Masayoshi Son warns of an AI "Black Ship Invasion": SoftBank teams up with OpenAI to launch new enterprise cybersecurity protection service

SoftBank Group President Masayoshi Son announced on the 16th that they will collaborate with OpenAI through a joint venture, SB OAI Japan GK, to launch a cybersecurity service called "Patching as a Service," targeting approximately 3,000 critical infrastructure companies in Japan, including finance, power, and transportation sectors.
(Background: SoftBank pledged OpenAI shares in exchange for a $10 billion loan, with Son betting on AI and leveraging debt)
(Additional context: Anthropic completed a $65 billion Series H funding round, with a valuation of $965 billion, surpassing OpenAI)

Table of Contents

Toggle

• Turning attack tools into defense products, the logic of same origin in offense and defense
• From 50 to 1,000 people
• The "Black Ship" analogy and Japan’s structural choices in the AI race

The same cutting-edge AI models can be used as cyber weapons to paralyze power grids and financial systems, or to scan for vulnerabilities and assist in defense. On June 16, Son Son, at a corporate event in Tokyo, unusually used the analogy of the "Black Ships" arriving, elevating the threat of AI cyberattacks to a national security level, while simultaneously announcing the launch of a new business.

Turning attack tools into defense products, the logic of same origin in offense and defense

In an official announcement, SoftBank stated that the joint venture SB OAI Japan GK, established by SoftBank Corp and OpenAI in November 2025, will launch a cybersecurity protection service called "Patching as a Service."

This service has two core components: first, using AI models to scan client systems for potential vulnerabilities; second, a technical team will accompany the process all the way until patches are applied. The entire process is a one-stop service, so companies don’t need to determine which vulnerabilities are most dangerous or how to fix them safely.

SoftBank is not directly selling an untested concept externally. Before launching publicly, Son Son said SoftBank had conducted large-scale internal drills on its own systems, using OpenAI’s cybersecurity technology to identify vulnerabilities. The testing results were "quite effective," and only then was the service announced externally.

From 50 to 1,000 people

Currently, about 50 technical staff are responsible for this service, and SoftBank plans to expand this team to 1,000, a twentyfold increase. The target clients are approximately 3,000 large enterprises operating critical infrastructure within Japan, including airports, power companies, and transportation operators.

Son Son’s calculation is straightforward: AI investments need practical scenarios for monetization, Japan’s government and industry are highly anxious about cybersecurity, and OpenAI’s technological capabilities have been validated within their own systems. Combining these elements results in "Patching as a Service."

On June 16, Son Son also met with Japan’s Finance Minister, Shōzō Katayama, alongside OpenAI executives. After the meeting, Katayama stated that both sides agreed on the need to overcome cybersecurity threats posed by AI and expressed hope to establish future cooperative frameworks. This scene itself signals that the service has received at least verbal endorsement from the Japanese government.

The "Black Ship" analogy and Japan’s structural choices in the AI race

Son Son’s use of the "Black Ships" metaphor is very clever. In 1853, U.S. Navy Commodore Matthew Perry led a fleet to Japan, forcing the country to open its ports. This event is remembered in Japan as a traumatic moment of "forced change of destiny" by external forces. Son Son frames the threat of AI cyberattacks within this analogy, with a clear message: Japan’s policymakers and business leaders should feel the same existential pressure.

A deeper structural issue is that Japan does not have its own top-tier frontier models in this AI race; the dominance of large models is held by a few American companies. Son Son straightforwardly said, "We trust OpenAI and will thoroughly utilize it to prevent cyberattacks." The implicit message is that, without a domestic leading model, Japan can only choose sides.