Why 3K Industries Plus AI Is the Best Startup Bet Now

Shibata Naoki and Tadokoro Masayuki argue that unglamorous, labor-short industries beat already-digitized sectors for AI startup opportunity.

• Target “3K” industries — tough, dirty, dangerous — because they are un-digitized and will accept 60% AI accuracy over zero human capacity.
• Avoid digital marketing and ad-tech; Google, Facebook, and Amazon spent 20 years optimizing those with the best engineers before AI existed.
• Shibata personally invested in Freeworks, an AI agent that matches shipping companies with truck drivers via chat and voice, replacing phone/email dispatchers.
• Freeworks example: AI can charge premium rates for low-demand routes (Gunma to Akita) and negotiate discounts on high-demand routes (Tokyo to Osaka), raising logistics company margins structurally.
• Frame AI as a revenue multiplier, not a cost cutter — cost framing makes employees fear job loss and slows adoption; revenue framing makes everyone, including staff, want it to succeed.
• Japan’s 2024 trucking-hours regulation tightened driver supply while cargo volume rose, making logistics a structurally understaffed and high-value AI target.
• Long-term competitive moat in the LLM era comes from proprietary fine-tuning data accumulated through sales and operations, not from the model itself.

2025-11-21 · Watch on YouTube