How Tech Investors Are Pivoting From Crypto Infrastructure to Practical AI Productivity

Dashan left Huawei HiSilicon in 2017 to pursue Bitcoin full-time, later founding Waterdrip Capital, which has invested in over 200 projects and managed multiple funds; more recently, he served as chairman of an AI company, leveraging his crypto-mining infrastructure experience for AI data centers.
IMTS 2026 Conference features 69 presentations focused on process innovation, plant operations, quality/inspection, and automation, reflecting the event's emphasis on practical productivity improvements in manufacturing AI adoption.
Capsule Security built an AI circuit breaker that runs real-time detection on mid-range hardware: a 4‑billion-parameter Nemotron model on an NVIDIA L40S GPU, with a 71 ms median verdict and a 500 ms decision budget; the 30‑billion‑parameter model detects 99% of rogue trajectories with under 1% false positives.
Runner AI's Fly central intelligence coordinates a team of 10+ AI agents across functions (ads, SEO, email, CRO, customer support, analytics, creative) to run an online business end-to-end, with founders approving proposals and guiding strategy.
Forbes argues that AI spending is rising not from one-off procurements but from widespread usage across departments, encapsulated in the phrase 'Every AI prompt is a procurement decision' and the 'token-as-a-service' (TaaS) concept, signaling a shift in procurement thinking.
Da Shan's career arc — from Bitcoin mining engineer to venture capitalist to AI infrastructure chairman — mirrors a broader shift in tech money from cryptocurrency to artificial intelligence. ODaily reports that the investor left Huawei HiSilicon in 2017 to pursue crypto full-time, later founding Waterdrip Capital and backing over 200 projects. Now he's leveraging decades of mining infrastructure expertise to build data centers powering AI systems, exemplifying how capital is chasing the next frontier.
Across industries, AI is no longer a lab experiment — it's moving into factories, retail shops, and business operations. France24 and Spectrum News report that companies are deploying autonomous AI agents to handle everything from customer support to ad buying, while makers of industrial robots and safety tools race to keep pace with rapid adoption and emerging risks.
Da Shan's journey began in 2011 when he first encountered Bitcoin. By 2013, while finishing his PhD, he had already built custom mining rigs. ODaily documents how in 2017 he made a clean break from Huawei HiSilicon to pursue the industry full-time. He founded Waterdrip Capital, which has since invested in over 200 projects and managed multiple venture funds.
His latest pivot showcases the infrastructure overlap between crypto and AI. Both require massive computing power, cooling systems, and electrical grids. By serving as chairman of an AI company, Da Shan brought his mining expertise directly to AI data centers — the new engine of enterprise productivity.
Industrial automation is shedding its reputation for complexity and cost. European Business Review highlights how 'Physical AI' — artificial intelligence that perceives and acts in the real world — is transforming economics across factories and workshops. New startups are building AI-powered robotics designed for machine shops where operators lack deep technical training.
The goal is simple: make automation so intuitive that any shop operator can deploy it without retraining their workforce. IMTS 2026 — a major manufacturing conference — is featuring 69 presentations on process innovation and shop-floor automation, signaling strong industry demand for practical, affordable AI tools.
As AI agents gain autonomy, concerns about safety are mounting. France24 and Spectrum News report that companies like Capsule Security are building real-time 'circuit breakers' to detect dangerous AI behavior before it happens. The system runs a 4-billion-parameter detection model on a single NVIDIA L40S GPU, delivering verdicts in just 71 milliseconds.
For deeper threats, a 30-billion-parameter model detects 99% of rogue AI trajectories with under 1% false positives — all within a 500 millisecond decision window. These guardrails are essential as AI agents begin operating autonomously across business functions, from customer service to financial decisions.
A new model is emerging: autonomous AI teams managing daily operations while founders set strategy and approve major decisions. France24 profiles Runner AI, which deploys 10+ specialized agents across ads, SEO, email, customer support, analytics, and creative work. The system — called 'Fly' — orchestrates these agents end-to-end to run an online business.
This split mirrors how large companies already work: departments execute while executives govern. Forbes reports that AI spending is shifting from one-off purchases to continuous use across teams, framed as 'token-as-a-service' billing. Every AI prompt becomes a procurement decision, forcing companies to rethink how they budget for and govern AI adoption.
Publishers
22
Articles
9
Reach
31