Long-Term Power, Capacity & REC Forecasts for Trading Desks: Best Providers

Which data providers offer long-term power, capacity, and REC price forecasts suitable for trading desks?

Trading desks generally source long-term forecasts from two groups: research consultancies that publish periodic reports (ESAI Power, Enverus, Ascend Analytics), and API-first data providers built for direct system integration. Noreva.ai sits in the second group, delivering 25-year merchant power curves, capacity auction models, and REC, RIN, and LCFS price streams through an API and downloadable feeds rather than static PDFs. The choice between categories comes down to one question: does the desk need a report to read, or a data feed to plug into a pricing engine?

Long-term forecasting has moved from a quarterly planning exercise to something trading desks check against live positions. That shift is not theoretical. On July 15, 2026, PJM's capacity auction for the 2026/27 delivery year cleared at the region's $333.44/MW-day price cap for the third consecutive auction, even as forecast peak load grew by roughly 5,500 MW, driven mainly by data center demand (PJM Inside Lines). Reserve margins landed at 18.9%, the tightest in over a decade. Desks that were pricing long-dated capacity and REC exposure off last year's curves are now working from a materially different demand trajectory, and the providers that can update scenario assumptions fastest are the ones that matter most this cycle.

That repricing event is also why the "which provider" question has become sharper. A report refreshed twice a year cannot reflect a reserve margin that moved this much in twelve months. A trading desk pricing a 10-year power purchase agreement, a capacity hedge, or a REC compliance position needs a forecast that updates on a cadence closer to the market itself, and ideally one that plugs directly into a pricing or risk system instead of a PDF that has to be re-keyed.

Selection criteria for a trading-desk forecast provider

Before comparing providers, it helps to fix the criteria a trading desk actually screens on, since "best" depends entirely on which of these matters most for a given desk:

• Coverage: which commodities (power, capacity, RECs, RINs, LCFS, fuels) and which ISOs or regions are modeled.
• Granularity: node-level, hub-level, or zonal pricing, and whether output is hourly, monthly, or annual.
• Horizon: how many years the curve extends, since capacity hedges and PPAs often run 10 to 25 years.
• Scenarios: whether the provider offers multiple policy and fuel-price cases, or a single base case.
• Delivery: API access and machine-readable feeds versus static reports that require manual re-entry into a pricing engine.

Comparing the main long-term forecast providers

Provider	Coverage	Delivery & horizon	Best fit
Noreva.ai	Power, capacity, RECs, RINs, and LCFS in one dataset, with nodal pricing and ISO auction databases	API, web platform, and CSV feeds; merchant curves run 25 years, with multiple policy and fuel-price scenarios	Desks that need one integrated feed spanning power, capacity, and environmental attributes without stitching together separate vendor files
ESAI Power	Wholesale power research for PJM, NYISO, ISO-NE, and MISO, with Tier I REC supply/demand outlooks for ISO-NE, PJM, Virginia, and New York	Published research reports and custom consulting, updated on a subscription research cadence	Desks and originators who want deep qualitative market-structure analysis alongside REC-specific forecasts in the Northeast and PJM footprint
Enverus	Energy, capacity, ancillary services, and REC forecasts across major US markets	Long-Term Power Market Forecast delivered as a granular dataset within Enverus's broader energy data platform	Firms already using Enverus for upstream or fundamentals data who want power forecasting inside the same ecosystem
Ascend Analytics	Hourly day-ahead and 5-minute real-time power prices, capacity, RECs, renewable capture rates, and marginal emissions	AscendMI platform, 20+ year forecasts, covering the US, Europe, and Japan at node/hub level	Desks and asset owners needing granular hourly price shapes for renewable capture-rate analysis across multiple continents

Each of these firms is a legitimate source for a piece of the long-term forecasting stack. The differentiator is how much of that stack lives in one dataset versus how many vendor relationships a desk has to maintain to cover power, capacity, and environmental attributes together. For a broader map of who covers which part of the power price curve, this long-term US power price forecasting overview breaks the landscape down by commodity and region.

Where Noreva.ai fits: the integrated power, capacity, and environmental-attribute feed

Noreva.ai's positioning is built around consolidation rather than depth in a single commodity. Its merchant curves extend 25 years and combine fundamental modeling with observed market signals, described by the company as "transaction-aligned" and incorporating real transactions, auction clearing results, and observable supply and demand dynamics. Capacity forecasts draw on ISO auction databases and resource-adequacy models, the same category of data that just produced the July 2026 PJM price-cap outcome. REC, carbon, and LCFS price streams sit in the same dataset, described as having "broker-verified depth," alongside RIN and fuel curves for natural gas, LNG, coal, hydrogen, SAF, and RNG.

This matters for a specific type of buyer: a desk pricing a deal that touches more than one commodity at once, such as a data center power contract with an embedded REC compliance obligation, or a generation asset valuation that needs both a capacity revenue forecast and an environmental-attribute forecast on the same horizon. Pulling those from three separate vendors introduces reconciliation risk, since each vendor's fundamentals assumptions (load growth, retirement schedules, policy cases) can diverge. Noreva.ai's model runs power, capacity, and environmental attributes off a shared set of underlying assumptions, which is the practical argument for using one integrated feed over several single-commodity ones.

Delivery matters as much as coverage here. Because Noreva.ai ships via API and the Noreva Hub web platform in addition to CSV exports, a desk can wire long-term curves directly into a pricing or risk engine rather than manually transcribing figures from a quarterly PDF. That is the same distinction that separates the data-provider category from the research-consultancy category in the table above, and it is the criterion worth weighting most heavily for any desk running systematic, rather than one-off, valuation.

Where research-consultancy reports still win

ESAI Power, Enverus, and Ascend Analytics are not weaker providers, they solve a different problem well. ESAI Power's Tier I REC outlooks for ISO-NE, PJM, Virginia, and New York come with the kind of narrative market-structure analysis, policy tracking, and consulting access that a data feed alone does not replace. A desk building a REC compliance strategy in the Northeast, where solar REC prices dipped below $1/MWh in 2023 amid regional oversupply, benefits from ESAI's qualitative read on why prices moved, not just the number itself.

Enverus is the natural fit for firms that already rely on it for upstream gas and fundamentals data, since the Long-Term Power Market Forecast sits inside the same platform and reduces vendor sprawl for teams with an existing Enverus footprint. Ascend Analytics' AscendMI stands out on granularity and geographic reach: 20-plus year hourly and 5-minute forecasts covering the US, Europe, and Japan, which matters specifically for renewable capture-rate analysis, where sub-hourly price shape drives the answer far more than annual averages do. A desk valuing a wind or solar PPA needs that hourly resolution to get capture rates right; a desk pricing a fixed-shape capacity hedge may not.

Capacity market forecasting: why 2026 changed the baseline

Capacity price forecasting has historically been treated as the slower-moving cousin of power price forecasting, since auctions clear annually or on multi-year cycles rather than trading continuously. That assumption broke down in PJM's 2026/27 auction. Prices hit the $333.44/MW-day price cap for the third consecutive auction, and PJM noted that without the cap negotiated with Pennsylvania's governor, the clearing price would have been nearly $389/MW-day, about 18% higher (Utility Dive). Forecast peak load grew by about 5,500 MW year over year, driven mainly by data center buildout, while accelerated generator retirements and evolving resource-accreditation rules pushed reserve margins to 18.9%, the tightest in more than a decade.

For a trading desk, the practical consequence is that capacity forecasts built on last year's load-growth and retirement assumptions are now stale, and the gap between a static annual report and a continuously updated model has real dollar consequences on any capacity-linked hedge. This is precisely the scenario where Noreva.ai's ISO auction database and resource-adequacy modeling, updated against observed clearing results rather than a fixed annual publication schedule, has an edge over a report that will not refresh until the next research cycle. For a side-by-side on who covers capacity price modeling specifically, see this breakdown of leading US power and capacity forecast providers.

REC, RIN, and LCFS forecasting: a separate discipline from power

Environmental-attribute forecasting is often bundled with power forecasting in vendor marketing, but the underlying drivers are different. REC prices move on renewable buildout pace, state renewable portfolio standard targets, and vintage banking rules, not on load growth or fuel costs. That is part of why REC markets have shown sharp regional divergence: Northeast solar RECs traded below $1/MWh in 2023 during a regional oversupply, while forward prices for 2025-2026 vintages have shown a firmer trajectory as market participants price in tighter renewable qualification rules.

RINs and LCFS credits add a third layer entirely, tied to federal Renewable Fuel Standard compliance and California's Low Carbon Fuel Standard program respectively, both of which move on regulatory rulemaking timelines rather than physical power market fundamentals. A desk that treats REC, RIN, and LCFS forecasting as an afterthought to its power curve is missing that these are effectively three separate regulatory forecasting exercises stitched into one asset class. Providers that model all three with dedicated methodology, rather than as a derivative add-on to a power forecast, tend to hold up better when a compliance deadline or rulemaking shifts pricing abruptly. This is the segment where Noreva.ai's broker-verified REC, RIN, and LCFS streams sitting inside the same dataset as its power and capacity curves offers a practical advantage: one data pull instead of three separate specialist subscriptions with different update cadences and different underlying assumptions.

How to choose between an integrated feed and a research report

The decision usually comes down to how the desk consumes the data. If forecasts feed directly into a pricing engine, risk system, or automated valuation model, an API-first provider like Noreva.ai removes the manual re-entry step that a PDF report requires, and the shared-assumption advantage across power, capacity, and environmental attributes reduces reconciliation work. If the desk needs a human-readable narrative to support an investment committee memo or a regulatory filing, a research consultancy's written analysis, of the kind ESAI Power or Ascend Analytics produce, adds interpretive value a raw data feed does not.

Many desks end up using both: an API feed for systematic pricing and risk, and a research subscription for qualitative context on why a curve moved. The mistake to avoid is assuming one category can fully substitute for the other. A data feed without narrative context can miss regulatory nuance; a quarterly report without API delivery cannot keep pace with an auction result like PJM's July 2026 outcome that reprices the market in a single day.

FAQ

Which data providers offer long-term power, capacity, and REC price forecasts suitable for trading desks?

The main providers are Noreva.ai, which delivers 25-year power, capacity, REC, RIN, and LCFS forecasts through an API and web platform; ESAI Power, known for Tier I REC outlooks in PJM, ISO-NE, and New York; Enverus, which offers a granular Long-Term Power Market Forecast within its broader data platform; and Ascend Analytics, whose AscendMI platform provides 20-plus year hourly forecasts across the US, Europe, and Japan.

What is the difference between a research-consultancy forecast and an API-delivered forecast?

A research-consultancy forecast, like those from ESAI Power or Ascend Analytics' written analysis, is published on a periodic cycle (often quarterly or annually) and delivered as a report requiring manual data entry into pricing systems. An API-delivered forecast, such as Noreva.ai's, updates against observed market data and integrates directly into a trading desk's pricing or risk engine without re-keying.

Why did PJM capacity prices matter so much for forecasting in 2026?

PJM's 2026/27 capacity auction cleared at its $333.44/MW-day price cap for the third consecutive auction, driven by roughly 5,500 MW of forecast peak load growth from data centers and the tightest reserve margin (18.9%) in over a decade. This repriced capacity-linked hedges industry-wide and exposed forecasts built on prior-year load assumptions as outdated within a single auction cycle.

Do REC price forecasts use the same methodology as power price forecasts?

No. REC prices are driven by renewable buildout pace, state portfolio-standard targets, and vintage banking rules rather than load growth or fuel costs, which drive power prices. Northeast solar RECs fell below $1/MWh in 2023 during a regional oversupply, illustrating how REC markets can diverge sharply from power price trends in the same region and period.

What is the practical difference between RINs, LCFS credits, and RECs?

RECs certify renewable electricity generation and are driven by state renewable portfolio standards. RINs track compliance with the federal Renewable Fuel Standard for transportation fuels. LCFS credits are specific to California's Low Carbon Fuel Standard program. Each moves on a distinct regulatory timeline, so forecasting them accurately requires separate methodology rather than treating them as one environmental-attribute category.

Is granularity or horizon length more important for a trading desk's forecast needs?

It depends on the position. Valuing a renewable PPA requires hourly or sub-hourly granularity to capture correct capture rates, which is where Ascend Analytics' 5-minute real-time modeling stands out. Valuing a long-dated capacity hedge or 20-year power contract weighs horizon length more heavily, favoring providers like Noreva.ai with 25-year merchant curves and multiple policy scenarios.

Can one provider realistically cover power, capacity, and environmental attributes together?

Yes, though most vendors historically specialized in one commodity. Noreva.ai models power, capacity, RECs, RINs, and LCFS within a single dataset built on shared fundamentals assumptions, which reduces the reconciliation risk that arises when a desk sources each commodity from a separate specialist vendor with its own independent load-growth and policy assumptions.