HVDC’s “unmapped toll” is landing in converter stations and XLPE cable, not reactors—because PJM is already pricing grid pain at ~$6B/H1 and AI will have to pay it again

AI load growth doesn’t just increase total megawatts. It increases the frequency and severity of “stress events” that expose long-haul transfer limits—then markets price those limits.

On PJM, the largest U.S. power grid, congestion/transmission-constraint costs rose to about $6B in the first half of 2026 (up 43% year over year), according to Reuters citing PJM market monitoring data. The key investor takeaway is not the magnitude alone; it’s the direction: grid pain is rising faster than ordinary demand growth, which pushes utilities toward the fastest controllable capacity additions—often HVDC long-distance transfer rather than incremental AC upgrades.

An HVDC link typically requires two converter stations (one at each end) plus an HVDC cable/line segment in between. In practice, the “attention bottleneck” is rarely the conceptual circuit diagram; it’s delivery of the physical modules with long procurement lead times and scarce qualification slots.

The article’s core supply-chain claim is simple: the value pool shifts toward converter stations and HVDC XLPE cable orders rather than toward the most visible “reactor” part of an HVDC schematic. Reactors matter for performance and filtering, but the buildout constraint most investors can verify quickly is that cable qualification, conductor/cross-link manufacturing capacity, and converter-station integration drive multi-year schedules—exactly the kind of constraints that turn into pricing power when demand is “lumpy” (data-center clusters) and reinforcement is time-bound.

XLPE HVDC cables are designed to operate at high conductor temperatures and high DC stresses, which forces suppliers through strict qualification and production ramp steps. That matters for investors because it links grid deadlines to supplier throughput—not just engineering.

NKT’s disclosures provide two investor-grade datapoints. First, in March 2023 NKT announced record orders for the world’s first 525 kV XLPE HVDC submarine cable projects for offshore wind, with cable production scheduled to start between 2024–27 and commissioning expected in 2028–30. Second, on Aug. 24, 2026, NKT announced it qualified a 525 kV HVDC subsea cable system for operation at 90°C conductor temperature—directly tying performance improvements to the operational envelopes utilities will contract for.

Converter stations bundle power electronics, transformers, AC/DC switchgear, control systems, and the integration work needed to turn “link design” into “deliverable capacity.” When an AI-driven load requires long-distance transfer, utilities will typically prioritize the HVDC end-to-end path that unlocks capacity transfer.

This is why the value capture chain is station-centric: station-level EPC scopes are where schedules are most difficult to compress because multiple subsystems must clear testing, factory acceptance, and grid-side commissioning sequencing.

ABB’s FY2025 revenue was $33.22B, with EBITDA of $7.12B (both from company financial statements). That kind of scale matters because it suggests the market expects electrical equipment vendors to remain funded through long-cycle grid capex.

But investors should ask the tighter question this piece focuses on: when utilities pay for long-distance transfer, which line items actually behave like backlog-driven, multi-year revenue—converter stations and HVDC XLPE cable scope, where the physical constraints are easiest to prove with qualification and commissioning timelines.

Reactors can be engineered, procured, and tuned—but for most investors the real question is “what is time-bound and qualification-bound?” In HVDC links, that tends to show up in converter-station integration and HVDC cable production/qualification.

So the non-obvious link here is: PJM’s congestion costs rise → utilities prioritize faster controllable transfer → procurement shifts toward the HVDC end-to-end path → converter stations and XLPE cable scopes capture the timeline-limited spend. Even if reactors are required technically, they’re less likely to dominate the multi-year delivery bottleneck compared with stations and XLPE cable systems that must meet strict operational envelopes.

Short-term (days to quarters): watch for HVDC project award timing that shows utilities treating congestion as a solved problem rather than an acceptable cost. If procurement accelerates, cable orders and station integration milestones should follow first because those are the schedule-critical steps.

Medium-term (1–3 years): track commissioning windows and qualification progress for XLPE HVDC cables. NKT’s 90°C qualification disclosure is exactly the type of “operational envelope” improvement that can expand utilization and support future award decisions.