🛢 Energy & Commodities mixed · 6–18 months
A what‑if from the future

What if AI-efficiency breakthrough collapses datacenter power-demand forecasts?

A step-change in model efficiency sharply lowers electricity per unit of compute, undercutting utility load-growth projections and triggering a de-rating of power names tied to the AI-demand thesis.

21%
our model probability
over 6–18 months
prediction markets — wisdom of the crowd
loading live odds…
Empirically anchored 21% · 90% range 6–35% · 37 analogues · measured class tech_ai_bear 72% in 18 mo · 3% held back for the unknown
how we built this number — every step
Measured class rate — tech_ai_bear ≈0.8595/yr → 72% in 18 mo72%
Analyst prior · editorial share 30% of the class22%
Pooled · weight 86%21%
Crowd — no liquid market
Reserve 3% · no extremizing (×1.0)21%
Published21%

The class rate is measured from our dated, sourced event library (decade-normalized Poisson — the full table is public at base_rates.json). The variant’s share within its class is the analyst’s editorial call, published so you can audit it. A wider range means thinner precedent. Full recipe: methodology · scored at Reality Check.

The butterfly cascade

How this trigger trickles across markets, left → right — the root shock, its first‑order moves, then the ripple effects. Drag any node; tap a market for its real price history.

Resolution timeline — how this probability is moving

Our model's odds (gold) over time vs the crowd's (Polymarket, blue), from the past toward the 6–18 months horizon. Each dot is a real macro event that nudged the probability — green pushed it up, red pushed it down. Tap a dot for the source. The gold path is an illustrative reconstruction anchored to today's estimate — real dated events, not a live re-estimate history.

loading the timeline…

What it would mean

If this plays out, it is a mixed shock. A step-change in model efficiency sharply lowers electricity per unit of compute, undercutting utility load-growth projections and triggering a de-rating of power names tied to the AI-demand thesis. The trigger decomposes into signed root‑shocks — AI capex ▼ · Industrial demand ▼ · Risk appetite ▼ — which propagate through our causal graph to the markets below.

If it happens — the markets it would move

Biggest moves first. Projected moves are cascade-model priors; hist A–B% = what comparable past events actually did (measured abnormal returns), and model prior · unmeasured marks markets with no analogue backing yet. Tap any market for its price history.

MarketClassProjected move
1Nvidia NVDAon Hyperliquid 📈 chartEquity▼ -0.4%
hist -3.8–+1.2% · other way +6.83% (n=12)
2Freeport (copper) FCX 📈 chartEquity▼ -0.4%
hist -0.65–+0.69% · other way +2.41% (n=12)
3Solana SOLon Hyperliquid 📈 chartCrypto▼ -0.3%
hist -0.26–-0.03% · other way -5.52% (n=10)
4Broadcom AVGOon Hyperliquid 📈 chartEquity▼ -0.3%
hist -0.2–-0.06% · other way +0.76% (n=11)
5Micron MUon Hyperliquid 📈 chartEquity▼ -0.3%
hist -2.77–+1.08% · other way +0.43% (n=12)
6Hyperliquid (HYPE) HYPEon HyperliquidCrypto▼ -0.3%
model prior · unmeasured
7Semiconductors SMHon Hyperliquid 📈 chartEquity▼ -0.2%
hist -0.14–-0.09% · other way +0.88% (n=12)
8MicroStrategy MSTRon Hyperliquid 📈 chartEquity▼ -0.3%
hist -1.13–+2.46% · other way +18.0% (n=12)
9Ether ETHon Hyperliquid 📈 chartCrypto▼ -0.2%
hist -7.39–+2.59% · other way +6.05% (n=10)
10Nasdaq 100 NDXon Hyperliquid 📈 chartIndex▼ -0.2%
hist -0.15–-0.01% · other way +0.67% (n=12)
11Copper XCUon Hyperliquid 📈 chartCommodity▼ -0.2%
hist -0.48–+0.89% · other way -0.83% (n=12)
12Bitcoin BTCon Hyperliquid 📈 chartCrypto▼ -0.2%
hist -4.36–+0.63% · other way +10.62% (n=11)
13Tech sector XLK 📈 chartEquity▼ -0.1%
hist -0.11–-0.05% · other way +0.8% (n=12)

Probable recommendation

If the scenario above plays out, the probable cross‑asset positioning → a scenario‑conditional read, not personalized investment advice
Long
For a common-man portfolio: Mixed for a typical portfolio — the move is more about rotation than direction. Favour the winners over the losers below rather than net exposure.
Also moves (not yet on Hyperliquid): Freeport (copper) -0.4% · Tech sector -0.1%

Historical precedent — what analogous events actually did

Across 37 analogous events (overlap‑weighted), as abnormal returns — market beta stripped, so it's the event's own effect, not the market backdrop. Shown at 20 days (persistent) and 5 days (immediate); ↺ fades = the two horizons disagree. Confidence = consistency × sample × significance.

PJM grid emergency during Winter Storm Elliott 2022-12 Texas grid failure during Winter Storm Uri 2021-02 Northeast blackout cascading grid failure hits ~55 million 2003-08 Gold tops $4,000 and silver spikes past $50 in historic squeeze 2025-10 Gold tops $3,000 for the first time amid tariff and rate-cut fears 2025-03 TSMC slumps as DeepSeek roils AI-chip demand assumptions 2025-02 Micron's weak FQ2 guidance sparks a sharp December selloff 2024-12 Henry Hub natural gas hits a 25-year low amid record US production 2024-11 ASML bookings-miss crash 2024-10 Waha hub natural gas prices crash to record negative on Permian glut 2024-08 Gold tops $2,500 for the first time on Fed rate-cut bets 2024-08 Megacap AI-capex doubt selloff 2024-07 Trump 'Taiwan should pay for defense' chip selloff 2024-07 Niger coup d'etat 2023-07 Netflix subscriber-loss crash 2022-04 Meta 2022-02 Didi removed from China app stores after NYSE IPO 2021-07 California rolling blackouts during a record heatwave 2020-08 Gold closes above $2,000/oz for the first time 2020-08 Norilsk Nickel Arctic diesel spill 2020-05 Nvidia crypto-glut guidance crash 2018-11 North Korea sixth nuclear test 2017-09 North Korea 'fire and fury' nuclear scare 2017-08 Gold futures velocity-logic flash crash 2014-01 Gold all-time peak of $1,921/oz 2011-09 Egyptian revolution / Mubarak uprising 2011-01 Henry Hub natural gas spot price peaks during 2008 commodity boom 2008-07 Platinum hits all-time record near $2,290 on South African power crisis 2008-03 South Africa Eskom power emergency spikes platinum/PGMs 2008-01 Amaranth Advisors natural-gas blowup 2006-09 Soviet August coup attempt against Gorbachev 1991-08 Chernobyl disaster 1986-04 Silver Thursday 1980-03 Gold peaks at $850 1980-01 Iran hostage crisis / US freezes Iranian assets 1979-11 Three Mile Island partial meltdown 1979-03 1979 Iranian Revolution oil shock 1979-01
AssetHistory saysAbnormal (20d · 5d)HitnConfidencevs cascade
Bitcoin BTCSHORT-3.6% · 5d -1.7%74%22 0.36✓ matches cascade
Volatility VIXLONG+4.8% · 5d +0.1%65%31 0.24·
10y yield DGS10SHORT-3bp · 5d +1bp ↺ fades60%37 0.17·
NVDA NVDASHORT-3.3% · 5d -5.2%59%30 0.13✓ matches cascade
ETH ETHSHORT-6.8% · 5d -3.7%60%20 0.13✓ matches cascade
Gold XAULONG+0.1% · 5d -0.3% ↺ fades57%30 0.13·
MSTR MSTRLONG+2.5% · 5d -2.9% ↺ fades57%30 0.12⚠ differs
XCU XCULONG+0.9% · 5d -0.1% ↺ fades56%30 0.11⚠ differs
High-yield credit HYGLONG+0.2% · 5d +0.1%57%28 0.10·
SMH SMHLONG+0.0% · 5d -1.0% ↺ fades54%30 0.07⚠ differs
MU MUSHORT-2.5% · 5d -1.9%54%32 0.06✓ matches cascade
NDX NDXLONG+0.1% · 5d -1.1% ↺ fades54%32 0.06⚠ differs
FCX FCXLONG+0.9% · 5d +0.2%46%30 0.00⚠ differs
SOL SOLLONG+0.1% · 5d -5.7% ↺ fades40%19 0.00⚠ differs

Methodology. Probability and impact are anchored to history and scored against what actually happens — wins and losses, in public, at Reality Check. Crowd odds live from Polymarket & Kalshi. By Vikas Singh, Quantitative Strategist. Updated 2026-07-03.