On August 31, Nodal Exchange plans to list 168 hourly power futures — one contract for every hour of the day across seven ERCOT and PJM locations. The launch does more than add granularity. It turns the shape of the day into a directly tradable curve.

Posted July 21, 2026 — six weeks before launch.

The average is no longer the market

Power risk used to be packaged into broad blocks: a month, a day, on-peak, off-peak. That framework was built for a system dominated by thermal generation and relatively predictable load.

Solar and storage changed the geometry.

The important question is no longer simply, “What will power cost tomorrow?” It is, “Which hour owns the scarcity, which hour gets crushed by oversupply, and how wide can the distance between them become?”

Nodal’s planned launch completes a granularity ladder it has been building through monthly, daily, and now single-hour futures. The exchange plans to list all 24 hours at ERCOT North, South, West, and Houston, plus PJM Western, AEP-Dayton, and Northern Illinois, with the next five days available at all times.

But Nodal is entering a market whose hourly risk already has multiple wrappers:

• Nodal: an exchange-cleared strip of individual hourly futures.
• ElectronX: bounded, fully collateralized hourly futures and $100-or-nothing binaries.
• ICE: TB4 and shaped blocks that package the economics of intraday spreads.

One day can contain two different power markets

At ERCOT North in summer, real-time hour-ending 20 averaged roughly $205/MWh over the three-year sample. The 2–5 a.m. hours averaged about $21/MWh.

That is a 10:1 ratio inside the same day, before isolating a scarcity event.

The daily top-four-minus-bottom-four spread — ICE’s TB4 geometry — averaged $85/MWh at ERCOT North, but the median was only $34/MWh. The gap between those two numbers tells the story: the distribution is not merely volatile; it is sharply right-skewed.

The 99.7th-percentile day reached $3,133/MWh. The largest day in the sample, September 6, 2023, printed $3,771/MWh.

That distinction matters for every new hourly product. A contract can reference the same hub and the same hour while producing very different economics depending on whether the exposure is unbounded, capped, floored, packaged into a block, or converted into a binary payout.

Day-ahead is not a neutral strike

Across the sample, ERCOT North real-time cleared above day-ahead in only about 41% of hours. At PJM Western, the rate was closer to 32%.

That is the day-ahead premium in one number: sellers of day-ahead power are paid, on average, for accepting real-time tail risk.

It is also a direct fair-value input for ElectronX’s binaries, which pay $100 when real-time settles above the day-ahead strike. A flat 50-cent prior is not neutral. It is systematically wrong — and the degree of wrongness changes by hour.

Overnight, real-time beats day-ahead more frequently. By mid-morning, the hit rate falls. The binary is therefore not just a view on “power up or down.” It is a piece of hour-specific insurance.

Selling the binary is selling that insurance. The dashboard shows what the premium has historically been worth.

The band is part of the trade

ERCOT West makes the bounded-futures question especially visible.

In the historical sample, real-time power cleared below zero in roughly one of every six midday hours. In the evening peak, it breached $100 in approximately 9–12% of hours.

A bounded future with a $0 floor removes the negative tail. A $100 ceiling removes part of the positive tail. That means the standard-band contract can be worth structurally less than an unbounded hourly future precisely when the underlying hour is most interesting.

The floor and ceiling are not simply collateral settings. They change the instrument.

Three venues, three ways to own the hour

Four trades the hourly strip makes possible

Weather lights the match. Load decides whether it catches.

Across a full year, daily maximum temperature explains only part of TB4 variation. The relationship strengthens in summer, and PJM’s intraday spreads tend to track peak load more closely than ERCOT’s.

Heat is necessary, but not sufficient.

That is why the spike premium survives: the market is not pricing temperature alone. It is pricing the interaction among temperature, system load, renewable output, outages, congestion, and the hour in which they collide.

The opposite regime matters too. Mild, windy shoulder-season stretches can compress TB4 into its bottom decile for days at a time. A long spread position does not need a dramatic loss to hurt; it can simply bleed while nothing happens.

What to watch before August 31

Final contract specifications

Nodal’s self-certification should clarify the settlement index, contract size, tick, and other final mechanics. The launch announcement frames the contracts as tools for managing real-time positions, but the exact settlement language will determine fair value and hedge fit.

Portfolio-margin offsets

The most important operational question may be whether Nodal Clear recognizes offsets between hourly, daily, and monthly positions at the same hub.

If the strip receives meaningful offsets, it can trade as a curve. Without them, it risks behaving like 24 capital-intensive orphans.

The storage feedback loop

Every additional megawatt of battery capacity is a potential seller of the same intraday spread these contracts make tradable.

That structural short can compress ordinary-day spread medians even while scarcity tails remain extreme. The result is a market with quieter typical days and still-violent exceptional ones — a difficult distribution, and exactly the kind of distribution that rewards precise contract design.

Drive the market before it launches

Choose a location. Flip between day-ahead and real-time. Find the peak hour, the crushed hour, and the spread between them. Then change the wrapper — unbounded, bounded, binary, or TB4 — and watch the economics change.

The contract is not merely the reference price.

Data and methodology. ERCOT MIS day-ahead and real-time hub settlements (reports 13060/13061) and native-load archives; PJM Data Miner 2 day-ahead and verified real-time hub LMPs (da_hrl_lmps, rt_hrl_lmps, hrl_load_metered); Open-Meteo historical weather. July 1, 2023 through July 20, 2026. Approximately 374,550 hourly settlement observations. Contract descriptions are based on Nodal Exchange, ElectronX, CFTC, and ICE materials available as of July 21, 2026. Nodal hourly specifications remain subject to final regulatory filings and may change. Analytics are historical and precomputed. Not investment advice.