Summary

  • NERC says preliminary results show the foundational computational-load standards passed an initial ballot; it still has to validate the result and publish the comment report.
  • The vote is not FERC approval, an effective date, a registration decision or proof that any particular data-centre project may interconnect and operate.
  • The investable response is to preserve design and evidence optionality now, while keeping final recurring compliance cost outside the base case until later clocks resolve.

The decisive word in NERC’s 19 September announcement is not “passed”. It is “preliminary”. That qualifier does two jobs at once. It tells data-centre developers, utilities and investors that a standards perimeter around computational loads is becoming more plausible. It also prevents them from treating an unvalidated initial ballot as an enforceable obligation.

This is a familiar market error: a procedural milestone is compressed into a legal result, and a legal result is then compressed into project readiness. Here those are different claims, controlled by different institutions and supported by different evidence.

NERC said the foundational standards developed under Project 2026-02 had cleared an initial ballot on preliminary results. It expected to validate the vote and issue a report of comments the following week. The announcement also acknowledged that important stakeholder perspectives remained unresolved. Those facts make the ballot a meaningful drafting signal, not a finished rulebook.

Nine clocks sit between a ballot and operating compliance

The first clock is the preliminary result itself. The second is validation and the publication of comments. The third covers disposition of those comments, any further ballot and final adoption within NERC’s standards process. Only then does the planned federal track become concrete.

The fourth clock is filing. FERC’s July action under section 215(d)(5) of the Federal Power Act directed NERC to submit reliability standards and Rules of Procedure revisions, including criteria for a Computational Load Entity, by 31 December 2026. NERC has described a filing in the final quarter. A direction to submit a package is not advance approval of that package.

FERC’s disposition is the fifth clock. The Commission can approve, require changes or otherwise act on what NERC files. An implementation plan and effective date form the sixth. Neither was created by the September ballot announcement.

The seventh clock is entity-specific registration under whatever criteria survive the process. The eighth belongs to the site: interconnection studies, an agreement, funded upgrades, tested models, commissioning and permission to energize. The ninth is operating evidence after applicable requirements take effect—records, disturbance performance, communications and an auditable compliance practice.

These clocks can influence one another, but none substitutes for the next. A company may eventually meet a registration threshold and still lack an executable interconnection path. A project may have utility approval to energize and still need new compliance arrangements later. Market analysis should state which clock a piece of evidence actually moves.

The Level 3 Alert is evidence, not the standard

NERC’s 4 May Level 3 Alert asked existing registered reliability and transmission entities to take Essential Actions while registration and standards work continued. Its subjects—model quality, studies, instrumentation, commissioning, operational coordination, protection and controls—show where technical cost is likely to accumulate.

But the alert and the Project 2026-02 standards are different instruments. The alert gives current entities immediate risk-mitigation work and supplies a technical record. It is neither a final Project 2026-02 standard nor a FERC approval order. Treating it as the final text would overstate legal certainty; ignoring it would understate engineering notice.

The project’s Standards Authorization Request is deliberately broad. It considers data sharing, interconnection practices, modelling, protection, high-resolution monitoring, commissioning and operating response. It also leaves room for later work across existing standards families. The programme is therefore better read as a phased reallocation of responsibility than as one threshold taking effect on one date.

Registration will not be an interconnection permit

The proposed Rules of Procedure would add a Computational Load Entity category. That matters because registration determines who enters the compliance system and which approved requirements can apply. It does not certify a site’s queue position, system impact, upgrade funding, equipment settings, commissioning result or operating performance.

Thresholds discussed in an April proposal may also change. They should not be modelled as immutable eligibility tests. A responsible investment case separates the probability of eventual registration from the probability and timing of a particular project’s interconnection.

Price optionality, not a finished rule

The low-regret expenditure is not a guessed compliance programme. It is the preservation of choices that become expensive to recover after procurement. Projects can reserve panel and network capacity for high-resolution measurement; specify time synchronisation and event recording; require vendors to deliver usable dynamic models and settings data; retain access to firmware and test records; schedule staged commissioning; and assign 24-hour operating contacts.

Contracts deserve equal attention. Model rights, OEM cooperation, data ownership, retest obligations, change control and responsibility for corrective work can determine whether a technical requirement costs weeks or quarters. A cheap specification that withholds the model needed by a transmission planner is not cheap once the site is waiting to energize.

The right base case therefore has two columns. One funds foreseeable design optionality and evidence production now. The other holds final recurring compliance staffing, a definitive registration threshold and a binding effective date as unresolved until NERC and FERC create them. That is neither delay nor regulatory arbitrage. It is disciplined staging by evidence.

Sources