East Asian Technology Intelligence
Japan & China tech news — translated, contextualized, and delivered to your inbox.
Free. Unsubscribe anytime.
3 Takeaways This Issue
- Nissan and Honda are abandoning their historical reliance on in-house development to partner with autonomous driving startups like Wayve and Helm.ai, a strategic shift aimed at closing the technology gap with Tesla and Chinese competitors.
- Japan’s share of global patent applications fell from 25% to 8% over the last two decades, reflecting how the country’s R&D apparatus has fallen behind the United States and China in the critical fields of artificial intelligence and software.
- Snowflake’s record fourth-quarter revenue of $1.39 billion highlights how enterprise data infrastructure providers are successfully monetizing the corporate rush to deploy autonomous AI agents.
This Issue’s Analysis
The Signal
Google’s Gemma Strategy in Japan: How Open-Weight Models Are Winning Conservative Enterprises
Google announced that its open-weight Gemma model family has exceeded one billion cumulative downloads and unveiled a curated repository called “Awesome Gemma” on GitHub to organiz
Robotics & Automation
Nissan and Honda’s Autonomous Pivot: Partnering with Startups to Rival U.S. Robotaxis
Japanese automakers Nissan and Honda are abandoning their traditional preference for in-house technology development to partner with specialized autonomous driving startups. This s
Robotics & Automation
Unitree’s 18,000-Unit Milestone: How China’s Scaling Hits Western Humanoid Robotics Vendors
Chinese robotics pioneer Unitree has made a spectacular debut on Shanghai’s STAR Market, with its valuation soaring to RMB 445 billion ($62.3 billion USD) after shares surged 629%
AI & Machine Learning
Alibaba’s 75% Capex Surge: Driving High-Performance Edge AI Despite Profit Drops
Alibaba has dramatically increased its capital expenditure by 75% to build out AI infrastructure despite a 76% drop in quarterly net income, while simultaneously releasing highly o
🧩 Pattern This Issue
- Japan: automakers Nissan and Honda abandon in-house autonomous development for startups
- China: Alibaba increases capital expenditure by 75 percent to fund open-source AI
- Korea/Taiwan: Samsung reports record H1 revenue driven by surging Chinese semiconductor demand
East Asian industrial giants are abandoning closed-loop R&D pipelines to aggressively subsidize open platforms and specialized startups, a shift that will erode the software-moat advantages of US tech incumbents as hardware and infrastructure become heavily commoditized.
Also This Issue
🗾 Japan Radar
What Japanese media is reporting that Western outlets miss
🗾 Startups & Funding
Anthropic Could Make IPO Prospectus Public This Month in Record-Breaking Listing, US Media Reports
AI startup Anthropic is expected to publicly release its IPO application documents as early as August, following a confidential draft submission to the SEC in June. The company’s valuation reached $965 billion in May, and its market capitalization upon listing could exceed $1 trillion—and potentially top $2 trillion—positioning it to surpass SpaceX as the largest IPO in history. Competitor OpenAI is also preparing for a public listing, with plans to go public by 2027.
Why it matters: A public listing of this scale will establish a concrete, market-tested valuation benchmark for generative AI pioneers, shifting the sector from private venture-backed hype to public market scrutiny. This liquidity event will trigger a massive capital reallocation across the global tech ecosystem, forcing competitors to accelerate their own monetization timelines.
For Western readers: Western investors should assume that the window for private mega-rounds in foundational AI is rapidly closing, as market leaders transition to public markets to fund their capital-intensive compute requirements.
🗾 Enterprise & Cloud
Snowflake Reports Record Earnings as CEO Argues Differentiation Has Become Easier Amid AI Agent Push
Snowflake posted record Q4 revenue of $1.39 billion (+33% YoY) and raised full-year product revenue guidance to +31%, driven by rapid adoption of its new Agentic Control Plane products, CoCo (a coding agent for developers) and CoWork (an enterprise knowledge worker agent). CEO Sridhar Ramaswamy positioned the company’s shift toward direct user-facing AI interfaces as a strategic moat, arguing that standardized data infrastructure makes differentiation simpler when competitors chase identical architectures.
Why it matters: Snowflake’s explicit pivot captures the telemetry layer that determines which providers retain workflow automation spend as traditional SaaS margins compress under generative AI compression. Domestic Chinese and Japanese cloud vendors tracking this will prioritize agent-to-database integration over model development, while Western SaaS firms lose direct user touchpoints if they refuse to embed grounded interfaces directly into their stacks.
For Western readers: If you evaluate enterprise data stack investments in APAC subsidiaries, assume Snowflake’s agentic layer will dictate integration requirements for local LLM deployments within 6–9 months; prioritize vendors that explicitly support structured telemetry export over those pushing closed-agent ecosystems.
Cross-Regional Analysis
Japan’s Share of Global Patent Filings Shrinks to 8% Amid US-China AI Dominance
A new Nikkei analysis reveals Japan’s share of global patents has plummeted from 25% to 8% over the past two decades. This decline is driven by the rapid acceleration of artificial intelligence breakthroughs and foundational technology filings by players in the US and China. Japan now ranks fourth in core AI mechanism and foundational technology patents, trailing behind China, the United States, and South Korea.
Why it matters: Japan’s declining patent share forces its domestic tech giants to rely on foreign AI foundations, reducing their leverage in licensing negotiations and increasing their long-term IP royalty burdens. To counter this, Tokyo is funding domestic AI champions like SoftBank to build sovereign models, but these state-guided efforts are defensive plays rather than bids for global market leadership.
For Western readers: Western enterprises negotiating joint ventures with Japanese conglomerates should expect diminished domestic IP leverage from their partners and should independently secure their core software and AI licensing pathways rather than relying on Japanese partners to provide them.
🇨🇳 China Watch
China’s technology moves, framed for Western readers
Policy & Regulation
Chinese Antitrust Regulator Imposes $1 Billion Fine on Meituan
China’s antitrust regulator, the State Administration for Market Regulation, is imposing a $1 billion fine on food-delivery giant Meituan for anticompetitive practices. The penalty targets the platform’s ‘choose one of two’ exclusivity agreements that forced merchants to list solely on its platform. This action aligns with Beijing’s broader, multi-year regulatory crackdown aimed at curbing the monopolistic influence of its domestic internet majors.
Why it matters: Beijing is establishing a strict compliance baseline that prioritizes state-guided economic order over unchecked platform growth. Meituan will have to dismantle its exclusivity agreements, immediately opening the door for rivals like ByteDance’s Douyin to aggressively capture market share in local life services and food delivery.
For Western readers: If you are a Western investor or brand operating in China’s digital ecosystem, stop assuming platform exclusivity is a viable defensive moat; plan for a highly fragmented, multi-platform distribution strategy as merchant lock-ins are legally dismantled.
🔺 The Triangle
Where US, Japan, and China technology interests intersect
Semiconductors & Hardware
Samsung Posts Record H1 2026 Results Driven by Semiconductor Boom and Surging Chinese Demand
📊 Featured Chart
Source: Samsung H1 2026 filing
Samsung Electronics reported record consolidated revenue of $205.6 billion for the first half of 2026, with its semiconductor division contributing 97% of total operating profits at $96.2 billion. China consolidated its position as Samsung’s largest market, accounting for 34% of overall revenue, which represents $59.7 billion in sales.
Why it matters: China’s rising share of Samsung’s revenue—up from 25% in 2023 to 34% in early 2026—shows that Beijing’s legacy and consumer hardware industries are absorbing massive volumes of memory and system chips. While Washington focuses on cutting off advanced AI silicon, Chinese device makers are successfully securing the fundamental hardware layers necessary to dominate global consumer and industrial electronics.
For Western readers: Do not assume that export controls are successfully decoupling the East Asian hardware supply chain; instead, expect Beijing to leverage its position as Samsung’s largest customer to secure favorable pricing and allocations on mature-node memory and components.
Semiconductors & Hardware
Quantum technologies rush out of US labs to match China’s growth
Washington is directing $1.4bn via CHIPS Act grants to IBM and GlobalFoundries while building quantum manufacturing capacity ahead of China’s rapid expansion in cryogenic equipment supply chains. The push prioritizes near-term sensor deployment over fault-tolerant computing as the US attempts to secure critical hardware infrastructure before Beijing dominates the physical layer.
Why it matters: Washington’s grant structure and Pentagon timeline show that quantum will follow the same industrial policy playbook as classical semiconductors, where government capital de-risks manufacturing while private firms race for scale in tooling. China’s leap from zero to ten refrigerator manufacturers since 2021 proves state-guided supply chain capture outpaces academic milestones, meaning Western buyers should expect Beijing to control cryogenic infrastructure before fault-tolerant chips arrive.
For Western readers: If you source quantum hardware or invest in deep tech manufacturing, assume Chinese equipment makers will secure initial deployment contracts by 2027 and prioritize partnerships with Japanese vacuum and optical instrument firms that can fill the US capacity gap.
Policy & Regulation
US Patent Rulings Threaten AI-Generated Drug Pipelines of Chinese Biotech Pioneers
Chinese biotechnology pioneer Insilico Medicine is facing intellectual property hurdles in the US after filing patents for its AI-discovered pulmonary fibrosis drug under human names to comply with US patent law. Despite the company’s marketing focus on its generative AI platform, US courts maintain that only humans can be legally recognized as inventors, creating potential legal vulnerabilities for AI-driven drug portfolios.
Why it matters: Chinese AI drug discovery firms are highly exposed to this regulatory gap because they aggressively market their ‘end-to-end AI’ capabilities to secure venture funding, yet must legally downplay the AI’s role in patent filings to ensure enforcement in Western markets. This discrepancy opens the door for competitors to invalidate patents by proving that human software operators did not contribute enough to meet the legal definition of joint inventorship.
For Western readers: Western pharmaceutical firms licensing drug candidates from Chinese or global AI-native biotech partners must audit the target’s patent filings to ensure human contribution is thoroughly documented, otherwise they risk acquiring assets with highly fragile IP protections.
Semiconductors & Hardware
Chiplets Offer a Scalable Compute Strategy for Software-defined Vehicles
As automotive compute demands outpace traditional monolithic system-on-chip scaling, chiplet architectures are emerging as the viable path forward for software-defined vehicles. This shift allows manufacturers to decompose large processors into multiple specialized dies within a single package, balancing high neural processing demands with strict automotive reliability and cost constraints.
Why it matters: The shift from monolithic SoCs to automotive chiplets moves the competitive bottleneck from front-end logic design to advanced packaging capacity. Legacy tier-1 automotive suppliers who fail to secure packaging partnerships with specialized foundries risk losing control over the central compute architecture of next-generation vehicles.
For Western readers: Western automotive OEMs relying on monolithic silicon roadmaps must pivot to chiplet-ready software architectures within the next design cycle or face severe cost and thermal efficiency disadvantages compared to East Asian rivals adopting modular packaging.