Private Credit Meets AI: The Data Center Opportunity

In Brief:

As artificial intelligence reshapes the global economy, institutional investors face a critical question: how to capture the AI opportunity without betting on which software company will emerge victorious. Alexey Teplukhin, Managing Director at Blue Owl Capital, argues the answer lies in owning the physical infrastructure rather than picking winners—data centers representing the "airports of information" that every competitor must use. Speaking on the InsuranceAUM.com podcast with host Stewart Foley, Teplukhin makes the case that hyperscale data centers have become essential utilities underpinning modern society, with demand accelerating far faster than most investors appreciate.

Big Picture Drivers:

• Exponential data growth: More data has been created in the last 18-24 months than in all of human history combined, driven by richer media formats and AI applications

• Infrastructure scarcity: Large contiguous land parcels with sufficient power access are extraordinarily rare, creating natural barriers to entry

• Hyperscaler concentration: Microsoft, Amazon, Google, Meta, and Oracle represent the primary tenants, offering institutional-grade credit quality

• Physical constraints: Chip efficiency improvements are decelerating as manufacturing approaches atomic limits, ensuring continued infrastructure demand

Key Themes:

• Infrastructure over speculation: Blue Owl's investment thesis centers on owning the physical backbone rather than attempting to identify which AI software company will dominate—a strategy that sidesteps winner-take-all risk

• Essential utility status: Data centers have evolved from telecom connection points to mission-critical infrastructure analogous to airports and toll roads, making them effectively non-optional for modern commerce

• The Jevons Paradox at work: Historical patterns show that as technology becomes more efficient, consumption increases rather than decreases—video replaced audio, which replaced text, each requiring exponentially more data

• Power and land as moats: The multi-year process of assembling Manhattan-sized parcels and negotiating gigawatt-scale power agreements creates durable competitive advantages

Key Insights:

• Atomic limits approaching: Chip manufacturing has reached 17 atoms of width for imprinting, approaching the physical limit of one atom, which means efficiency gains are decelerating and cannot eliminate infrastructure demand

• Grid reliability is non-negotiable: Independent power sources create single points of failure, as demonstrated during Texas winter storms when data centers remained operational while standalone facilities failed

• Scale defies imagination: Blue Owl's recent Meta partnership involves a Louisiana campus the size of Manhattan drawing approximately 2 gigawatts—more power than the city of New Orleans consumes

• Winner identification is the real risk: The existential concern isn't whether data centers remain relevant, but whether investors can avoid backing the wrong software horse in a market that will inevitably consolidate

• Development expertise creates winners: The ability to assemble land, negotiate with utilities, and navigate rezoning over multi-year timelines separates successful data center developers from competitors

• Hyperscale emergence was recent: Large-scale data centers only became viable roughly 10 years ago when fiber optic infrastructure, cloud software platforms, and server technology converged

Memorable Quotes:

• "In the last 18 to 24 months we've created more data just in that period than all of humanity before it." - Alexey Teplukhin, explaining the exponential acceleration of data generation

• "The data center has replaced the airport. It's replaced these kinds of meetings and allows us to speak virtually because it's hosting the data on that server." - Alexey Teplukhin, describing how data infrastructure has transformed communication

• "We don't want to have to pick the winner to be successful." - Alexey Teplukhin, articulating Blue Owl's investment philosophy of owning infrastructure rather than betting on specific AI companies

• "There will definitely be AI companies that succeed and there will definitely be big AI companies that fail. Like I think that is something that everyone can agree on." - Alexey Teplukhin, on market rationalization ahead

• "I think the concept that people are driving cars is going to become outdated in 20 or 30 years. It's going to be a horror that there are people who are allowed to drive cars." - Alexey Teplukhin, on AI's transformative potential beyond data centers

The Wrap:

Teplukhin's perspective offers private credit and institutional investors a framework for participating in AI growth without the binary risk of picking software winners. By focusing on the physical layer—land, power, and buildings—Blue Owl positions itself to benefit regardless of which hyperscaler or AI platform ultimately dominates. The strategy echoes historical infrastructure plays: those who built railroads and telegraph lines profited whether Standard Oil or Carnegie Steel won their respective markets. For insurance asset managers and private credit allocators seeking exposure to digital transformation with real asset characteristics, data centers may represent the rare convergence of secular growth and defensive positioning.