👀In Case You Missed It
Anthropic dropped Claude Opus 4.5, breaking 80% on SWE-Bench Verified, cutting API prices by ~66%, and pushing deep into agentic workflows with Chrome, Excel and Claude Code integration — a full-stack productivity play.
OpenAI launched Shopping Research, a GPT-5 mini–powered assistant that handles preference-finding, cross-site scanning and recommendation generation in a single workflow, with near-unlimited usage baked in.
Google’s TPU moment is here, as Meta explores renting Google’s chips from 2027 — the first serious sign that a hyperscaler might diversify away from Nvidia at frontier scale.
The global chip supply chain is flashing stress signals, from DRAM price spikes and TSMC wafer losses to Europe funding new fabs and investors debating whether the GPU cycle is peaking — all driven by relentless AI demand.
Read more here
🧠 Cloud & Infrastructure Watch
Ilya Sutskever says the “age of scaling” is ending as frontier progress shifts from compute to pure research, signalling a strategic reset for long-horizon AGI labs.
Read MoreAnthropic’s new study claims AI could double U.S. labour productivity growth by accelerating knowledge work and automating multi-step workflows across every major services sector.
Read MoreNVIDIA fired back at rising TPU adoption, asserting Blackwell remains “a generation ahead” in performance, flexibility, and fungibility for multi-model inference.
HP will cut roughly 6,000 jobs by 2028 as part of a large AI-driven operational overhaul across devices, cloud tools, and enterprise support functions.
Read MoreMicrosoft’s Copilot is exiting WhatsApp after rule changes banning general-purpose AI chatbots from the platform, underscoring new governance controls over messaging ecosystems.
Read MoreTencent open-sourced HunyuanOCR, a state-of-the-art document understanding model optimised for enterprise-scale text extraction.
Read MorePerplexity launched a personal shopping assistant with memory-based recommendations and PayPal-powered instant checkout.
Read MoreOpenAI is rolling out unified text–voice conversations, letting users speak and type in the same interface with real-time streamed responses.
Read MoreGoogle Cloud and NATO finalised a major sovereign cloud deal enabling classified AI-assisted training, modelling, and threat analysis under strict European isolation controls.
Read MoreAmazon employees warn the company is adopting an “all-costs-justified” approach to AI deployment, raising internal safety and governance concerns.
Read MoreFleet Space used AI-enhanced satellite imaging to identify a massive lithium deposit in Quebec, confirming AI-geophysics as a new industrial exploration frontier.
Read MoreGoogle Cloud warns that demand for Nvidia GPUs and its own TPUs is rising simultaneously, suggesting dual-track infrastructure growth rather than a single-vendor trajectory.
Read MoreChina’s major tech giants are increasingly shifting AI model training offshore to bypass domestic chip restrictions and access Nvidia hardware more freely.
Read MoreAlibaba Cloud revenue jumped 34% as the company struggles to meet exploding Chinese AI compute demand.
Read MoreChina reportedly blocked ByteDance from using Nvidia chips domestically, accelerating Beijing’s push for local silicon alternatives and sovereign AI pipelines.
Read MoreOpenAI’s massive RAISE Act lobbying fight in the U.S. shows rising tension between state lawmakers, industry, and federal policy on AI governance.
Read MoreMultiple London councils activated emergency plans after cyberattacks disrupted digital services, reflecting increasing municipal vulnerability to AI-enabled intrusions.
Read More
💻 Hardware & Semiconductor Moves
Black Forest Labs launched Flux.2, a next-gen image model suite with multi-reference consistency, 4MP outputs, improved typography, and lower cost than rivals.
Read MoreNVIDIA reiterated Blackwell’s leadership as Wall Street questioned whether Google’s accelerating TPU roadmap is eroding GPU dominance.
Read MoreDRAM shortages trigger a surge in RAM pricing, with enterprise memory expected to remain supply-constrained into 2026 due to AI server demand.
Read MoreDell reported AI server revenue surging more than 100% YoY and expects another 50% jump next year as hyperscalers expand GPU-dense clusters.
Read MoreSamsung’s Exynos 2500 is now running fully on-device generative AI via Nota AI optimisation, signalling more offline AI work across mobile silicon.
Read MoreChina’s tech giants are turning to overseas training clusters—and alternative chip vendors—as domestic restrictions tighten.
Read MoreHP’s layoffs come as the company aggressively pivots toward AI-optimised PCs and edge devices.
Read MoreIQM will invest over €40m to expand quantum chip production, reinforcing Europe’s bid to stay competitive in high-performance compute.
Read MoreMeta, Google, and Amazon warn of a deepening memory-chip crunch driven by hyperscale AI server builds.
Read MoreNew evidence suggests China is slowly breaking free from Nvidia dependence via alternative accelerators, domestic fabs, and offshore compute deals.
Read More
