JAPAN WUT PODCAST 201 "DARK FACTORY"
Should Japan be concerned about China’s patent for a undersea cable cutting device? Who stands to gain from Dark Factories? Can Japan’s AI startup hubs compete on the global stage?
SHOW NOTES
JAPAN
ANA places mega-order of 77 new jets
SUPPLY CHAIN WAR
Japan, Philippines OK strategic talks between forces amid China rise
Chinese Navy penetrates deep into the Tasman Sea
A towed submarine cable cutting device and its cable cutting method
China might be using cutting-edge AI to tilt the global power balance toward BRICS and away from the G7, zeroing in on Taiwan and the South China Sea to control supply chains. The twist? They could be slicing undersea cables to choke rival data flows—because in 2025, data’s the new oil. Let’s break it down with the evidence.
First, the cable incidents. Taiwan’s seen five undersea cable cuts this year, a sharp uptick from three each in 2023 and 2024, according to Taiwan’s Digital Ministry. On January 3, the Trans-Pacific Express cable—linking East Asia to the U.S.—snapped near Keelung Harbor at 7:51 a.m. A Chinese-crewed ship, the Shunxin-39, was spotted nearby, possibly dragging its anchor, though bad weather let it slip away to Busan. Then, on February 25, a cable to Penghu Islands went down at 3 a.m., with the Hong Tai 58—another Chinese-crewed vessel—detained after lingering suspiciously close. The other three cuts are murkier, likely hitting Matsu Islands cables in January or February, with repairs dragging into March, mirroring past Chinese vessel incidents in 2023. Taiwan’s 15 cables carry over 95% of its internet, so five hits in two months isn’t random—it’s a pattern screaming disruption.
Now, the AI angle. China’s got the tech to pull this off. The World Intellectual Property Organization reports they’ve filed over 38,000 AI patents from 2014 to 2023—six times the U.S.’s 6,276—spanning industrial applications like robotics and optimization. A 2020 patent from Lishui University for a cable-cutting device isn’t AI-driven, but pair it with over 3,000 AI patents since 2020 for underwater navigation—like Shanghai Jiao Tong’s 2022 seabed-mapping submersible—and you’ve got a system that could pinpoint and sever cables with precision. Huawei’s 2021 patent for network optimization could then reroute BRICS data, leaving G7 rivals scrambling. No patent explicitly says “AI cuts cables,” but the pieces fit—China’s AI could orchestrate this as a “grey zone” tactic, disrupting without declaring war.
Why Taiwan and the South China Sea? They’re supply chain goldmines. Taiwan’s TSMC pumps out 90% of the world’s advanced chips—vital for G7 tech—and the South China Sea handles 80% of global trade. Cut Taiwan’s cables, and you bottleneck G7 operations; control the sea’s sparse cables, and you redirect flows to BRICS-friendly nets like HMN Tech’s lines. Five cuts in 2025, all near Chinese ships, suggest intent—especially with China’s history of stalling rival cable projects there since 2022, per Foreign Policy. Data disruption hits G7 economies hard while BRICS builds alternatives, like the Belt and Road’s overland routes.
BRICS supply chain patents back this up. China dominates with roughly 60% of the bloc’s 25,000-30,000 AI-related filings since 2020—think Huawei and ByteDance’s logistics tweaks—designed to pivot flows post-disruption. Russia’s 2023 satellite monitoring and India’s 2021 port optimization add resilience. Compare that to the G7’s 30,000-40,000 patents—led by the U.S. and Japan with filings like IBM’s 2021 inventory rerouting or Toyota’s 2022 factory flow—which focus on surviving chaos, not starting it. China’s edge isn’t volume; it’s strategy—their tech could both cut and capitalize.
The theory holds water because it aligns with China’s goals: a BRICS-led order challenging the G7. Taiwan’s five incidents, tied to Chinese presence, match the capability of their AI ecosystem—over 54,000 total patents by 2023—and their supply chain ambitions. It’s not ironclad—no leaked AI directive proves it—but the spike from three to five annual cuts, paired with patents suggesting sabotage and rerouting, paints a picture. Data’s the lifeblood; disrupt it, and you shift the game.
SOCIETY 5.0
AI chip giant Nvidia reports blockbuster revenue
Dark factories are spreading. Factories are becoming unmanned and people are no longer there.
Toyota to build autonomous roads and underground passages in Woven City
Ericsson completes smart transformation of its Nanjing factory
Japan's digital deficit hits record ¥6.5 tril as U.S. tech reigns
Officials from each country meet in person for the first time to set rules for generative AI
Versclesia IDZ Chiyoda Iwamotocho Residential Condominium Units Sold Out
3 Japan teens arrested over fraudulent mobile contracts, aided by ChatGPT
Christie's first-ever AI sale angers some artists
The World As You Know It Is About to End
Goldman Sachs projects that AI will automate about 300 million jobs within the next decade, while the CEO of Anthropic, creator of leading AI platform Claude, predicts human-capable AI could replace 30% of human labor in just the next two years.
Last year alone, 150,000 tech workers lost their jobs across major companies, many due to AI-driven efficiencies. That’s still far from the projected 300 million jobs or 30% of global labor being replaced, but it’s where we’re headed. Consider these details…
JPMorgan’s AI system, COIN (Contract Intelligence), can process 12,000 commercial-loan agreements in seconds – a job that would take humans roughly 360,000 hours.
Morgan Stanley rolled out an AI assistant called Debrief for its 15,000 wealth advisors. It handles notetaking and meeting summaries, saving about 30 minutes per meeting. With around 1 million company Zoom calls a year, that adds up to 500,000 hours saved.
Companies using AI for their marketing strategies have seen response rates jump by 40% while cutting deployment costs by 25%.
At Goldman Sachs, AI can now draft 95% of an IPO prospectus in minutes –a job that used to take a six-person team weeks to complete.
Mark Zuckerberg, on a recent Joe Rogan podcast, said that this year – 2025 – Meta will have an AI assistant that works like a “mid-level engineer writing code.” To give you an idea of the savings, Meta has 15,000 mid- and low-level software engineers making $175,000 to $260,000 a year.
