AIToday
AI Regulation & PolicyTop Companies' AI MovesTop Companies AIPublished: Aug 22, 2026, 06:30 JST2 min read

7 Japanese carriers standardize AI-era low-latency network design at ITU-T

7 Japanese carriers standardize AI-era low-latency network design at ITU-T

Key takeaway

  • Seven Japanese telecom and equipment makers have won international standardization of design rules for all-photonics networks—a low-latency, energy-efficient architecture underpinning AI services.

  • The ITU-T standard, finalized August 2026, allows different carriers to interconnect reliably without proprietary lock-in.

  • A second, more detailed architectural standard is targeted for October 2028.

3 Key Points

  1. What happened

    KDDI, NTT, 1FINITY, NEC, Rakuten Mobile, Oki Electric, and Sumitomo Electric led the creation of design guidelines for low-latency, energy-efficient networks supporting all-photonics architecture. The framework was adopted as ITU-T Recommendation Y.3335 on August 13, 2026.

  2. Why it matters

    The guidelines enable different carriers to build compatible networks for AI and remote sensor processing without relying on a single vendor or organization. This removes a major barrier to global interoperability—until now, carriers outside existing consortia lacked a shared technical blueprint for all-photonics networks (APN).

  3. What to watch

    The seven companies plan to advance a detailed functional architecture recommendation by October 2028 at ITU-T, building on their domestically developed network technology and standardization expertise to drive global APN adoption.

Ask the AI about this article →

Context & Analysis

The standardization of ITU-T Recommendation Y.3335 represents a shift from proprietary consortia governance to open international standards for next-generation networks. Until now, all-photonics network knowledge had accumulated within specific organizations—notably IOWN Global Forum, which coordinates shared technical documents and optical network federation technology across participating carriers. However, this approach created a participation barrier: carriers and governments outside the consortium lacked a neutral reference design.

The seven Japanese companies—spanning carriers (KDDI, NTT, Rakuten Mobile), equipment makers (NEC, Oki Electric, Sumitomo Electric), and Fujitsu's network unit (1FINITY)—moved to codify their accumulated APN expertise into a framework that belongs to no single entity or implementation method. By embedding this knowledge into an ITU-T recommendation, they have created a public-domain baseline that any carrier or vendor can reference, lowering the cost and complexity of building multi-carrier APN services. The framework is grounded in research conducted under Japan's National Institute of Information and Communications Technology (NICT) Beyond 5G / 6G fund, ensuring it reflects real-world engineering experience rather than theoretical requirements.

FAQ

What is all-photonics networking (APN)?
APN is a network architecture delivering high capacity, low latency, and low power consumption. It enables real-time processing of data from remote sensors and robots, supporting next-generation services like physical AI.
Why was this international standard needed?
Global interoperability and widespread adoption required a common, vendor-neutral design framework. Until now, only participants in specific consortia (such as IOWN Global Forum) had access to shared technical specifications; carriers and regions outside those groups lacked a standardized blueprint.
What comes next?
The seven companies are working toward a more detailed functional architecture recommendation at ITU-T, targeted for October 2028, to further support global APN deployment.
Top Companies AIRead Original Article

Get the latest AI Regulation & Policy news every morning

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytime

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleDeepSeek launches multimodal model matching Anthropic's Opus 4.8