Data Center Trends in the AI Era: Joaquin Fagundo’s Perspective on the Roadmap to Exploding Compute Demand
- Joaquin Fagundo
- Sep 30
- 3 min read
Why AI Is Changing Everything
Artificial intelligence isn’t just another IT workload — it’s the single biggest force reshaping the global data center industry. As AI adoption accelerates, compute demand is skyrocketing, forcing operators to rethink design, power, cooling, and sustainability strategies.
Technology leader Joaquin Fagundo, from Parkland, FL, explains why hyperscalers, cloud providers, and enterprises are investing billions in AI-ready infrastructure and what the roadmap looks like through 2030.
AI as the Engine of Compute Growth
AI workloads differ drastically from traditional enterprise applications:
Power Density: Legacy racks drew 5–10 kW, while AI GPU racks now require 60–100 kW or more.
Networking: Massive parallel training and inference demand ultra-low-latency, high-bandwidth fabrics.
Storage: Training large models can involve petabytes of data and ultra-fast SSD/NVMe systems.
📊 Industry Forecasts:
McKinsey projects 33% annual growth in AI-ready capacity through 2030, with 70% of all demand AI-driven by then.
Goldman Sachs forecasts 165% increase in data center power demand by 2030, largely fueled by AI workloads.
“AI is not incremental, it’s exponential,” notes Joaquin Fagundo. “This shift is rewriting the rules of infrastructure investment.”
Key Data Center Trends Driving the AI Era
1. Power Density & Grid Strain
Data centers are already one of the largest power consumers globally. AI workloads are expected to double or triple power needs per facility. Operators are moving toward on-site generation, grid partnerships, and renewable PPAs to meet demand.
2. Cooling Evolution
Traditional air cooling is insufficient for dense GPU clusters. Expect rapid adoption of:
Liquid cooling loops
Immersion cooling tanks
Hybrid cooling systems blending air and liquid
3. AI-Driven Operations
Ironically, AI is also helping run data centers:
Predictive load balancing
Real-time cooling optimization
Failure prevention via machine learning
4. Sustainability at Scale
With regulators and communities watching closely, sustainability is no longer optional. Expect stronger focus on:
Carbon-free energy sourcing
Water conservation in cooling
Transparency in emissions reporting.
How Industry Giants Are Preparing
Google (Alphabet)
Maintains a fleet-wide PUE of 1.09, one of the best in the industry.
Deploys custom TPUs optimized for AI workloads.
Expands renewable sourcing to support its AI-heavy Gemini model training.
Oracle & the Stargate Project
Partnering with OpenAI and SoftBank to build five new AI campuses in the U.S. under the Stargate initiative, targeting 10 GW of compute capacity.
Oracle committed $40B in Nvidia GB200 chips and invested $1B in European AI infrastructure.
Microsoft & AWS
Both are committing multi-billion-dollar expansions for AI-ready regions.
Investing in nuclear power partnerships, battery storage, and renewables to stabilize future AI workloads.
“These aren’t just technology bets,” Fagundo explains. “They’re shaping the digital economy for the next decade.”
Roadmap for Compute Demand: 2025–2030
Timeframe | Key Challenge | Industry Response |
2025–2027 | Bridging AI capacity gaps | Retrofits, modular AI pods, renewable sourcing |
2027–2030 | Scaling AI dominance | Multi-GW campuses, orchestration platforms, liquid cooling |
2030+ | Sustaining exponential growth | Next-gen accelerators, advanced cooling, metro AI clusters |
💡By 2030, McKinsey estimates that global data center capital expenditures will total $6.7 trillion, with AI infrastructure alone making up $5.2 trillion.
Risks Ahead for AI Data Centers
While growth is explosive, Fagundo warns of risks:
Power grid constraints and permitting delays
Water scarcity impacting cooling strategies
Volatile chip supply chains
Regulatory pressures on carbon and emissions
Unpredictable AI workload spikes
“These risks are already shaping design and site selection,” says Joaquin Fagundo from Parkland, FL.
“AI is the demand multiplier,” notes Joaquin Fagundo. “It’s reshaping how operators think about density, efficiency, and resilience.”
Closing Perspective
For Joaquin Fagundo, the message is clear:
“The infrastructure being built today is the backbone of tomorrow’s AI economy. Operators who adapt to power density, cooling innovation, and sustainability will define the next decade of digital growth.”
From Parkland, FL to Silicon Valley, the future of compute demand is already here, and data centers are at its core.