With the acceleration of artificial intelligence workloads, advanced wireless deployments, and surging cloud demand, unprecedented pressure is being placed on global energy systems and physical infrastructure. Addressing this rapid acceleration is no longer simply about meeting current needs; this transformation has direct and profound implications for network resilience, economic competitiveness, and the long-term viability of technological progress.
Taken from the April 2026 Marconi Society Executive Institutes Forum Report, we’re highlighting how participants focused on the pressing intersection of technology and energy, surfacing infrastructure sustainability and energy transformation from an operational challenge to a critical global priority.
Energy Demand as a Strategic Constraint
Data center expansion, AI model training, and the proliferation of connected devices are driving energy consumption at a pace that existing grid infrastructure and energy policies were not designed to accommodate. Participants discussed the limits of continuing with current energy generation models and pointed to nuclear, fusion, and expanded solar as the main categories of alternatives that could support future AI growth. Across these ideas, the shared view was that sustaining AI growth will require both new energy sources and more serious planning around how generation capacity is developed and deployed.
Infrastructure Resilience and Physical Vulnerabilities
Digital resilience is only as strong as the physical systems that underpin it. Vulnerabilities in power grids, subsea cables, and national interconnection points represent compounding risks that are often addressed in isolation rather than as an integrated system. The consensus was made abundantly clear: physical and digital infrastructure must be viewed as a unified, interdependent ecosystem.
Security and resilience should be considered just as critically as efficiency. Highly centralized infrastructure can create attractive single points of failure, particularly when energy generation and computing resources are concentrated together. Participants noted that a more distributed model may potentially be better for resilience, even if it introduces tradeoffs in cost or efficiency.
Another important consideration was how efficiency gains need to happen across the full AI stack, not only at the model level. Hardware, packaging, system design, software, and operational practices were all discussed as areas where incremental improvements could produce meaningful aggregate impact. Sustainable AI growth will depend not just on building more power capacity but also on creating stronger incentives for efficiency, optimization, and resilience throughout the ecosystem.
The Role of Policy and Investment
The infrastructure and energy transformation agenda requires alignment across the private sector, government, and regulatory bodies. The pace of private investment in data centers and wireless infrastructure is outpacing public policy frameworks for energy planning, environmental impact, and grid modernization. Without proactive policy engagement, the sector risks creating infrastructure bottlenecks that could constrain technological progress.
Participants further explored the pace of AI expansion and whether the current rate of investment is sustainable. Deployment appears to be moving faster than historical computing efficiency trends, while usage remains supported by subsidized economics and large flows of capital. One concept the group suggested was “tokens per watt” as a more practical way to think about AI efficiency. Rather than measuring progress only through larger systems and higher throughput, participants reiterated the importance of understanding how much useful output can be produced per unit of energy consumed.
Ultimately, who should bear the cost of the supporting infrastructure? Participants underscored the need for AI companies and data center operators to contribute meaningfully to the costs of generation and grid upgrades associated with their growth. Grid modernization will likely require stronger public-private coordination, and future projects will depend in part on whether local communities believe the benefits and burdens are being shared fairly.
This session made clear that energy demand has become a strategic constraint for technological innovation. Data center expansion, AI model training, and the growing proliferation of connected devices are driving energy consumption beyond what existing grid infrastructure and energy policies were designed to support. As participants noted, to sustain our digital future, industry must look beyond current energy generation models and seriously consider transformative alternatives—including nuclear, fusion, and expanded solar capacity—to ensure infrastructure can keep pace with the demands of an increasingly connected and AI-driven world.
This is the work the AI Institute was created to advance. At the 2026 Marconi Awards Gala & Institute Forums this November 4-6 in San Francisco, global leaders from industry, academia, government, and civil society will convene to translate these complexities into concrete frameworks, partnerships, and standards while recognizing leaders championing innovative solutions.
