Everything the newsroom published, in chronological order. Each item carries origin, sources and reading time.
CHIPS SK Hynix out of replacement SSDs in warranty claims; refunds original price amid shortage INDUSTRY Cloudflare Agent Tracing: A Debugging Tool, Not an Audit Trail INDUSTRY DoorDash Lifts Grocery Conversion 24% With Task-Aware Agent Memory
FUNDING Robotics funding surges to $18.8B in 2026; Chinese humanoids race toward IPOs with $400M rounds
MARKET Grok 4.6 reaches frontier parity with Fable 5 at 80-88% discount, pressuring Anthropic and OpenAI pre-IPO margins
FUNDING River AI closes $1.1B led by General Catalyst to democratize custom AI model training on open weights
MARKET Anthropic targets October IPO at $2T valuation, largest ever; $47B ARR growing to $100B-$120B
MARKET GPU spot rental prices double to ~$4/hr; RTX 5090 hits 135% markup as memory shortage bites
POLICY Modi sets India semiconductor timeline (7-8 fabs in 1-2 years) and 100 GW nuclear target by 2047
RESEARCH Grok 4.6 matches Fable 5 tier on agentic work at 85% discount; context-efficient benchmark win
CHIPS India targets 5–8 new semiconductor plants in 7–8 years; 5 nuclear reactors by 2033
MARKET India vows 5–7 new semiconductor plants in next 7–8 years; targets 100 GW nuclear capacity by 2047
FUNDING River AI closes $1.1B seed/Series A for open-weight model reinforcement learning at 15–20 min training
FUNDING Databricks closes $5B at $190B valuation on 80%+ YoY AI agent infrastructure momentum
RESEARCH Alibaba releases Qwen3.8-27B under Apache 2.0; dense model competes with Opus 4.6 Max on agentic tasks, runs on 32GB MacBook
INDUSTRY OpenAI's Enterprise Revenue Now Exceeds Consumer
FUNDING Oxylabs becomes Lithuania's unicorn after bootstrapping $350M ARR; Warburg backs data-for-agents play
RESEARCH Alibaba releases Qwen3.8-27B open weights; beats Muse Glimmer on coding, runs single GPU
FUNDING Cambridge Aerospace Raises $300M at $3.4B Valuation; UK Defense Startup Scales Skyhammer Low-Cost Interceptor
FUNDING Firmus raises $2B at $10.5B valuation; Australian AI data center builder doubles in 4 months
FUNDING CodeRabbit $143M Series C at $1.5B; launches Agentic Change Management for AI-generated code governance
MARKET Nvidia's SpaceX stake worth $21B after xAI acquisition; stock down 18% from IPO price
FUNDING CodeRabbit raises $143M at $1.5B, launches Agentic Change Management to govern AI code explosion
CHIPS Point2 Technology raises $136M Series B with Arm backing, targets AI interconnect bottleneck
FUNDING River AI lands $1.1B from General Catalyst, NVIDIA, AMD to democratize open-weight model training
MARKET HF Q3 2026: Chinese frontier models (754B–2.78T) now dominate; US shifting to optimization
POLICY Apple splits AI Intelligence by region; China model via Alibaba, Qwen, creates dev testing gaps
FUNDING Databricks closes $5B round at $190B valuation on 80% YoY revenue growth
RESEARCH Meta Muse Glimmer vs Alibaba Qwen3.8-27B: Local Agent Benchmarks Compete
MARKET OpenAI Enterprise Revenue Now Exceeds Consumer, $40B Run Rate
CHIPS SMIC Q2 breaches $3B revenue, triples profit on AI peripheral chips; CEO raises prices for Q3 orders as 93.7% utilization signals tightness
RESEARCH Alibaba opens Qwen3.8-27B weights: 27B multimodal model for single-GPU local inference, licensed Apache 2.0
BREAKING NVIDIA Opens Indonesia's First University AI Center; Eyes Southeast Asia Talent Pipeline
FUNDING Wayve hits $8.6B on $1.5B raise; embodied AI AV gets Microsoft, NVIDIA, Uber, 3 automakers; London robotaxi trial 2026
FUNDING Oxylabs lands $130M from Warburg Pincus at $3.6B; 11-year bootstrap reaches $350M ARR in web-data for agentic AI
INDUSTRY Production Use, Not Hype, Drives Open-Model Adoption
RESEARCH Z.ai GLM-5.3: 6x long-horizon coding gains, emergent cyber 84.5% (CyberGym) from post-training alone on frozen 743B base
FUNDING Databricks raises $5B at $190B (42% jump in 8 months); hits $7B ARR with 80% YoY growth, Lakebase $100M+
BREAKING Cloudflare adds MCP protocol detection, user tracking, portal-only access enforcement to Cloudflare One
RESEARCH Meta open-sources Muse Glimmer: 30B agentic model optimized for local GPU; 4-bit quantization, 3.1x speculative decoding