PowerLink
Appendix

Source Appendix

Citations [1]–[15]. Footnote markers [n] throughout the site resolve here. Figures carry the as-of dates shown. Reviewed June 2026.

Grid constraints and power markets

The interconnection, equipment, and capacity-price evidence behind the argument that electricity is the binding constraint on AI buildout.

  1. [1]

    ~55-month interconnection delay

    LBNL, Queued Up (2025 ed.): U.S. projects reaching commercial operation in 2024 spent ~55 months in queue; median request-to-COD exceeds 4 years. View source ↗
  2. [2]

    ~128-week transformer lead time

    Wood Mackenzie transformer survey (Q2 2025): standard large power transformers average ~128 weeks; GSU units ~144 weeks, up from ~50 weeks in 2021. View source ↗
  3. [3]

    +833% PJM capacity price

    PJM Base Residual Auction: 2025/2026 delivery year cleared $269.92/MW-day vs. $28.92 for 2024/2025, about +833%. View source ↗
  4. [4]

    ~2,600 GW interconnection queue

    LBNL, Queued Up: ~2.6 TW active queue at YE2023; ~2.29 TW at YE2024. U.S., requested capacity. View source ↗
  5. [13]

    Grid age / transmission timelines

    DOE / ASCE: much U.S. grid infrastructure is 40+ years old; new transmission commonly takes 10-15 years to build. View source ↗

Critical minerals and supply concentration

Processing and mining concentration, demand scenarios, and material-intensity figures for lithium, copper, cobalt, and rare earths.

  1. [5]

    90%+ rare-earth processing

    IEA 2024-26: China ~91% refined rare-earth output, ~94% sintered permanent magnets, and ~60% of mining. View source ↗
  2. [6]

    ~70% cobalt from DRC

    IEA / USGS / Cobalt Institute: DRC ~70-73% of global mined cobalt supply in 2024-25. View source ↗
  3. [7]

    +500% lithium demand to 2050

    IEA / Wood Mackenzie: lithium demand rises several-fold by 2050. IEA Net-Zero ~10x; WoodMac 5.6-13.2 Mt LCE. +500% is conservative-to-mid case. View source ↗
  4. [8]

    3-4x copper intensity

    Industry analyses (copper.org / USGS / WEF): AI / hyperscale facilities can run ~3-4x copper intensity vs. conventional per facility; per-MW differential is lower. View source ↗

Feedstock, flaring, and siting

Gas pricing at constrained hubs, flaring volumes, scrap-tire generation, and the solar and hydro resource data behind site selection.

  1. [9]

    Waha negative gas pricing

    EIA / S&P Global / NGI: Waha averaged near $0/MMBtu in 2024 and traded negative on ~42% of days; lows around -$7 in 2024 and -$9.52 in Apr. 2026. View source ↗
  2. [10]

    Tire volumes

    USTMA / EPA-class data: global tire production ~2.5B units/yr; U.S. scrap-tire generation ~280M/yr, ~4.2-4.5M tons. View source ↗
  3. [11]

    Bakken flaring

    EIA / VIIRS satellite flaring data: Permian is the leading U.S. flaring basin; Bakken is among the largest by flared volume. No point-rank. View source ↗
  4. [12]

    High-DNI corridor

    NREL / EIA solar resource maps: highest U.S. direct-normal irradiance is concentrated in the desert Southwest; Arizona sites qualify regionally. View source ↗

Eswatini

National generation capacity, peak demand, and the labour-market data referenced on the Hydro page.

  1. [14]

    Eswatini national peak demand

    EEC / ESERA / SACREEE 2024: national peak demand ~230-234MW; ~71MW installed domestic capacity, including ~60MW hydro and ~10MW solar; balance imported via SAPP. View source ↗
  2. [15]

    Eswatini youth unemployment

    World Bank / ILO-modeled 2024: youth unemployment ~54%; IMF cites ~48.7% for 2023. View source ↗

This site contains forward-looking statements that are subject to risk and uncertainty, including permitting, financing, site readiness, commissioning, feedstock availability, commodity and power-price volatility, and regulatory conditions.

← Back to home