (02:42):
    “These chips are located in servers that are housed in racks...because they’re really powerful chips, they produce heat...the new age data centers that accommodate AI need to basically have the ability to cool really hot racks of chip. And that involves a whole bunch of technical infrastructure being created around electricity distribution and cooling.”
  • Modern data centers require vast investments in cooling, power distribution, physical security, and supply chain partnerships.
  • Udhay Mathaiallagun (03:28):
    “We are talking about tens of billions of dollars worth of equipment going in here… it’s a step change both in terms of technical capability that’s required, but also the capital that needs to be deployed.”

3. Power as the Core Bottleneck and Opportunity

• AI’s Massive Energy Demand
  • Stuart Upson (04:29):
    Power is a “huge opportunity to invest into...It’s a big bottleneck around the world.”
  • Infrastructure investors are expanding into renewable, nuclear, and even gas to meet the acute power needs of AI.

4. Globalization of AI Infrastructure

• Not Just a US-Centric Race
  • Udhay Mathaiallagun (05:46):
    “It’s very healthy globally, the US clearly out in front...But it’s not just limited to the US...Europe has got some national champions...these conversations are not limited to just those markets. We've got a number of those happening across markets in Asia Pacific and in North America.”
  • Brookfield’s growing partnerships with governments worldwide aim to create “sovereign AI capabilities,” particularly in France, Sweden, Asia-Pacific, and North America (07:05).

5. The Partnership Imperative

• Hyperscalers and Sovereigns Working Together
  • Stuart Upson (08:49):
    “A 1 GW training center is $50 billion worth of capital, excluding the power...Why are there partnerships getting formed? It's because these are very, very hard things to do and very few people can do them.”
  • Partnerships bring capital, expertise, land, and speed to market—supremely important in the competitive AI landscape.
  • Example: Nvidia’s strategic investment and technology partnership with Brookfield (09:30).
• Udhay Mathaiallagun (10:23):
  Brookfield’s renewable power partnerships:
  • Up to 10GW with Microsoft in Europe and North America.
  • 3GW with Google for data center capacity.

6. Decarbonization and the “All-of-the-Above” Power Approach

• Stuart Upson (11:33):
  “What you’re seeing is really a move to an all of the above power solution approach...nuclear is now a really big focus energy transition business.”
• Brookfield (via Westinghouse) is building new nuclear capacity; tech majors like Microsoft and Google are reactivating mothballed US nuclear plants to supply data center clusters (12:11).
• Stuart Upson (12:35):
  “I think many of them have now recognized that nuclear and baseload power from nuclear is essential to enabling that to happen. And nuclear is obviously also a clean form of power. And so that’s a key part to solving that puzzle.”

7. The AI-Power Feedback Loop

• AI itself is expected to drive advances in energy efficiency and power management.
• Stuart Upson (13:21):
  “AI itself, the technology is the flywheel that is going to move technological advancements forward at a much faster pace... It will end up applying to the efficiency of our energy generation solutions.”

8. Is There an AI Bubble?

• Despite increasing doubts due to some disappointing financial results, both guests stress that any bubble doesn’t change the fundamental infrastructure demand.
• Stuart Upson (14:29):
  “Because something may or may not be in a bubble does not mean that that is not a valuable thing.”
• Udhay Mathaiallagun (15:13):
  “During a bubble... stuff gets built, but it gets used, it gets used... [need to be] knowledgeable and disciplined to get through this phase.”

9. Avoiding the “Stranded Asset” Trap

• The risk: rapid technical change leaving physical infrastructure obsolete.
• Udhay Mathaiallagun (17:14):
  “We’re not investing at the high risk end of technical change…data center designs instead of fresh set every 10 years...the core doesn’t change too much.”
• Stuart Upson (18:04):
  “Everything up to the compute doesn’t really have much of a technology obsolescence risk...we’re going to need more [data centers] in the future, not less.”

10. The Next Era: Embodied AI in Infrastructure

• Stuart Upson (19:14):
  “[In] the fullness of time where AI moves to is, AI becomes embodied and you end up having robotic labor that’s utilizing AI that can actually carry out manual tasks in the management process as well.”
• Brookfield’s partnership with Figure AI to pilot humanoid robots in infrastructure asset management and collect data for AI model advancement (20:26).

11. The Limitless Scope and Necessity of Digital Infrastructure

• Udhay Mathaiallagun (21:29):
  “Data has become such an integral part of everything we do… the rate at which data is getting produced and needing to be moved… is doubling every three years… There’s going to be a very, very long runway of opportunity for digital infrastructure.”
• Stuart Upson (23:13):
  “It’s the largest investment opportunity in nominal dollars that’s ever been seen... we think this spend is going to be in excess of $7 trillion across power, data centers, compute…”

Notable Quotes & Memorable Moments

| Timestamp | Speaker | Quote/Insight | |------------|---------|--------------------------------------------------------------------------------------------------------------------| | 01:24 | Upson | “We’re actually only at the very early stages of AI technology... endless demand in terms of those improvements...” | | 02:42 | Mathaiallagun | “These chips... produce heat. So... data centers... need to have the ability to cool really hot racks of chip.” | | 03:28 | Mathaiallagun | “We are talking about tens of billions of dollars worth of equipment going in here... a step change in capability…” | | 04:29 | Upson | "Power is a huge opportunity to invest into...It's a big bottleneck around the world." | | 08:49 | Upson | “A 1 GW training center is $50 billion worth of capital, excluding the power... very few people can do them.” | | 10:23 | Mathaiallagun | “We partnered with Microsoft to deliver up to 10GW of new capacity... another... 3GW... to Google...” | | 12:35 | Upson | “Many... have now recognized that nuclear and baseload power from nuclear is essential to enabling that to happen.”| | 14:29 | Upson | “Because something may or may not be in a bubble does not mean that that is not a valuable thing.” | | 15:13 | Mathaiallagun | “During a bubble...stuff gets built, but it gets used, it gets used...be disciplined to get through this phase.” | | 17:14 | Mathaiallagun | “We’re not investing at the high risk end of technical change… the core doesn’t change too much.” | | 18:04 | Upson | “Everything up to the compute doesn’t really have much of a technology obsolescence risk…” | | 19:14 | Upson | “AI becomes embodied... robotic labor that’s utilizing AI that can actually carry out manual tasks in the management process...” | | 21:29 | Mathaiallagun | “The rate at which data is getting produced and needing to be moved… is doubling every three years…” | | 23:13 | Upson | “We think the investment opportunity for AI infrastructure is absolutely enormous… in excess of $7 trillion...” |

Segment Timestamps

• 00:03 – 01:24: Introduction; importance of technology advances in private infrastructure
• 01:24 – 02:15: Upson on the early stage and future demand of AI infrastructure
• 02:15 – 03:28: Technical and financial demands of next-gen AI-ready data centers
• 04:09 – 05:46: Energy challenges and investment opportunities in meeting AI power needs
• 05:46 – 07:05: The globalization of AI infrastructure; government partnerships
• 08:05 – 10:18: Tech majors and hyperscalers forge large-scale partnerships
• 10:23 – 12:11: Renewable and nuclear power initiatives for data centers
• 13:10 – 14:29: AI growth cycles, the bubble question, and infrastructure discipline
• 15:52 – 18:04: Staying resilient to obsolescence risk; versatile infrastructure components
• 19:14 – 20:26: Embodied AI and the future of robotic automation in infrastructure
• 21:29 – 23:13: “Long runway” for digital infrastructure investment opportunities
• 23:54 – end: Closing remarks; power as the defining bottleneck and opportunity for AI

Final Thoughts

• Limitless Demand, Unprecedented Scale: AI is driving a “step change” in data center and energy infrastructure needs, with trillions in investment at stake.
• Key Risks and Mitigations: Investors must avoid speculative overbuilding and focus on disciplined, partnership-driven deployment anchored by real customer demand.
• Technology, Power, and Partnership Convergence: The AI era is marked by interdependencies among power innovation (including nuclear), global partnerships, and the physical infrastructure to support runaway digital growth.
• Looking Ahead: Robotics and embodied AI will further reshape labor and asset management, while the race to build—and power—AI has only just begun.

For infrastructure investors, the AI revolution presents both a generational opportunity and a complex, evolving challenge—one best met with technical expertise, sound risk management, and a collaborative approach.