Samanth Subramanian on the Undersea Cables That Keep the Internet Alive

Electricity demand is surging as data centers, new manufacturing, and broader electrification drive a wave of investment into the power sector. The article highlights a structural boost for energy infrastructure and related suppliers, including switches and cabling used in communications rooms. The tone is positive for power and infrastructure spending, though the piece is largely descriptive and contains no company-specific financial figures.

Analysis

The important second-order effect is not just higher electricity capex, but a longer-duration bid for the entire “power bottleneck” stack: grid gear, switchgear, transformers, conductors, and backup generation. The market tends to underappreciate how quickly data-center interconnection queues and load growth translate into real procurement, which means the winners are often the picks-and-shovels industrials with pricing power rather than the obvious software beneficiaries. In this phase, order books can outrun earnings by 2-4 quarters, so the setup is more about backlog visibility than near-term margin expansion.

The more interesting loser is not utilities broadly, but any business with electricity as a meaningful input and weak pass-through economics — especially energy-intensive manufacturers, cold storage, and legacy industrials in competitive end markets. A sustained demand boom can also tighten regional power markets, creating localized price spikes that lag headline inflation by months and then pressure margins through 2026. The second-order pressure point is supply-chain scarcity: lead times for critical grid equipment can extend, which paradoxically benefits incumbents with installed capacity and penalizes smaller entrants that cannot secure components on time.

The contrarian risk is that consensus is extrapolating a straight-line capex cycle, when the real path is likely lumpy: permitting, transformer shortages, and utility interconnect delays can push monetization out 12-24 months. That makes the trade less about chasing “AI electricity demand” and more about owning the bottlenecks with contractual pricing power. If power prices jump too quickly, demand elasticity and political intervention become the medium-term check, particularly in regulated markets where rate cases can dampen returns.

Net: the move looks under-owned in the industrial supply chain and over-owned in the generic electrification basket. The best risk/reward is to own the toll collectors, not the end-demand story, and fade businesses whose input costs rise faster than they can reprice.

Key Decisions for Investors

• Long ETN / PWR on a 6-12 month horizon: these are direct bottleneck beneficiaries from grid and data-center infrastructure spend, with backlog-driven earnings leverage and relatively defensive demand. Add on dips if the market treats the theme as a short-lived AI trade.
• Pair trade: long HUBB or RBC / short an energy-intensive industrial basket (e.g., CAT vs. a utilities-equipment name is less clean, so prefer long electrical-equipment leaders against a short in a high-power-cost manufacturer in your universe). The thesis is margin compression for input-takers versus pricing power for suppliers; target a 10-15% spread over 3-6 months.
• Consider a call spread in a grid equipment leader over 9-12 months rather than outright stock: the market is likely to re-rate backlog visibility before full EPS conversion, giving asymmetric upside with defined downside.
• Avoid chasing regulated utilities with stretched multiples unless they have explicit data-center load growth and favorable rate frameworks; the upside from demand growth is often offset by lagged pass-through and political scrutiny. Use as funding shorts against better supply-chain beneficiaries.