2026.10.20Issue #319

Grid Codes Are Turning Toward Data Centers

A term searched only 111 times a month still triggered two government meetings.

Business

Oswarld (Kwangseob Ahn)· Writer & editor of INLEVEL9

What a Search Volume of 111 Tells You — and What It Doesn’t

The Korean term “grid code” gets searched an average of 111 times a month (as of the September 2026 observation). That’s search count, not number of unique searchers, so the actual number of people looking this up could be lower. But if you look at the daily data, there’s a clear peak on September 18 — the highest point in the observation window. Normally the term barely registers, then on one day it spikes.

Public records don’t show what announcement or meeting happened that day. So I can’t just point at a single spike and call it “a trending term right now.” Instead, if you trace how this term has shown up in official announcements and trade publications throughout the year, a pattern emerges. The scope of grid code discussions is expanding — from power generation facilities to large electricity-consuming facilities like data centers.

What Kind of Rule Is a Grid Code

A grid code isn’t the name of a single document. It’s an umbrella term for the technical and operational standards that facilities and operators connecting to the power grid have to follow. In Korea, this includes the Ministry of Climate, Energy and Environment’s “Power System Reliability and Electric Quality Maintenance Standards,” the Korea Power Exchange’s “Electricity Market Operation Rules,” and Korea Electric Power Corporation’s “Transmission and Distribution Facility Usage Regulations.”

The reason these rules exist is that the grid has to match production and consumption at every single moment to maintain frequency and voltage. Coal and nuclear plants generate electricity by spinning massive rotating turbines, and the inertia of those turbines absorbs sudden fluctuations to some degree. Solar, wind, and battery systems connect to the grid through inverters¹. An inverter isn’t a rotating machine — it’s a power electronics device — so you have to define, through settings and rules, exactly how it should respond when voltage and frequency waver. That’s why the more renewable energy comes online, the more the grid code has to be rewritten.

In July, Korea Started Rewriting the Rules for Power Generation

On July 21, 2026, the Electricity Committee under the Ministry of Climate, Energy and Environment convened the “1st Grid Code Special Committee.” This body is a reorganization of the former “Power System and Reliability Special Committee.” The first meeting had four agenda items: improving Korea’s domestic grid code for inverter-based resources, safety and grid-connection standards for battery energy storage systems (BESS), large-scale blackout prevention and system restoration frameworks, and the mid-to-long-term direction of future grid operation standards. Behind all this sits the government’s goal of deploying 100GW of renewable energy capacity by 2030.

There are numbers showing that the rules haven’t kept pace with what’s happening on the ground. At a National Assembly discussion that same month, the Ministry of Climate Energy and Environment reported that renewable energy curtailment rose from 27 incidents in 2024 to 82 in 2025, with the amount of curtailed power jumping from 12.4GWh to 109.4GWh. Power market participants reportedly grew from 19 companies in 2001 to roughly 8,000 in 2026. The participant pool has expanded enormously, but the problem is that the system for setting the standards these participants must follow — and for verifying whether they actually comply — hasn’t grown to match. In ongoing government discussions about establishing a new Electricity Supervisory Service, improving the grid code and overseeing its enforcement have also been floated as core responsibilities.

This terminology didn’t just appear this summer out of nowhere. Back in April, at a power governance forum, the head of KEPCO’s Grid Planning Office already named “who writes and revises the grid code, and how roles get divided among institutions” as a key challenge. What changed in July is that this discussion moved into a formally named committee.

August: Regulatory attention turns to data centers

August 27 brought a meeting of a different flavor: the first session of the “AI Data Center (AIDC) Research Council,” chaired by Vice Minister Kwangseob Ahn—wait, chaired by Second Vice Minister Ho-hyun Lee of the Ministry of Climate, Energy and Environment (Korea’s second vice-minister post). At this meeting, the government announced plans to establish a dedicated grid code for AI data centers within the year. The idea is to define the minimum technical requirements a data center must meet to connect to the grid and draw power stably.

The government’s stated reason is the peculiar shape of AI data centers’ power demand. During training and inference, power draw can swing by megawatt-scale amounts on 1/1,000-second (millisecond) timescales, and hyperscale facilities can demand gigawatt-scale power. Just as instability in a power plant’s output can destabilize the grid, the explanation goes, demand-side swings at this speed can also disrupt frequency and voltage.

Among the things the Council said it would examine are facilities data centers already have. AI data centers run energy storage systems (ESS) and uninterruptible power supplies (UPS) to protect sensitive GPUs and maintain power quality. The government said it would review connection strategies and institutional arrangements that put this equipment to work stabilizing the grid, without compromising its original function. In other words, equipment installed purely to protect a company’s own servers could end up counting as equipment that contributes to grid stability.

This discussion is tied to the government’s broader AI data center vision. In June, the government unveiled a plan for 18.4GW of AI data center capacity by 2035. In a September 21 feature, the Energy Newspaper (Energy Shinmun) noted that this figure is merely an investment and build-out target—it doesn’t mean power supply contracts or grid connections have actually been finalized. The same article argued it’s premature to treat what the Council is reviewing as a done deal. Who the rules will apply to, when, how performance will be verified, and who will bear the cost—none of this has been decided yet.

America started by building an incident record

Overseas power authorities are grappling with the same problem. The North American Electric Reliability Corporation (NERC) reviewed 26 cases on the Texas grid, from January 2023 to September 2025, where large electronic loads failed to ride through fault conditions. Even when a transmission line fault was cleared normally, cryptocurrency mining facilities lost 17–95% of the power they were drawing within a few milliseconds. It was a shock on the demand side — the mirror image of a power plant suddenly tripping offline.

In May 2026, NERC issued guidelines for mitigating large load risk². The guidelines aren’t binding, but they address large-load operators and equipment manufacturers who hadn’t previously been subject to regulation. They recommend accounting for ride-through³ performance — the ability to stay connected rather than disconnect during a grid fault — and providing accurate information on load composition, protection settings, and model specifications.

In July, the US Federal Energy Regulatory Commission (FERC) went a step further. It ordered NERC to submit, by December 31, 2026, mandatory reliability standards addressing the grid-interconnection risks of “computational loads” such as AI data centers and cryptocurrency mining operations. What began as a recommendation is now moving toward becoming a mandatory standard.

Korea’s push to finalize an AIDC grid code within the year overlaps in timing with this trend. The Ministry of Climate, Energy and Environment has also noted that major countries abroad are moving to establish grid-stabilization performance requirements for large load facilities. Whether Korea’s own standards will become mandatory, as in the US, or remain technical criteria used only in interconnection negotiations, has not yet been announced.

How to Read a Small Search Volume

A number like 111 searches a month looks negligible on its own. Search data alone can’t even tell you who was searching. But overlay it against the official calendar for the same period, and the reading changes. In July, a specialist committee called the Grid Code launched. In August, came an announcement that data-center-specific standards would be drawn up within the year. In September, an energy trade publication identified grid interconnection as the variable that would determine success or failure for data center projects. The policy discussion has started, but the standards document itself hasn’t been published yet.

Technical terms at this stage are bound to show low public search volume — the pool of people actually searching is likely narrow: operators preparing for interconnection negotiations, equipment manufacturers, policy officials. For terms like this, the size of the spike matters less than what was on the official calendar around the date it occurred, and whether the subject matter the term covers has shifted. In the case of Grid Code, data centers entered its scope over the course of this past summer. What triggered the spike on September 18th remains a question still to be confirmed.

Oswarld’s Lens

What catches my attention in this trend is that the position of the data center is shifting. Until now, data centers were, from the grid’s perspective, customers who bought enormous amounts of electricity. What they needed to secure was capacity, rates, and timing of supply. Once a demand-side grid code exists, a data center becomes a participant that must prove how its own equipment responds at the moment of a fault, just to connect to the grid.

This change can reorder a company’s priorities. If the sequence used to be: pick a site, bring in servers, then negotiate the connection with KEPCO (Korea Electric Power Corporation), the grid code was paperwork you encountered at the very last step. But once the specifications of UPS and ESS, and the way load fluctuations are smoothed out, become conditions for connection, the grid code turns into an input you need at the earliest stage of design. Energy Newspaper also expects this direction to increase operators’ need to factor in grid requirements from the very start of design.

So the first thing to look at in the AIDC grid code expected by year-end isn’t the technical figures — it’s the scope of application. Depending on what scale, which installations by what connection date, and who verifies compliance and how, this standard could become a document that lets businesses predict their project timelines — or it could become just another layer of uncertainty.

Closing

What a term searched 111 times a month told us was the moment a rule had just begun to be applied. Once the standard document arrives and the first real-world case shows up, the people searching for this term will look different from who they are now.

💬 In the data center, renewable energy, or factory projects you’ve worked on, Reader, when did you first check the grid interconnection conditions? Was it before the site was chosen, or after? Let us know in the comments.

Your take shapes the next issue

What resonated most in this issue, or where has your experience been different?

References & Further Reading

The author is Oswarld (Kwangseob Ahn). Current roles: Adjunct Professor at Sejong University, Strategy Consultant at INLEVEL9. Career, research, books, and recent work are kept current on the About page. Latest · July 2026: HEMA-2: A Consolidation-Aware Tri-Memory Architecture with Multi-Channel Scheduling for Lifelong Conversational AI.

📝 Glossary

Footnotes

Inverter-Based Resource (IBR): A power source—like solar, wind, or battery storage—that converts direct current or variable-frequency power into grid-compatible electricity through power electronics (an inverter). Unlike spinning generators, its response to grid fluctuations is determined by control settings rather than physical rotation. ↩
NERC Reliability Guideline for Large Load Risk Mitigation: The “Reliability Guideline: Risk Mitigation for Emerging Large Loads,” published by the North American Electric Reliability Corporation in May 2026. It’s a recommendation document aimed at reducing the planning and operational risks posed by large loads such as data centers and crypto-mining facilities, and it carries no binding force. ↩
Ride-through: The capability of equipment to stay connected to the grid for a set duration during momentary voltage or frequency disturbances, rather than disconnecting immediately. ↩