Sustainability

The Grid Price Behind Every Head of Lettuce

Philadelphia's power grid just got a lot more expensive to run on, and I look at what that means for growing food indoors here.

By Harry Hayman 7 min read

Twenty eight dollars and ninety two cents. That is not what anything costs, it is what one megawatt-day of grid capacity was worth two years ago, before I had thought hard about what a data center in Virginia has to do with a flat of lettuce growing under LEDs in Philadelphia. Today that same megawatt-day costs $329.17. Eleven times more, on a grid I did not vote to change and neither did anyone else running an indoor farm on it. I read the auction report the way I would read a utility bill that looked wrong, twice, slowly, checking the math myself before I believed it. It held up. The number is real, the year it covers already started on June 1st, and it is already showing up in what Philadelphia pays for power. What follows is what that number means for an operation like ours, and why the honest answer was never going to be a complaint.

The Number PJM Just Set

PJM Interconnection runs the wholesale electricity market for a stretch of the country that runs from Chicago to the New Jersey shore, and Pennsylvania sits inside it. Each year PJM holds what it calls a Base Residual Auction, a forward market where power plants and other resources bid to be paid for standing ready to supply electricity years in advance. The result becomes the capacity price every utility in the footprint has to pay, and every utility passes that cost through to the people it serves.

On July 22nd, 2025, PJM announced that its auction for the delivery year running June 1st, 2026 through May 31st, 2027 had cleared at $329.17 per megawatt-day across the entire region, the price cap set by federal regulators. Utility Dive covered it under a headline that said exactly what it was: another record, a 22 percent jump from the year before. PJM’s own release put a number on the prior clearing price, $269.92 per megawatt-day, and RTO Insider did the subtraction for me: fifty nine dollars and twenty two cents more, a 22 percent increase.

I went back one more year, to the 2024/2025 delivery year auction results PJM itself published, which cleared at $28.92 per megawatt-day. Divide $329.17 by that number and the answer is 11.4. Eleven times, roughly, in two years. I could not find anyone who had printed that specific division, so I did it myself, twice, because a number that large deserved to be checked before I built an argument on top of it.

Everyone Points At The Same Thing

I looked for an account of this auction that blamed something else, and I could not find one that left data centers out. PJM’s own release said the demand forecast for the region grew by 5,250 megawatts year over year, and called that growth almost entirely driven by data centers. Utility Dive’s reporting on the auction put new peak load at roughly 5,500 megawatts, mainly from data centers. And PJM’s own market monitor, in a separate piece, went further and called data center demand the primary reason capacity prices are where they are.

I want to be precise here rather than dramatic. Data centers are not the only name in the account. PJM has also pointed to electrification and general economic growth, and the prior year’s jump was tied partly to power plant retirements and a change in how the grid operator credits capacity toward reliability. None of that changes the headline number. It changes how I talk about it. This is not a story about one villain plugging in more servers. It is a story about a demand curve that grew faster than the supply behind it, and a controlled environment farm sits nowhere on either side of that curve. We do not bid into PJM’s auction. We just pay whatever it clears at.

What This Has To Do With Philadelphia

PJM is not an abstraction to me. It is the grid Philadelphia has been part of since 1927, when Philadelphia Electric Company, the utility now known as PECO, joined two neighboring utilities to form the power pool that eventually became PJM, according to the grid operator’s own history. Almost a century later, the connection still runs in one direction: whatever the auction clears at becomes part of the bill.

The Philadelphia Tribune reported that PECO customers saw their bills rise by about five dollars a month starting June 1st, tied directly to capacity costs set at PJM auctions and being phased in over time. Five dollars sounds small next to $329.17. It stops sounding small when you are running lights over trays of greens around the clock, and the article is clear that even more recent record prices have not finished phasing in yet.

Nobody who runs an indoor farm in this city sat in on that auction. Nobody voted on the demand forecast behind it. The bill still arrives.

The Cost Line Nobody Votes On

Every operator in controlled environment agriculture already knows this math, even if the industry rarely says it plainly in public. Lighting alone can draw more than half of what an indoor farm pulls from the grid, with climate control close behind it. One industry analyst told Agritecture that the cost of lighting has been the single biggest production expense in the sector for two decades running, and better technology has not changed that ranking. Payroll is still the largest line on most of our budgets. Electricity is the largest line that is not a paycheck, and unlike payroll, we do not set that price. PJM’s auction does.

That is the part of this story that never makes the industry panels. We talk about crop yield, about light spectrum, about water recycling. We do not talk enough about the fact that our single biggest controllable input is priced inside a market built around thirteen states and roughly sixty seven million people, most of whom will never taste a leaf we grow. A hyperscale data center can absorb an auction result like this one and keep building. A break even indoor farm cannot outbid that kind of demand for the same megawatt, and should not have to try. The honest move is not to compete for cheaper power. It is to stop being surprised by the price.

What Concrete Greens Does With That

So here is what changes on our end, and it is not a mission statement, it is an operating decision. We treat the location of a facility as an energy decision first and a real estate decision second, because the megawatt-day price is now a bigger variable than the square footage price. We schedule the heaviest draws, the lighting cycles and the climate systems that can shift by a few hours without hurting a crop, into the parts of the day when demand on the grid is lower and cheaper. And we build the electric bill into the unit economics of every crop before we grow it, not after the invoice arrives, because a 22 percent capacity step is exactly the kind of number that turns a marginal crop into a losing one if nobody planned for it.

None of that is new to this post. It is the same approach that has kept Concrete Greens honest since we started: a revenue generating operation built to complement the farms already growing food outside city limits, not replace them, producing greens locally and year round, and directing what we do not sell to people in this city who need food, not into a growth projection. An operation built to break even can absorb a demand shock like this one and keep operating. An operation built on the assumption that costs would stay flat cannot.

PJM has already held the next auction, for the 2027/2028 delivery year, and it cleared at a new record too: $333.44 per megawatt-day, the price cap again, the third record in a row. I expect the number after that to be higher still. The instruction I keep coming back to, for myself as much as for anyone reading this, is simple: build the real cost of power into the plan before the next auction forces you to, not after.

Sources and references

Back to all of Harry's writing