Edition #2 - April 2026
Welcome back to Behind the Meter. If you're new here, this is where I share what's caught my attention in energy every couple of weeks - the stuff that actually matters for how we manage utilities, procure power, and plan portfolios.
Two weeks ago, the largest grid operator you've probably never heard of quietly became the first in the US to span two interconnections. On the same day, Australia's energy rule-maker was fielding submissions on whether data centres should have to prove they won't destabilise the grid before they're allowed to connect. These two events, on opposite sides of the Pacific, are telling the same story.
The Take: When the Load Becomes the Problem
For decades, grid planning has been a supply-side exercise. Build generation, connect it, keep the lights on. The technical standards, the market rules, the interconnection queues - they all face one direction: towards generators. Loads just showed up and consumed.
That era is ending. The AEMC's draft rule on technical standards for large inverter-based loads, released in March and open for submissions until 7 May, is one of the clearest signals yet. For the first time, the NEM is proposing that large loads, specifically data centres above 30 MW, must meet disturbance ride-through requirements. They need to stay connected during voltage and frequency events and recover power within defined timeframes. The same expectations we've historically reserved for generators.
This isn't bureaucratic box-ticking. It's a fundamental rethinking of what it means to participate in the grid. A single hyperscale data centre can draw more power than a small city. When that load trips during a grid disturbance, the cascading effect on frequency and voltage is real and measurable. The AEMC is rightly asking: if you're going to consume at that scale, you need to be a responsible grid citizen.
What's interesting is the global convergence. The proposed Australian standards largely align with what Texas, Ireland, and Finland are already implementing or proposing. FERC has simultaneously ordered PJM to create new tariff products for co-located generation and large loads. And when SPP expanded its RTO footprint on 1 April to cover 732,000 square miles across 17 states, becoming the first RTO to span two interconnections, it did so partly to manage the transmission planning complexity that these new load patterns create. The estimated $200 million in annual regional benefits tells you everything about the value of coordinated planning at this scale.
For anyone managing a multi-site utility portfolio, the implications are practical. Network charges are going to be reshaped by where large loads cluster. Demand patterns are shifting in ways that make interval data more valuable, not less. And the businesses consuming at scale are about to face connection requirements that look a lot more like what generators deal with today.
I'd encourage anyone in the consulting or procurement space to read the AEMC's draft determination closely. The 30 MW threshold matters, but the direction of travel matters more. The grid is being redesigned around the assumption that large loads are active participants, not passive consumers. That changes how you advise clients, how you model costs, and how you think about risk.
Quick Hits
NSW just approved $52 billion in data centres. Where's the power coming from? The NSW Government endorsed 15 data centre projects through its IDA pipeline on 27 March, worth a combined $51.9 billion. Another $40.7 billion of proposals were knocked back as premature. To put the scale in context, connection requests for NSW data centres over the past 18 months have totalled more than 10 GW - that's over half the state's peak demand. The grid wasn't designed for load growth at this velocity. If you're advising large energy users in western Sydney or the Hunter, the network investment needed to service these facilities is going to reshape charges for everyone on the same feeders. This is the demand-side problem The Take is about, playing out in real time.
Half the NEM's electricity is now renewable. The pricing implications are brutal. In Q4 2025, renewables and storage crossed 51% of NEM generation for the first time. South Australia hit negative prices in 46% of all dispatch intervals that quarter - a NEM record. Victoria wasn't far behind at 43%. This isn't a blip, it's structural. Daytime wholesale prices are heading towards zero in large parts of the grid. The AER's new Solar Sharer Offer, launching 1 July, essentially acknowledges this by requiring retailers to offer three hours of free midday electricity. If your clients' contract structures are still benchmarked against 2023-24 wholesale averages, they're overpaying in the middle of the day and underhedged for the evening ramp.
Australia's energy regulator goes solo. On 1 July 2026, the AER formally separates from the ACCC to become a standalone Commonwealth entity. The AER gets its own staff, its own budget, and its own accountability framework. For energy businesses, this means a regulator that's more focused and potentially more assertive on compliance, billing standards, and market conduct. If your compliance processes assume the old structure, now's the time to review them.
The grid is being rebuilt in real time, and the rules are following. Whether it's data centres proving they can ride through a frequency event, $52 billion of new load lining up behind a single state's connection queue, or a regulator cutting the umbilical cord to stand on its own, the theme is the same: the energy system is growing up, and it expects its participants to do the same.
I'll be watching the AEMC submissions closely. If you're preparing one, or if you're already seeing these load dynamics reshape your clients' cost profiles, I'd like to hear about it.
Until next fortnight,
Cohen
Know someone who'd find this useful? Send it their way. They can subscribe at utilified.com. You can also find me on LinkedIn.
If something in here sparked a thought, hit reply. I read every one.