Robotics

Today's Stories

1. 1

   DeepSeek's new reasoning model reaches #2 ranking among open-weights options, signaling a shift in AI competition toward efficiency and accessibility.

   DeepSeek released a new reasoning model that ranks #2 among open-weights reasoning models. The model uses 27% of FLOPs compared with DeepSeek-V3.2 and supports 1M tokens (up from 128K in V3.2), while achieving stronger performance on reasoning benchmarks. The result suggests that smaller, more efficient models can match or exceed the capabilities of larger systems. This matters to businesses because it may reduce the cost and complexity of deploying advanced AI, shifting competitive advantage toward companies that optimize for efficiency rather than raw scale.

   The model's practical performance on real-world tasks beyond benchmarks, and whether other AI developers follow DeepSeek's focus on reducing computational requirements while maintaining reasoning quality.

2. 2

   Are You Ready to Automate?

   Are You Ready to Automate?

3. 3

   Analysis of China's 362 million workers shows clerical staff face the highest AI exposure risk despite being far outnumbered by craft workers, suggesting AI disruption won't follow workforce size.

   A researcher mapped AI exposure across China's occupational groups using ILO workforce data and occupation-based AI exposure models. Clerical support workers (33.6 million people) score 8.5/10 on AI exposure, while craft and related trades workers (93.6 million people) score only 2.5/10, and professionals (81.8 million people) score 6.5/10. The data reveals a mismatch between workforce size and AI vulnerability—China's largest occupational group faces relatively low AI disruption risk, while smaller groups of clerical workers face much higher risk. The country's weighted average AI exposure is 4.48/10, indicating that AI adoption will affect workers unevenly rather than across the board.

   A separate finding shows plant and machine operators score 3.0/10 on AI exposure but 7.5/10 on robotics risk, suggesting AI and robotics deployment follow different patterns. The analysis notes these AI exposure scores are modeled estimates based on occupation tasks and are not official government statistics.

4. 4

   The Global X Robotics & AI ETF (BOTZ) holds $3.54 billion（約5700億円） in assets, offering retail investors direct access to robotics and AI companies they might otherwise struggle to buy.

   The Global X Robotics & Artificial Intelligence ETF (NASDAQ:BOTZ) is the largest pure-play robotics fund in the U.S. market, with roughly $3.54 billion（約5700億円） in net assets spread across foreign-listed automation giants, U.S. AI chipmakers, and surgical robotics specialists. Most retail investors cannot easily purchase the international and specialized robotics companies held in this fund directly. The ETF solves that access problem by bundling these holdings into a single tradable security.

   The fund's size and composition mean it serves as a concentrated vehicle for exposure to robotics and AI hardware — a notably narrow focus compared to broader technology or AI-focused funds.

5. 5

   Tesla's stock has gained 2,630% over 10 years, but at a price-to-earnings ratio of 363, today's valuations leave little room for error.

   A $10,000 investment in Tesla 10 years ago would be worth $273,000 today, reflecting a 2,630% gain. The company's 2025 total revenue was $94.8 billion（約15兆円）, down year over year, even as the stock trades 19% below its December 2025 peak. Tesla has transformed from an unproven auto industry newcomer into a globally recognized tech company positioned at the premium end of the EV (electric vehicle) market. However, the current price-to-earnings ratio of 363 reflects market expectations so high that future returns depend heavily on the company delivering on its self-driving and robotics ambitions.

   The company's growth over the past decade has been tremendous, but investors seeking similar gains over the next 10 years should be cautious given how lofty current valuations are relative to near-term earnings.

6. 6

   More than 50 drones seized near U.S. World Cup venues, most violations from careless operators rather than security threats.

   Federal officials have seized more than 50 drones near FIFA World Cup event sites across the United States. Atlanta accounted for a large share, with the FBI seizing 26 drones as of June 18. Los Angeles reported at least 28 seizures, Dallas at least 33, Miami at least 28, and Seattle five. The FAA established No Drone Zones around stadiums (within three nautical miles and up to 3,000 feet during matches) and fan areas (within one nautical mile and up to 1,000 feet). The main problem is not sophisticated security attacks but careless or uninformed drone operators flying in restricted areas. Violations have been reported at all eight active U.S. stadium sites. Unauthorized drone flights can result in civil penalties of up to $75,000 per violation, criminal fines of up to $100,000, confiscation of the aircraft, federal charges, and possible arrest—serious consequences that suggest federal teams are actively enforcing restrictions around one of the world's largest sporting events.

   The distinction between careless operators and malicious ones may not matter to security teams managing the World Cup. Many of the confiscated aircraft appear to belong to people testing new equipment or attempting quick aerial shots rather than posing intentional threats, but enforcement remains active throughout the tournament.