For large C&I operators evaluating co-location of load with generation, the regulatory middle ground — large-scale behind-the-meter configuration with grid backup — has been eliminated by FERC, and the Talen-Amazon Susquehanna restructure is the precedent that governs. If your project's economics depend on netting out consumption behind the meter to avoid transmission and standby charges, that assumption needs to be reexamined now.

This post is for energy and operations decision-makers at data centers, large manufacturing facilities with onsite generation, industrial parks, and green hydrogen producers who are evaluating whether to configure load behind or in front of the meter. If you are trying to determine whether a behind-the-meter arrangement can still eliminate your transmission and standby charge exposure — or whether FERC has closed that door at your scale — this is built for you. By the end, you will know what behind-the-meter and front-of-meter configurations actually mean in regulatory terms, when each was viable, what killed the large-scale BTM path in PJM, and the four interconnection mechanics FERC has approved as the replacement framework.

What Co-location of Load with Generation Actually Is

Three configurations are in play, and operators conflate them constantly. The distinctions are precise and consequential.

Behind-the-meter (BTM) means the generator sits on the consumer's side of the utility meter. It serves the load through a private line. No interconnection agreement with the grid is required, because power never crosses the meter in the direction of the grid. The generator is invisible to the transmission system.

Co-location — the BYOG (Bring Your Own Generation) model — means a generation facility is sited adjacent to the large load and interconnects directly to the bulk power grid. The generator serves the load contractually through a power purchase agreement, and power flows through the meter. The load takes a transmission service product for whatever residual it draws from the grid. In a typical BYOG arrangement, onsite generation covers roughly ninety percent of the load, with about ten percent supplied from the grid.

Front-of-meter (FTM) means the system is on the utility side of the meter, fully inside the regulated grid. The operator takes power as any other transmission customer would, from generation that the regulated grid delivers.

The practical consequence of these distinctions shows up on your bill in two ways. First, in whether you owe transmission charges at all. Second, in whether you face standby charges for backup service when your onsite generator is offline. If you have not read the TEG post on utility standby charges for on-site generation, that is directly relevant context — because whether you are behind or in front of the meter completely determines how standby charges hit you, or whether they hit you at all.

Why the BTM vs. Front-of-Meter Question Exists — and Why FERC Is Forcing It Now

The reason this decision is in front of large operators right now is PJM's interconnection queue. It currently stretches into the early 2030s. If you are a large load that needs power in 2026 or 2027, waiting in that queue for a new interconnection agreement is not an option. BYOG — co-locating your own generation adjacent to your facility and interconnecting it to the bulk grid — emerged as the fastest legitimate path around that bottleneck.

Behind-the-meter was the even faster path: no interconnection agreement at all. For decades, BTM rules worked because small onsite generators had no meaningful financial consequence for the bulk transmission grid. A 500 kilowatt rooftop solar array netting out behind the meter is not a transmission-cost issue. A two-gigawatt data center doing the same thing is a fundamentally different situation.

When a load that large nets out its consumption behind the meter, it avoids transmission and grid charges — and those charges get redistributed to every other customer on the system. FERC's position is that large-scale BTM arrangements do not adequately reflect grid infrastructure and upgrade costs, and they are correcting that. This is not a prediction about where policy is heading. The Talen-Amazon outcome and the April 2026 order together make it a current regulatory fact in PJM.

When BTM Helped — and What Ended It for Large Loads

The behind-the-meter configuration was genuinely useful for large loads that could run their onsite generation reliably enough to minimize grid interaction. No interconnection agreement meant no queue wait. No net grid withdrawals meant no transmission tariff exposure. And if you could design the system to serve your entire load without backup draws, you could theoretically avoid standby charges too.

The design worked at smaller scale because regulators either did not notice or did not consider the cost-shifting material enough to act. At two gigawatts, FERC noticed.

The Talen-Amazon Susquehanna case is the clearest data point in the market right now. Talen Energy and Amazon Web Services structured a 1,920 megawatt behind-the-meter arrangement at the Susquehanna nuclear plant in PJM. Other utility companies in PJM challenged it on the grounds that it would unjustifiably redistribute transmission costs to every other customer on the system. In April 2025, FERC upheld the rejection of that arrangement. By June 2025, Talen and Amazon had restructured to a 1,920 megawatt front-of-meter power purchase agreement.

The largest attempted BTM deal in PJM history failed and was restructured to FTM within weeks of the ruling. That is the governing precedent. Any project modeling at a scale comparable to a large data center or industrial complex needs to start from that fact.

When Front-of-Meter Co-location Works — FERC's Four Approved Interconnection Mechanics

FERC's April 2026 order (Docket ER26-1088-000) did not close the door on co-location entirely. It closed the large-scale BTM door and replaced it with a structured framework. Four interconnection mechanics are now approved for co-located loads in PJM.

Sub-full-capacity service lets a data center or large load use unused interconnection rights at an existing host generator. If the host generator has more interconnection capacity than it currently uses, a co-located load can step into that unused capacity without triggering a new full interconnection study.

Queue acceleration allows co-located loads to advance ahead of the standard interconnection queue at defined checkpoints. This does not eliminate the queue entirely, but it creates a mechanism to move faster than the standard multi-year process for a new greenfield connection.

Provisional Interconnection Service provides interim transmission service while the full interconnection study is being completed. For loads that cannot wait for the full process, this creates a path to start taking power while the permanent interconnection is finalized.

Surplus Interconnection Service is arguably the most commercially interesting of the four. It allows a load to use pre-studied, unused interconnection capacity at an existing generator's interconnection point — without triggering a new full study. Retiring coal plants and underutilized nuclear facilities in PJM's footprint are the primary candidates. If you are evaluating a host site with unused interconnection rights, Surplus Interconnection Service is the mechanic worth the most detailed scrutiny.

None of these four paths is frictionless. Each has conditions, qualification criteria, and timelines that require interconnection counsel and careful grid modeling. But they represent the legitimate options FERC has opened in place of the large-scale BTM path it has closed.

When Co-location of Load with Generation Becomes a Liability — Gross Demand Charges and Reliability Obligations

Here is where operators consistently underestimate their exposure.

Once connected under the co-location framework, PJM's tariff requires eligible customers taking Network Integration Transmission Service on behalf of co-located loads to pay on a gross demand basis — not a net basis. That distinction matters enormously. Your onsite generator may be serving ninety percent of your load. Your net grid withdrawal may approach zero in many hours. You still pay transmission charges based on your gross load, not your net withdrawal.

The same principle extends to ancillary services. Because co-located loads are synchronized with the transmission system and consume ancillary services — specifically regulation service and black start service — cost-causation principles require they pay for those services on a gross demand basis. Even in hours where net energy withdrawals are near zero, regulation service charges and black start charges apply to the gross load.

The operators who get hurt here are the ones who built their pro forma on a behind-the-meter netting assumption and then had to restructure into the co-location framework without updating their cost model. Near-zero net withdrawal does not mean near-zero grid charges. It means transmission and ancillary service charges on your gross demand, in every hour.

The reliability enforcement dimension is equally unforgiving. FERC's April 2026 order establishes that co-located loads carry explicit operational obligations — including load-shedding and generator-tripping requirements — that are less forgiving than standard transmission service. PJM's compliance framework is still being finalized, but the Commission has made clear that violations carry material consequences, up to and including termination of transmission service. One reliability failure can strand a billion-dollar asset. Confirm the current enforcement mechanism with your interconnection counsel before you finalize any co-location structure.

For context on how the broader category of on-site generation decisions intersects with standby charge exposure, the TEG post on C&I microgrids for industrial facilities covers the reliability and cost control architecture that underlies these decisions.

Vendor Pitches, Red Flags, and Questions That Smoke Out BS

The co-location space attracts vendors who built their models when the behind-the-meter netting assumption was still legally viable. Some of those models have not been updated. Here is how to find out which ones you are looking at.

The core red flag is a pro forma that shows transmission cost avoidance through BTM netting without explicitly addressing the Talen-Amazon precedent and the April 2026 FERC order. If the model was built before June 2025 and has not been updated, you are looking at a financial projection built on a regulatory assumption that no longer holds.

A second red flag is any vendor who presents gross vs. net demand charges as an implementation detail rather than a structural cost driver. If they are not explaining exactly how regulation service and black start charges apply on a gross basis at your co-located load, they either do not understand the tariff or are choosing not to explain it.

Questions to ask in the room:

• Has this model been updated to account for the April 2026 FERC order and the Talen-Amazon restructure? Can you show me where?
• Is the transmission cost projection based on gross demand or net demand? Walk me through the math.
• How does this model price regulation service and black start charges in hours when net grid withdrawal approaches zero?
• Which of FERC's four approved interconnection mechanics does this project use, and what are the qualification conditions for that mechanic?
• If we are relying on a host generator's unused interconnection rights, has Surplus Interconnection Service been formally evaluated?
• What does curtailment frequency look like under the non-firm tariff product during peak conditions, and what is the compound outage risk if the onsite generator and grid supply fail simultaneously?

What You Can Do This Week

1. Audit your current project model for the BTM netting assumption. If your cost projections show transmission avoidance through behind-the-meter netting, that assumption needs to be tested against the current regulatory framework before you commit further capital.

2. Get a gross vs. net demand analysis. Ask your interconnection counsel or project advisor to run your load profile against PJM's NITS gross demand requirement and quantify the ancillary service charges — including regulation service and black start — in hours where net withdrawal is near zero.

3. Map your project to one of FERC's four approved interconnection mechanics. Sub-full-capacity service, queue acceleration, Provisional Interconnection Service, and Surplus Interconnection Service each have distinct qualification conditions. Know which one your project relies on and what can disqualify it.

4. If you are evaluating a host site with unused interconnection rights, put Surplus Interconnection Service on the agenda with your interconnection counsel immediately. Retiring coal plants and derated nuclear units in PJM have pre-studied interconnection capacity. That capacity can be accessed faster than a new greenfield study — but only if the mechanics are structured correctly.

5. Confirm the current reliability enforcement framework with your interconnection counsel. PJM's compliance structure under the April 2026 order is still being finalized. Do not sign a co-location agreement based on a version of the enforcement framework that predates the final order.

The Bottom Line on Co-location of Load with Generation

Co-location of load with generation — specifically the behind-the-meter vs. front-of-meter configuration decision — is now a binary. The regulatory middle ground is gone.

Behind-the-meter at scale is a diminishing legal option in PJM. The Talen-Amazon restructure is the governing precedent. For large data centers, industrial loads, and green hydrogen producers evaluating this decision in PJM's footprint, the co-location framework FERC approved in April 2026 gives you four interconnection mechanics to work with — but none of them eliminate the reliability obligations, the gross demand charges, or the penalty exposure that comes with the framework.

The operators who come out ahead here are the ones who do the grid modeling before they sign anything, understand exactly what curtailment risk looks like during peak conditions, price gross demand charges correctly for ancillary services, and structure their load and generation under a mechanic they actually qualify for.

The operators who lose are the ones who built their pro forma on a behind-the-meter netting assumption that no longer holds — and discovered that after committing the capital.

Frequently Asked Questions: Co-location of Load with Generation

Q: What is the difference between behind-the-meter and front-of-meter co-location of load with generation?

A: Behind-the-meter co-location puts the generator on the consumer's side of the utility meter, serving the load through a private line with no interconnection agreement required. Front-of-meter co-location — what FERC now calls the BYOG (Bring Your Own Generation) model — puts the generator on the bulk grid side, interconnected directly to the transmission system, with the operator taking a transmission service product for the residual load the onsite generator does not cover. The key regulatory consequence is that behind-the-meter loads avoid transmission tariff exposure in theory, while front-of-meter co-located loads pay transmission and ancillary service charges on a gross demand basis regardless of how much the onsite generator covers.

Q: Can a large data center or industrial load still go behind the meter in PJM after the Talen-Amazon ruling?

A: For large loads in PJM, the behind-the-meter path is a diminishing legal option. FERC upheld the rejection of Talen Energy and Amazon Web Services' 1,920 megawatt BTM arrangement at the Susquehanna nuclear plant in April 2025, and Talen and Amazon restructured to a front-of-meter PPA by June 2025. That restructure is the governing precedent. Any large-scale project in PJM that models transmission cost avoidance through BTM netting needs to be tested against that precedent and the April 2026 FERC order before further capital is committed.

Q: What are FERC's four approved co-location interconnection mechanics under the April 2026 order?

A: FERC's April 2026 order (Docket ER26-1088-000) approved four interconnection mechanics for co-located loads in PJM: sub-full-capacity service (using unused interconnection rights at an existing host generator), queue acceleration (advancing co-located loads ahead of the standard queue at defined checkpoints), Provisional Interconnection Service (interim service during the full interconnection study), and Surplus Interconnection Service (using pre-studied, unused interconnection capacity at an existing generator's point without triggering a new full study). Each mechanic has distinct qualification conditions that require interconnection counsel to evaluate.

Q: Do co-located loads pay transmission charges even when net grid withdrawal is near zero?

A: Yes. PJM's tariff requires eligible customers taking Network Integration Transmission Service on behalf of co-located loads to pay on a gross demand basis, not a net basis. This means transmission charges apply based on the total load, not the difference between load and onsite generation. Beyond that, ancillary service charges — specifically regulation service and black start service — also apply on a gross demand basis because the co-located load is synchronized with the transmission system and uses those services regardless of net withdrawal. Near-zero net withdrawal does not mean near-zero grid charges.

Q: What is Surplus Interconnection Service and why does it matter for co-location projects?

A: Surplus Interconnection Service allows a co-located load to use pre-studied, unused interconnection capacity at an existing generator's interconnection point without triggering a new full interconnection study. It is the most commercially attractive of FERC's four approved co-location mechanics for operators looking to move faster than the standard PJM queue process. Retiring coal plants and underutilized nuclear facilities in PJM's footprint are the primary candidates, because they have existing pre-studied interconnection capacity that may be available to a co-located load.

Q: What is the binary choice FERC's co-location framework creates for large C&I operators?

A: Large C&I operators now face a clear binary: either accept the co-location framework with its four interconnection mechanics, gross demand charges, and reliability enforcement obligations — or commit to true islanding with no grid backup at all. The regulatory middle ground that allowed large-scale behind-the-meter configuration with grid backup has been eliminated. For most data centers, hospitals, and critical manufacturing loads, full islanding is operationally unacceptable, which means the co-location framework is not optional — it is the only viable path for loads that need both onsite generation and grid backup.

If you are an Indiana C&I operator working through a co-location project, an onsite generation evaluation, or any decision where the interaction between your load and the grid is driving cost and risk, the TEG Energy Decision Blueprint is built for that. It is available to qualified Indiana C&I operators spending five figures or more on electricity each month. You can also watch this episode of The TEG Podcast on YouTube for the full breakdown of this topic with Daniel's direct commentary on the Talen-Amazon case and the April 2026 FERC order.