Can Infrastructure Keep Pace With AI?

Executives from Tillman Global Holdings, NTT DATA, Infosys and Brookfield Asset Management discuss the physical constraints of scaling AI at the 2026 Bloomberg New Economy Forum in New Delhi. The article provides no specific figures, company developments or market reactions.

Analysis

The investable signal is a potential shift in AI bottlenecks: if power availability, grid connections, cooling, or permitting constrain deployments, value may accrue less to AI software in the near term and more to owners and builders of enabling infrastructure. That is a conditional mechanism, not evidence from this panel of contracted demand or near-term earnings upside.

For Brookfield Asset Management, the second-order opportunity is the possibility of more capital deployment into power and digital infrastructure; the counterweight is that long development timelines, permitting, and construction costs can delay fee generation and returns. Verify actual fund commitments, deployment pace, and project economics before underwriting earnings impact. For Infosys, physical constraints could defer some client AI rollouts, but could also sustain advisory and integration work; this discussion does not establish a change in bookings or revenue outlook.

The time horizon is structural (6–18 months), not a standalone near-term catalyst. The contrarian point: infrastructure scarcity can be bullish for asset owners but not automatically for asset managers—returns depend on entry prices, execution, and whether power and grid capacity arrive on schedule. No clear trade follows from the event alone.

Key Decisions for Investors

• No event-driven position in BAM or INFY on this discussion alone; treat it as a thesis prompt, not new earnings evidence.
• Put BAM on a conditional watchlist for evidence of AI-linked infrastructure fundraising, deployment, and contracted project economics. Reassess if deployment is delayed by permitting or grid access, or if reported returns weaken.
• For INFY, monitor large-deal bookings and management commentary on AI implementation timelines. A sustained slowdown in AI-related client programs would challenge the view that services work offsets delayed deployments.
• Track power availability, grid-connection queues, cooling capacity, and data-center project delays over the next 1–3 months; improving access would weaken the scarcity thesis, while persistent bottlenecks would support further diligence on infrastructure beneficiaries.