Breaking

SoftBank and OpenAI Launch AI Cyber Defense 'Patching as a Service' in Japan

June 21, 2026 at 15:25 EDT

Security
Sales & Business
AI Agents

SoftBank Group and OpenAI announced on June 16, 2026, at a corporate event in Tokyo the launch of "Patching as a Service," an AI-powered cybersecurity offering. Delivered through their joint venture SB OAI Japan GK, it will roll out progressively over the year to roughly 3,000 Japanese companies operating critical infrastructure such as airports, power and transport.

June 16, 2026 · Tokyo · SoftBank × OpenAI

SoftBank and OpenAI launch "Patching as a Service" to defend Japan's critical infrastructure

An AI-powered cybersecurity service — from vulnerability assessment to remediation — aimed at power grids, airports, banks and telecoms. Masayoshi Son calls AI-driven attacks "the greatest crisis since the arrival of the Black Ships."

~3,000

critical infrastructure operators targeted across power, transport & finance

50:50

SoftBank–OpenAI JV (SB OAI Japan GK), established Nov 2025

End 2026

target for full-scale rollout after the June 16 launch

130

companies (~200 executives) at the joint Tokyo event

How the service works — an end-to-end loop

Vulnerability
identification

→

Risk
analysis

→

Remediation
recommendations

→

Planning &
implementation

Future expansion is planned beyond patch delivery into threat response and system improvement. Pricing, benchmarks and supported environments are not yet disclosed.

Son's warning

"The greatest crisis since the arrival of the Black Ships."

A reference to Commodore Perry's 1853 arrival, which forced Japan to abandon isolation. The message: generative AI can dramatically expand the speed and scale of cyberattacks — while Japan's infrastructure defenses have long lagged.

The case for

The threat of AI cyberattacks is real and growing. The historical metaphor is seen as a timely way to underscore the urgent need for AI-powered defenses — backed by exclusive access to OpenAI models for Japanese firms.

The open questions

Just announced, so real user evaluations barely exist. Pricing, benchmarks, supported environments and concrete use cases remain undisclosed — the warning may also read as dramatic.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.