Valuing AI Infrastructure Through Physical Constraints and Automation

The AI infrastructure boom and the decline of IT consulting are not separate stories. They are two sides of a shift toward automated efficiency. While the market focuses on immediate earnings, it ignores the physical risks of data center concentration and the threat AI poses to human-led advisory services. Investors who rely on historical data to value these firms miss the compounding costs of climate exposure and the rapid obsolescence of traditional consulting models. Success requires looking past the current quarter to find companies that own essential, non-replicable resources rather than those selling human labor that AI now performs for free.

The Hidden Liability of Data Center Concentration

The current rush into AI infrastructure relies on a physical build-out that assumes environmental stability. However, research indicates that nearly 90% of global data center capacity faces elevated risk from climate hazards like wildfires and flooding. This creates a feedback loop: as we demand more compute for AI, we concentrate more value in vulnerable locations.

The immediate benefit of building massive facilities, such as the $200 billion Meta project, is the rapid expansion of AI capabilities. But the downstream consequence is an insurance and operational problem. Because these facilities are designed for 20 to 30 year lifespans, they are tethered to local infrastructure, such as power grids and water supplies, that are increasingly stressed.

"Most underwriting when you're buying insurance for real estate it still uses historical data which isn't doing a good job of predicting how climate events perform today."

-- Matt Frankel

The system is currently routing around these risks by focusing on the easy wins of building cooling solutions, yet the physical constraints of water and power remain. Investors who ignore these environmental externalities are betting that historical climate patterns will hold, a premise that the data increasingly refutes.

The IT Consultancy Apocalypse: Why Human Input is Devaluing

The decline of IT consultancies like Accenture is often framed as a temporary market dip, but it represents a shift in the value of human-led advice. As AI models like ChatGPT become more adept at providing personalized, real-time guidance, the bespoke advice traditionally sold by consultants is losing its premium.

The pattern is clear: businesses that previously turned to humans for implementation strategies are now turning to AI to learn how to implement itself. This creates a death by a thousand cuts for the consultancy model. Even when these firms attempt to pivot through aggressive acquisitions, such as Accenture’s recent $4 billion investment in cybersecurity, they often overpay for non-profitable assets, further pressuring future earnings.

"I do think that personalized advice from a human person is a dying art form and for better or for worse, I would personally say probably a little bit more towards the worst but it is being replaced by AI."

-- John Quast

The conventional wisdom suggests that these stocks are value plays due to their low earnings multiples. However, this relies on the assumption that revenue will stabilize. If AI continues to erode the demand for human expertise, these companies are not cheap. They are legacy businesses facing structural obsolescence.

Competitive Advantage in the No-Go Zone

When an entire industry faces disruption, the most durable advantage goes to companies that own the picks and shovels of the new era or those that possess diversified revenue streams. Pure-play IT consultancies are currently landmines. They lack the leverage to compete with the very technology they are supposed to be implementing.

Instead, the system rewards those who control the physical constraints of the AI revolution, such as water management and power efficiency. Companies like Badger Meter or Vertiv (VRT) provide essential, non-discretionary services that the AI build-out cannot function without. Unlike the consultancies, which sell a service that AI can now replicate, these firms sell products that solve the physical bottlenecks of the AI age. This is the difference between selling a commodity that is being automated and owning the infrastructure that the automation requires.

Key Action Items

• Audit your Value holdings: Review any IT consultancy or legacy tech investments. If their core revenue relies on human-led advisory, recognize that this is a shrinking moat. (Immediate)
• Shift toward physical constraints: Look for companies managing resource scarcity, such as water or power management, rather than service-based firms. These provide the picks and shovels for the AI build-out. (Next 6-12 months)
• Re-evaluate insurance exposure: If you hold large-cap tech or REITs, investigate their data center footprint. High concentration in climate-risk zones, such as Northern Virginia or the Asia-Pacific region, is a hidden long-term liability. (Next quarter)
• Prioritize diversified tech: For exposure to AI, favor companies like IBM that have legacy infrastructure, such as mainframes or software, alongside new growth drivers like quantum computing, rather than pure-play consultancies. (12-18 months)
• Engage in regulatory oversight: Participate in the SEC comment process regarding quarterly reporting (#savethe10q). Maintaining high-frequency data is a structural advantage for individual investors against institutional players. (Deadline: July 6th)