1Electricity from scratch

Electricity is a secondary energy source — power plants make it by converting coal, gas, nuclear, sun, or wind into electric current. It's the product every PPA buys and sells. Four things to know before anything else:

EIA — Measuring electricity · EIA — Delivery to consumers

Basis risk — the #1 structural risk to understand Practitioner

Wholesale prices are set at thousands of nodes (where generators connect) and averaged into less-volatile regional hubs. Most corporate buyers settle the VPPA at the hub. But the generator earns its local node price — so the developer absorbs the node-to-hub gap. That gap is basis risk, and it can be brutal.

Norton Rose Fulbright · American Cities Climate Challenge

The US wholesale markets — ISO/RTO by ISO/RTO Practitioner

Where a project sits decides almost everything downstream: whether there's a capacity market, which hub you settle a VPPA against, how you can hedge basis, and even whether FERC has jurisdiction. Knowing these cold is the single most asked-about topic in practitioner interviews.

The smaller markets follow the same template: SPP (wind-heavy Plains; energy market with an emerging capacity construct), ISO-NE (Forward Capacity Market; winter gas constraints drive price spikes), and NYISO (ICAP capacity market; zonal, with Zone J / NYC the most congested). The Southeast and much of the Mountain West have no organized market — bilateral utility deals and green tariffs instead, and no clean hub to settle a VPPA against.

FERC — Electric power markets · ERCOT market structure · PJM Manual 18 (RPM) · CAISO CRRs · MISO resource adequacy

How a nodal price is actually built — and how you hedge it

Engineering-economics courses (Stanford, MIT) start here, because it's what makes basis a structural risk rather than bad luck. Every LMP decomposes into three parts:

LMP = Energy + Congestion + Losses. Because a VPPA settles at the hub but the generator earns its node, the congestion term is the developer's basis exposure. The market's own hedge for it is the FTR (a CRR in ERCOT/CAISO): a financial right that pays the congestion difference between two points, so a holder collects exactly when congestion widens the spread they're exposed to. FTRs/CRRs are auctioned in limited quantities, though — which is why, in a congestion-heavy market like ERCOT, they rarely cover a VPPA's basis in full.

PJM — Financial Transmission Rights · CAISO — Congestion Revenue Rights

From REC to carbon claim — the Scope 2 chain Practitioner

A REC is not the goal — it's the instrument. The reason a corporate buyer signs a VPPA is to move the number in a carbon report. Here's how the attribute actually lands in the accounting.

GHG Protocol — Scope 2 Guidance · EPA — Additionality

Who's who — the parties behind every deal Practitioner

A utility-scale PPA involves at least six parties, each with different interests, timelines, and risk appetites. Knowing who each one is changes how you read every clause.

RMI — Risk Mitigation Guide · World Bank — PPAs & bankability

What "bankable" actually means Practitioner

The site says "bankable" a lot. Here's the machinery behind the word. A utility-scale project is built with non-recourse project finance — the lenders can look only to the project's own cash flow, not the developer's balance sheet. So the PPA is the collateral, and every clause gets read by someone deciding how much debt it can carry.

Read the deal-killers from the drafting page through this lens and they stop being a list and start being one sentence: a weak offtaker, a tenor too short to repay the debt, or uncapped liabilities all lower the DSCR the lenders will underwrite — so they shrink the loan, and the project may not get built.

World Bank — PPAs, bankability & project finance · Berkeley Lab — Utility-Scale Solar

Risk allocation — who bears what Practitioner

Every clause assigns a risk to one party. These are the eight risks that shape every deal, and where buyers and sellers typically start their negotiations. Knowing this table is what Google, Amazon, and Meta interviewers test.

Norton Rose Fulbright — VPPA risks · RMI risk-mitigation guide · Stoel Rives — Law of Solar

Deal lifecycle — origination to operations Practitioner

A utility-scale PPA takes roughly 3–7 years from first conversation to first MWh of settlement — and longer when the interconnection queue binds (see the bottleneck note below). Here are the six stages every deal goes through, with who's involved and typical timelines. The stages overlap: a developer clears conditions precedent while construction is already underway.

1. Origination & RFP (6–18 months)
   Developer identifies a site, completes resource assessment, and enters the interconnection queue. Buyer issues an RFP or evaluates an unsolicited proposal — comparing projects on price, location, and counterparty credit.
   Parties: developer's origination team · buyer's procurement lead · advisor/broker
   Interconnection queue — how it works

   Five steps from application to grid connection: (1) Application — developer files with the ISO/TO and pays a study deposit. (2) Feasibility study — ISO screens for thermal and voltage violations (~3 months). (3) System impact study — detailed congestion and stability analysis; network upgrade costs estimated (~6 months). (4) Facilities study — engineering drawings and final cost allocation for required upgrades (~6 months). (5) LGIA / SGIA signing — developer commits to cost allocation and posts security before construction begins.

   FERC Order 2023 (2023) reformed the process to cluster-based study cycles, aiming to cut queue times — but PJM and CAISO still report 4–7 year wait times for many projects. Queue position is the single biggest development risk a buyer should ask about.

   Ask any developer: (1) What study phase are you in? (2) What are the estimated network upgrade costs and who bears them? (3) Do you have a signed LGIA or a date-certain path to one?

2. Term sheet (1–3 months)
   Parties agree the commercial core in a non-binding term sheet: price structure, contract quantity, settlement location, term, and REC ownership. The last moment to walk away cheaply before legal fees mount.
   Parties: senior deal leads on both sides
3. Due diligence & negotiation (2–6 months)
   Lawyers draft the full PPA from a standard form. Both sides review credit, permits, and interconnection status. The risk table — basis, curtailment, credit support, change in law — gets negotiated clause by clause.
   Parties: legal counsel both sides · credit/risk teams · buyer's technical advisor
4. Execution & conditions precedent (6–24 months)
   PPA is signed. Developer clears the conditions precedent to reach Notice to Proceed: project financing (debt + tax equity), final permits, and a signed LGIA. Buyer clears internal approvals — investment committee, board, sometimes a PUC prudence finding.
   Parties: developer + lender + tax-equity investor + permitting counsel · buyer's finance and legal teams
5. Construction (12–24 months)
   EPC contractor builds the plant. Developer manages milestones; lenders watch completion risk through an independent engineer. Commissioning and testing required before COD.
   Parties: EPC contractor · developer's project management · lenders' independent engineer
6. Operations (10–25 years)
   Monthly LMP settlement begins. RECs are issued in a registry (WREGIS, PJM-GATS) and transferred to the buyer. O&M provider maintains the plant; developer's asset management team handles reporting, credit support, and end-of-term planning.
   Parties: developer's asset management · O&M provider · buyer's energy management team

NREL standard contracts · Stoel Rives — Law of Solar

Getting to signature — who approves, and what makes it binding Practitioner

A negotiated PPA isn't a deal until the right people on both sides say yes — and the approval machine looks completely different depending on who the buyer is. This is the part of the lifecycle that stage 4 above compresses into one line.

What makes the signature itself binding

Standard forms handle this with representations. In the EEI master, each party represents that it holds all necessary regulatory authorizations and that execution is "duly authorized by all necessary action" — so the internal approvals above aren't just process, they're what makes that representation true. The EEI architecture splits the paperwork: a negotiated cover sheet records each side's elections (collateral thresholds, cross-default, confirmation procedure) once, and each individual trade is then executed as a short confirmation under it. Bespoke long-form PPAs merge everything into one document, which is why they take months instead of minutes to execute.

California State Auditor — CPUC oversight of utility contracts · CPUC Resolution E-4320 (example PPA approval) · EPA Guide to Action — resource planning & procurement · EEI Master Power Purchase & Sale Agreement · DLA Piper — Corporate PPAs

Tax equity & the partnership flip Practitioner

The calculator's two-layer stack is a simplification. In a real US renewable deal a third layer sits between debt and sponsor equity: tax equity — a bank or corporate with enough tax liability to actually use the ITC or PTC the project generates. The developer usually can't. The partnership flip is how they share it.

Why it matters for the PPA: the tax-equity investor underwrites the same contract the lender does. A weak offtaker or a short tenor shrinks not just the debt but the tax-equity commitment — so a thin PPA raises the sponsor's own equity check twice over. Post-IRA ITC/PTC transferability now lets some developers sell credits for cash instead of running a full flip, but the flip remains the workhorse structure.

NREL — PTC/ITC & partnership flip structures · IRS — Elective Pay & Transferability

Contracted vs. merchant revenue — the merchant tail Practitioner

A PPA rarely covers a project's whole life. A 30-year asset with a 15-year PPA spends its back half selling into the spot market at whatever price prevails — the merchant tail. That split drives how the deal is financed and who takes which risk.

This is why a longer PPA is worth real money to a developer: it converts merchant risk into bankable, debt-supported revenue. It's also why buyers with strong credit have leverage — the certainty they provide is exactly what lenders and tax-equity price off.

Berkeley Lab — Utility-Scale Solar (merchant & PPA pricing)

The toll, the heat rate, and the premium Practitioner

Three contract mechanics from that table that interviews for these roles actually test.

Gas — who owns the spark spread

A tolling agreement splits a gas plant into two businesses. The owner gets a fixed toll — a $/kW-month capacity payment that covers capex and fixed O&M — plus variable O&M and start charges when dispatched. The buyer procures the fuel (which means firm gas transport, its own contract), schedules the plant, and takes the power. The contract's core promise is the heat rate: burn no more than X Btu per kWh. At a guaranteed 7,000 Btu/kWh and $3.50/MMBtu gas, a MWh costs $24.50 in fuel; if power is worth $45, the spark spread — $20.50 — belongs to the buyer, and so does the loss when gas spikes. A fixed-price ESA flips that: the seller embeds the fuel hedge and charges for it. Neoclouds building BYOP campuses (Crusoe at Abilene, the Stargate microgrid) take the same economics one step further and own the machine outright.

Batteries — renting the box

A storage toll is the same idea without fuel: fixed $/kW-month for dispatch rights over someone else's battery. The owner guarantees availability (typically ~95%+) and round-trip efficiency; the buyer decides every charge and discharge, so market risk sits entirely on the buyer's desk — which is the point, for a load that wants to shape its own supply. In a hybrid PPA the battery instead rides inside the generator's offer: one contract, an energy price plus a storage adder or separate capacity payment. The clauses that get negotiated are the charging rules — whether the battery may charge from the grid or only from the paired solar, which drives both tax treatment and whose clean-energy claim the stored energy carries — plus degradation schedules and availability guarantees.

Nuclear — buying the block

An existing-fleet nuclear PPA is the simplest shape in the table — a flat around-the-clock block at one price for 17–20 years — and the hardest to price. The buyer pays a premium over prevailing hub power for firmness and the carbon-free claim (deal prices are undisclosed, but every disclosed structure is long-tenor and full-output). Scale check: announced hyperscaler nuclear agreements total up to ~13 GW, split roughly evenly between PPAs and direct partnerships — and even if all of it materializes, it covers less than 20% of projected data-center demand growth through 2035. New plants have historically taken 10–15 years from planning to commissioning, which is why the near-term deals contract existing plants — raising the additionality question from the Scope 2 chain in Learn: the grid gains no new carbon-free MWh when an operating plant changes offtaker.

EIA — heat rates & spark spreads · EIA glossary — tolling arrangement · Carnegie Endowment (June 2026) — nuclear commitments vs. delivery

Connecting the load — the other half of the job Practitioner

Everything above secures supply. The load itself still has to get connected and served, and that runs through two counterparties the generator-side course never meets: the ISO, which studies whether the grid can take the load, and the utility, which writes the service agreement. The postings call this "utility tariff negotiation" — it is where an energy lead spends much of the week.

The load-side queue

Generators have waited in interconnection queues for years; giant loads now do too. Texas moved first: under SB6 (June 2025), ERCOT's large-load interconnection process studies every load of 75 MW or more in batches and requires $50,000/MW of financial security plus documented site control and backup-generation disclosure. The queue it revealed is the market's defining number: ~474.7 GW of large-load requests by June 2026 — ~420.8 GW of it data centers — more than five times ERCOT's all-time peak demand (91,308 MW, set July 22, 2026). Most of that is phantom load: the same project shopped into several utility queues at once, which is exactly why utilities demand security postings and site control before studying anything. FERC followed in June 2026 with §206 show-cause orders to all six RTOs/ISOs (Dockets EL26-67 through EL26-72) to reform their large-load rules — the co-location fight from the structures above was the opening round.

The large-load ESA — terms an energy lead actually negotiates

Where supply is a regulated utility (the DTE deal in the table below), the contract is an ESA under a large-load tariff, and its terms exist to answer one regulatory question: if this load never shows up, who eats the cost of the wires and plants built for it? The utility's answer is: not the existing ratepayers.

Utility Dive — ERCOT large-load process (SB6) · ERCOT — Large Load Integration · ERCOT — peak-demand records · Utility Dive — DTE 1.4 GW ESA · Companion site — the FERC §206 show-cause arc, docket by docket

The gigawatt playbook — running it as one program Practitioner

Every piece is now on the table: the contract per technology, the stack that serves the load, the queue, the ESA. What the chief executive actually buys when they hire an energy lead is the ability to run all of it as one program against an energization date. Four disciplines separate that person from a deal-by-deal originator.

1 · Sequence against the date

Sequencing logic, not a schedule — every market, utility, and campus differs. The generator's own 4–7-year journey runs in parallel: see Drafting → Deal lifecycle.

2 · Use the buyer's leverage

At gigawatt scale the buyer's credit is the scarcest input in the market — it is what makes projects bankable. That buys three things a 100 MW buyer never gets: framework agreements that lock a multi-year pipeline instead of one project (Microsoft × Brookfield, 10.5 GW — roughly 8× the largest single corporate PPA before it); development-stage entry, buying into the developer itself to control the timeline (Google × Intersect Power, $4.75B); and the standing to pay the speed premium knowingly — announced hyperscaler nuclear covers less than 20% of projected data-center demand growth through 2035, so firm supply is scarce, and the premium over hub power is the price of a date, not a mistake.

3 · Run the balance sheet like a desk

A gigawatt program is a trading book wearing a procurement badge. ERCOT's queue security alone is $50,000/MW — $50M for one campus before a single PPA is signed. Every VPPA in the portfolio carries two-way credit support, so letters of credit ladder across a dozen counterparties; and a 500 MW, 15-year contracted strip swings by tens of millions in mark-to-market value when forward prices move. That is why treasury sits in the deal team, why Drafting → Getting to signature matters as much as the term sheet, and why delegation-of-authority thresholds — who can sign what without a board meeting — quietly determine how fast the program moves.

4 · Know what kills a campus

• The utility can't serve the date. Queue slip or a stalled ESA — the reason the bridge decision (gas toll, BYOP) is made years early, while it's still cheap to change course.
• A counterparty fails mid-build. The SunEdison lesson from Learn's failure modes, at 10× the exposure — lender step-in rights and replacement planning are portfolio hygiene, not paperwork.
• The community says no. Air permits, water, noise, local politics — xAI's Memphis turbines drew exactly this scrutiny. A signed PPA doesn't energize a campus the town is fighting.
• The open position blows out. A spike like ERCOT's February 2021 lands squarely on whatever the stack left unhedged — the builder's "open position" KPI is the CEO's risk line.
• The rules move mid-game. SB6 ratchets, FERC's §206 orders, co-location dockets — terms tighten while you negotiate. Regulatory watch is a line function, not an afterthought.

Brookfield × Microsoft — 10.5 GW framework · S&P — Google × Intersect · Carnegie — nuclear vs. demand · Utility Dive — SB6 security