Daily Special Report
The day looks less like a single headline and more like a transition period: AI is moving from demos to workflows, chips and robots are still constrained by geopolitics, and markets are pricing in both earnings resilience and macro volatility. The next watchpoints are regulation, physical-world automation, and who can turn AI into durable revenue.
Today's coverage centers on a simple shift: AI is moving from models to workflows, hardware is moving from chips to the packaging stack, and policy pressure is rising across robots, platforms, and safety. Markets and consumer tech are reacting in parallel, but the biggest signal is that execution, regulation, and monetization now matter more than hype.
AI & Agentic Software
Agentic software is moving from novelty to workflow infrastructure. The set spans tabular foundation models, multimodal codecs, corpus relevance, RAG design, agent memory, and model-specific products from OpenAI, SKT, Krafton, Kimi, Cursor, and LearnVector.
The pattern is less about one breakthrough and more about specialization. Teams are now competing on task fit, retrieval quality, evaluation, latency, and localization instead of raw model scale alone.
That matters for monetization. The winners are likely to be the teams that bind models to a concrete surface such as coding, transcription, search, learning, or enterprise workflows, while also proving they can manage safety and reliability.
Article sources
- Latest AI Uses Tabular Foundation Models To Turn Columnar Data Into Vital Insights
- Truth is not a direction: a Tarski attack on LLM probes | Hacker News
- Hubble: Open-source notetaking app for you and your agents | Hacker News
- LearnVector – Andrew Ng's AI company building one-to-one learning experiences | Hacker News
- Entrepreneurs’ Response To AI: ‘Show Us The Money’
- Altman's latest interview: Don't expect AI to usher in a 4-hour workweek - 36Kr
- Mage-VL: An Efficient Codec-Native Streaming Multimodal Foundation Model
- A New Role for Relevance: Guiding Corpus Interaction in Agentic Search
Semiconductors & Advanced Hardware
The semiconductor story is increasingly about the stack, not just the transistor. Huawei's long-run chip math, TGV glass substrates, HBM4, 321-layer NAND, and domestic GPU adaptation all point to a market where packaging and memory matter as much as headline process nodes.
The real bottleneck is coordination across the full hardware chain. Substrates, inspection equipment, memory bandwidth, and integration now sit beside design as the places where advantage is won or lost.
That makes the race more capital-intensive and more geopolitical. Domestic substitution can create leverage, but only if yield, scale, and ecosystem depth keep improving.
Article sources
- Huawei Turns the Only Escape Route Into a Grand Road: Tau Law Time-Scaling Surpasses Geometric Shrinkage as 381 Chips in 6 Years Redefine Semiconductor Survival
- Domestic TGV Glass Substrate Industry Explodes: Chinese Manufacturers Race to Scale Through-Glass Via Technology as AI Chip Packaging Demands New Substrate Solutions
- Mineng Technology Raises Tens of Millions for SNN Brain-Like Chip: Becoming the Upstream Brain for Medical Devices With Spiking Neural Network Neuromorphic Computing
- Moore Threads announces completion of Kimi K3 adaptation: MTT S5000 domestic GPU powers a 28 trillion-parameter large model - IT Home
- Samsung weighs using Chinese-made DRAM to expand share in China’s budget smartphone market
- Gigavis signs a 64 billion won inspection equipment supply contract with a Chinese semiconductor substrate maker
- Foreign media test Intel platform DDR4/DDR5 memory performance: average gap of 11% to 14% across 15 games - IT Home
- SK hynix eyes over 200 trillion won in operating profit this year with HBM4 and 321-layer NAND
Robotics, Mobility & Physical AI
Physical AI is clearly past the concept stage. Internet giants, robotaxi pilots in London, personal eVTOL, embodied AI funding, and the World Robot Conference all show that robotics is becoming an investable product category.
The other half of the story is policy friction. U.S. robot bans, FCC blacklisting, and security framing around Chinese humanoids make robotics feel like both an industrial opportunity and a national-security issue.
The near-term winners are likely to be narrow deployments: delivery, driving tests, event demos, education, and defense-adjacent use cases. General-purpose autonomy is still far away, but the commercialization path is getting clearer.
Article sources
- Chinese Internet Giants Accelerate Into the Physical World as Meituan, JDcom, ByteDance and Baidu Build Robotics, Smart Delivery and Embodied AI Empires
- From Tsinghua Lab to Low-Altitude Market: Huaxi Aviation WorthyAero R1 Ultra-Light Personal eVTOL Weighs 115kg, Aims to Make Personal Flight Affordable
- 28 Days to Go | Intelligent Innovation, Integration, Manufacturing, and Fun: The Four Exhibition Halls of the 2026 World Robot Conference Are Here - 36Kr
- Embodied AI’s “20 Billion Club”: 8 Companies, Hundreds of Billions Raised, Betting on One Future - 36Kr
- An early investor in Unitree bets on a brain-computer interface company
- A Sector That Grew 40-Fold in 5 Years: Shooting Brake Is Entering Its “Smart Era” - 36Kr
- Baidu and Lyft Begin Robotaxi Tests in London, Intensifying Europe’s Autonomous Driving Race
- HiFi-UMI: Learning Deployable Manipulation Policies from High-Fidelity UMI Data Alone
Security, Regulation & AI Governance
The risk surface is widening. Worker petitions about slowing AI, nudify-app fights, healthcare warnings, and the AI robot Sally controversy all point to a society asking for more restraint.
Security is also moving deeper into the stack. Post-quantum attack research, MCP proxy controls, X's disaster-platform scrutiny, and Homeplus bond disputes show how model risk, legal risk, and disclosure risk are converging.
The takeaway is simple: governance is no longer a side topic. For companies shipping AI or operating data-rich platforms, compliance, transparency, and incident response are becoming part of product quality.
Article sources
- AI company employees petition US government to facilitate industry slowdown after security incident
- xAI Is Challenging a New Minnesota Law Banning 'Nudify' Apps
- Healthcare Is Undergoing a Cybersecurity Crisis, and It’s Not Slowing Down
- British Medical Association: Misleading “AI doctors” pose a major threat to public safety - IT之家
- [Securities Firms, Homeplus Short-Term Bonds] Largest issuance just before rehabilitation…Shinyoung Securities faces ‘design responsibility’ allegations
- [Securities Firms’ Homeplus Commercial Paper] How Far Does the Risk Disclosure Go… Mis-Selling Under Scrutiny
- Anthropic's Claude Mythos Preview discovers an attack technique that halves post-quantum cryptographic strength
- After the Kumamoto earthquake in Japan, it only showed recommendations X's role as a disaster platform comes under scrutiny
Markets, Finance & Corporate Earnings
Markets are being pulled in opposite directions. The KOSPI slump, Bank of Korea commentary, and tighter debt or ETF rules show policy trying to stabilize a nervous tape, while semiconductor earnings still provide some downside support.
The corporate line is mixed rather than broken. Rights offerings, order wins, profit beats, and growth resets sit next to losses, bond-risk disputes, and financing concerns. Capital is still available, but it is being priced more carefully.
That makes this a stock-picker market. Balance-sheet strength, refinancing risk, and policy sensitivity are likely to matter more than broad macro optimism until volatility settles.
Article sources
- How AVILOO Is Solving The Used EV Market Battery Health Trust Problem
- Copper And Aluminum Plug A Potential Profit Hole For Rio Tinto
- Global Top Beauty Group Returns to Growth - 36Kr
- About 9 new Hong Kong property developments, totaling around 2,581 units, to launch in August, up 40% from the previous month - 36Kr
- How Changsha Has Revitalized Service Consumption - 36Kr
- Global universities are flocking to China - 36Kr
- Alpen Nabs 10M Series B Funding Round
- Bank of Korea: “Semiconductor earnings support the stock market’s downside, but external uncertainty remains a variable”