Q1. Could you start by giving us a brief overview of your professional background, particularly focusing on your expertise in the industry?

I am an IT Infrastructure lead with 22+ years of experience specializing in Data Center Operations and Management, Projects, and System Administration. Having hands-on experience in Core Banking Operations. Professional, creative, flexible with proven analytical skills. Diversified skills include client relations, resource planning, recruiting, project management, and administrative support. Ability to oversee and manage a variety of individuals while ensuring timely completion of project deadlines, all while remaining on or under budget.

Q2. What percentage of your 2026-27 capacity is currently under active construction with secured power permits, versus 'announced' capacity that is still awaiting financing or grid-interconnection?

While Alphabet does not publicly disclose the exact percentage split of its 2026–2027 data center pipeline due to competitive secrecy, industry data reveals the broader trend. Roughly 50% to 60% of hyperscale capacity slated for this window is under active construction with secured power. The remaining "announced" backlog faces severe grid-interconnection queues that routinely drag on for 3 to 7 years. To bypass this bottleneck, Google is aggressively funding projects with pre-secured permits, highlighted by its multi-billion-dollar acquisition of Intersect Power to lock in ready-to-build capacity.

Q3. Are the expansion pipelines of enterprises weighed toward centralized gigawatt-scale campuses (training) or regional latency-optimized hubs (inference), and how does that match current tenant pre-leasing trends?

The enterprise data center pipeline is heavily weighted toward centralized gigawatt-scale campuses to meet the massive power demands of AI training. However, a structural pivot is underway toward regional hubs to prepare for low-latency inference workloads as AI goes operational. Current tenant pre-leasing trends match this perfectly: hyperscalers are aggressively locking down entire multi-gigawatt mega-campuses years in advance, driving vacancy rates to historic lows. Meanwhile, traditional enterprises are pre-leasing high-density, 20-50MW slices in regional colocation hubs to future-proof their upcoming localized inference needs.

Q4. With electricity consumption from AI expected to nearly double by 2030, what percentage of the portfolio relies on 'behind-the-meter' (on-site) generation versus the public grid, and how can this be hedged against potential regulatory grid-disconnection?

Currently, 85-90% of data centers rely on the public grid, but 20-25% of the 2030 pipeline is shifting to "behind-the-meter" (BTM) on-site generation like natural gas turbines and fuel cells. To hedge against regulatory grid-disconnection, operators are building "islanded" microgrids co-located directly at energy sources like nuclear plants or gas-well heads. They are also implementing dynamic load-shedding systems with large-scale battery storage (BESS) to voluntarily drop grid draw during peak strain. Finally, developers use dual-interconnection across distinct utility boundaries to structurally pivot power sourcing if local regulations restrict large loads.

Q5. What percentage of the current portfolio is 'liquid-ready,' and what is the projected Capex per MW to retrofit legacy halls for high-density AI clusters?

Across major data center portfolios, approximately 15% to 25% of current capacity is natively "liquid-ready" for AI workloads. Retrofitting the remaining air-cooled legacy halls costs approximately $2 million per MW, which is roughly 80% cheaper than greenfield construction. However, these retrofits require navigaIn 'water-stressed' coastal hubs, what cooling technologies can be used to minimize water consumption, and how can this mitigate the risk of local government water-diversion mandates?ting significant physical bottlenecks, including structural floor-loading limits and the spatial footprint of Coolant Distribution Units (CDUs).

Q6. Want to look at the CapEx and PUE tradeoffs of deploying adiabatic systems over pure dry coolers?

In water-stressed coastal hubs, data centers utilize closed-loop direct-to-chip cooling, adiabatic dry coolers (which use minimal water only on peak hot days), or seawater air conditioning (SWAC) to reduce water consumption to near zero. By eliminating evaporative water loss, these technologies completely decouple the facility from the municipal water grid. This operational independence ensures that the data center remains fully operational and immune to local government water-diversion mandates or emergency drought curtailments.

Q7. If you were an investor looking at companies within the space, what critical question would you pose to their senior management?

• High Operational Costs
• Environmental Effects
• Cost Advantage