★1 is a system that confirms the minimum requirements for a single product (such as changing the default password and providing update methods), and there are no continuing obligations after acquisition under the guideline. On the other hand, what determines the scale of damage is the capacity that can operate simultaneously from one distribution server, one administrator authentication, and one control plane. A label for the product alone does not evaluate that. If the underwriting and review side confirms these five items, it can reflect risk in pricing without relying on the outward appearance of the label, and it also gives facility owners and contractors an incentive to respond.
Explanatory video (approx. 6.5 minutes)
Other language versions: Japanese / English / Simplified Chinese / Playlist / Second installment (concentration of supply and target structure, about 8 minutes): Cyber risks hidden in power grid labels / How to submit comments: Submit your comments by November 6
1. The content of the requirements and the matters the requirements do not assess
The proposed revision requires JC-STAR★1 label acquisition for devices with IP communication functions that interconnect with the grid from April 2027 onward, including PCS, battery storage, EV chargers, and monitoring devices. In July 2026, IPA, the scheme operator, replied in writing to our inquiry as follows (the original is in the response dataset).
JC-STAR does not introduce the concept of “boundary defense” for the “equipment” it covers.
The effect of a labeled product does not change depending on where it is installed.
“Conformity label obtained” does not mean “safely installed and operated.”
In the same month, the Agency for Natural Resources and Energy also replied in writing that there are no documents comparing boundary defense or alternative measures such as IEC 62443, and no documents evaluating the impact on existing equipment. In other words, there are no materials on the system side showing the size of the label’s effect, nor materials showing safety at the facility level.
2. Five items we want to confirm
Presented in a form that can be used directly as questions to equipment owners, installers, and manufacturers.
- Capacity (kW) and number of units belonging to the same firmware distribution system If the distribution server or signing key is compromised, updates will be delivered simultaneously to all units through the legitimate path. One incident will not stop at one site. Question: From which manufacturer and which distribution system do this facility's PCS and battery storage receive updates? Approximately how many units and how many kW in Japan belong to the same system?
- Upper limit on the capacity that can be operated simultaneously with a single administrator credential If there is no upper limit, a single credential leak can move the entire control plane. Question: Can output control, shutdown, and setting changes be performed from the manufacturer's cloud or maintenance remote connection? Is there an upper limit on the number of units and total capacity that can be operated simultaneously with one credential?
- Whether the minimum operating point reachable by remote commands and the limit on output ramp rate are fixed on the device side If they are fixed, even if compromised, the scale of grid disturbance is capped by the device. If they can be changed remotely, there is no cap. Question: Can remote commands reduce output to 0%, or is a certain percentage of rated output the lower limit? Is the maximum ramp rate fixed in the device, or can it be rewritten remotely?
- Behavior when communication is lost (whether the last command value is retained or the device returns to a safe default operating mode) Behavior during recovery affects secondary damage, such as other equipment dropping out due to a frequency dip. Question: What does the device do when external communication is lost? Is that described in the specifications?
- Whether the facility is configured so that control points cannot be reached from external networks Boundary defenses using gateways or zone segmentation determine device-level reachability, which product labels do not evaluate. Question: Is there an in-house device at the site network boundary? Are the communication destinations the device can reach externally listed? Are the system operator's command path and the manufacturer's maintenance path separated?
3. Differences from overseas systems
| System | Assessment unit | Ongoing obligation |
|---|---|---|
| JC-STAR★1 (Japan, revised draft) | Product-level label (self-declaration of conformity) | No ongoing obligations after label acquisition under the guidelines |
| EU Cyber Resilience Act (CRA, Regulation (EU) 2024/2847) | Product + manufacturer | Early warning of exploited vulnerabilities within 24 hours, and an obligation to provide updates during the support period (reporting obligation applies from September 11, 2026) |
| U.S. NERC CIP-002-5.1a | Generating capacity controlled by one cyber system (1,500 / 3,000 MW) | Management requirements and audits according to capacity |
| UK PSTI Act (effective April 29, 2024) | Product + manufacturer | Prohibit initial passwords, vulnerability disclosure contact point, and clear update support period |
Whether the unit of evaluation is the product itself, or the facility, operator, or capacity. Whether there are obligations that continue after acquisition. Japan’s revised proposal differs from others on these two points, and unless the underwriting and review side fills the gap, it becomes a blank that no one will look at.
4. Our position
We are not opposed to making the requirements mandatory themselves. Our concern is that by treating the presence or absence of the label alone as meaning "countermeasures completed," the five items above will be left out of the review. If these five items are included in underwriting conditions, special clauses, and loan screening items, facility owners and contractors will have an incentive not only to obtain the label but also to choose a configuration that protects the equipment. This is a measure that can be started on the market side without waiting for a revision of the system.
The content on this page has been sent as information to the General Insurance Association of Japan, the corporate divisions of major non-life insurers, and the renewable energy financing desks of government-affiliated financial institutions. The date of sending and whether it was received are noted in the ledger.
What this document does not say
We are not saying that the risk of specific products or operators is high. Nor are we saying that ★1 is ineffective. Eliminating default passwords and providing update methods are meaningful against intrusion into a single device. We are also not proposing specific levels for insurance rates or loan conditions. What we are stating is only the fact that there are items not evaluated by the product label, that those items determine the scale of damage, and the competent ministry’s response that no materials exist on the system side to fill that gap.
Source
- Agency for Natural Resources and Energy, "Draft Revision of the Grid Connection Technical Requirements Guideline Related to Securing Power Quality" and public comment instructions (published October 6, 2026, e-Gov case number 620340012, submission deadline November 6, 2026)
- Written responses to our public questions from the Agency for Natural Resources and Energy, Electricity Infrastructure Division (July 13, 16, and 30, 2026), and from the IPA Security Center, Technical Evaluation Division (July 19, 25, and 26, 2026). The original text is in the answer dataset (CC BY 4.0)
- NERC CIP-002-5.1a; Regulation (EU) 2024/2847 (Cyber Resilience Act); UK Product Security and Telecommunications Infrastructure Act 2022
- Our "There is no basis for thinking we will not be targeted" "From Cybersecurity to Cyber-Resilience" "Verification of the proposed revision"
益田 周防海(Suomi Masuda)/ I-S3 Co., Ltd. Based on public materials and received response documents. Not advice on individual insurance or financing decisions.