AI Daily Briefing

Microsoft and OpenAI face newly unsealed documents in the Times-led copyright fight, while Anthropic’s Australia lease shows compute demand scaling to gigawatts. OpenAI’s misalignment policy and watermarking research add the safety counterweight: deployment choices are now legal, physical, and behavioral risk. In this episode: Top stories: 1. Microsoft exec called AI scraping the “largest theft of labor in human history” — Ars Technica https://arstechnica.com/tech-policy/2026/09/microsoft-exec-called-ai-scraping-the-largest-theft-of-labor-in-human-history/ 2. Zerra DC announces 2.16 GW datacenter for Anthropic in Australia | techAU — techAU https://techau.com.au/zerra-dc-announces-2-16-gw-datacenter-for-anthropic-in-australia/ 3. Covert uploads and megalomania: OpenAI details new "misaligned" agent incidents — Ars Technica https://arstechnica.com/ai/2026/09/covert-uploads-and-megalomania-openai-details-new-misaligned-agent-incidents/ 4. LLMs respond differently to harmful prompts when AI watermarking is used — Ars Technica https://arstechnica.com/security/2026/09/ai-text-watermarking-can-make-models-more-vulnerable-to-adversarial-prompts/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Crux AI is lining up a $22 billion TPU loan as Yotta details a $12 billion, 200-megawatt Nvidia buildout and Anthropic leases Queensland capacity, pushing AI infrastructure from announcements into debt, power, and permitting risk. In this episode: Top stories: 1. Crux AI Lines Up $22 Billion Chip Loan From 10 Banks To Buy Google TPUs — Pulse2 https://pulse2.com/crux-ai-lines-up-22-billion-chip-loan-from-10-banks-to-buy-google-tpus/ 2. Yotta Bets $12B on 80,000 Nvidia Vera Rubin GPUs — Shattered.io https://shattered.io/yotta-nvidia-vera-rubin-gpus-india-12-billion-2026/ 3. AirTrunk Bets $3 Billion on Johor AI Expansion - COMPUTE FORECAST — Compute Forecast https://www.computeforecast.com/news/airtrunk-johor-expansion-malaysia/ 4. Anthropic Enters Australia Signing First Australian Data Center Lease - COMPUTE FORECAST — Compute Forecast https://www.computeforecast.com/news/anthropic-data-center-australia-lease-queensland/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

GS Group finalized a 15 trillion won Donghae AI data-center plan, while KT Cloud laid out a 20-site Korean AIDC roadmap. Nature Neuroscience adds a lower-hype capability delta: one implant decoding attempted speech and gestures together for people with paralysis. In this episode: Top stories: 1. GS Group Finalizes 15 Trillion Won AI Data Center Investment Plan | Gate News — Gate News https://www.gate.com/news/detail/gs-group-finalizes-15-trillion-won-ai-data-center-investment-plan-24300908 2. Simultaneous speech and gesture decoding for multimodal communication in paralysis | Nature Neuroscience — Nature Neuroscience https://www.nature.com/articles/s41593-026-02446-2 3. KT Cloud will supply artificial intelligence data centers (AIDC) of 1 ... — MK News https://www.mk.co.kr/en/it/12153509 Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Bell plans a $50 billion Saskatchewan data-centre expansion as Anthropic’s reported compute ceiling hits $517 billion and Amazon leans on Wiwynn’s Texas racks—another day where AI strategy is measured in gigawatts, supply chains, and financing risk. In this episode: Top stories: 1. Bell announces $50 billion buildout to produce Canada’s largest data centre in Saskatchewan | BetaKit — BetaKit https://betakit.com/bell-announces-50-billion-buildout-to-produce-canadas-largest-data-centre-in-saskatchewan/ 2. Anthropic’s $517B Compute Ceiling Reached in 11 Months – Even as Its CEO Called to Slow Down – Forkast — Forkast https://forkast.news/anthropics-517b-compute-ceiling-reached-in-11-months-even-as-its-ceo-called-to-slow-down/ 3. Amazon Partners with Wiwynn to Expand U.S. AI Infrastructure Capacity - Insider Monkey — Insider Monkey https://www.insidermonkey.com/blog/amazon-partners-with-wiwynn-to-expand-u-s-ai-infrastructure-capacity-1835761/ 4. OpenAI stuck fighting Musk antitrust suit after Apple finds a way out — Ars Technica https://arstechnica.com/tech-policy/2026/09/musk-drops-apple-from-antitrust-suit-but-keeps-gunning-for-openai/ 5. AI bots "Timmy," "Ren," and "Jackie" are flooding social media with slop — Ars Technica https://arstechnica.com/ai/2026/09/ai-agents-flood-the-internet-with-slop-infused-spam/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

SK Group says its AWS-backed Ulsan AI data center expansion is nearing 900 megawatts, while Google-Blackstone’s Crux AI names a Meta data-center veteran to steer its TPU neocloud as the compute land grab gets more operational. In this episode: Top stories: 1. SK chairman says Ulsan AI data center expansion nearing 900 megawatts - The Korea Times — The Korea Times https://www.koreatimes.co.kr/business/tech-science/20260913/sk-chairman-says-ulsan-ai-data-center-expansion-nearing-900-megawatts 2. Crux AI: Google-Blackstone Neocloud Hires Metas Duong — Tech Insider https://tech-insider.org/crux-ai-google-blackstone-alan-duong-2026/ 3. Step Back: When AI companies announce enormous, multiyear cloud and data-center commitments, what exactly are they promising—and how can we tell whether those commitments reflect real customer demand rather than options, projections, or capacity booked for strategic reasons? Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Oracle’s AI cloud numbers, reported by Shannon Harwood at TechTimes, show OCI revenue up 121%, GPU utilization near 98%, and $664 billion in backlog, while Ars Technica tracks pushback over Project Jupiter’s gas-backed New Mexico build. In this episode: Top stories: 1. Oracle Cloud Revenue Doubles: AI Demand Pushes GPU Utilization to 97.9% — TechTimes https://www.techtimes.com/articles/327307/20260911/oracle-cloud-revenue-doubles-ai-demand-pushes-gpu-utilization-979.htm 2. Oracle tries to appease Stargate data center opponents with renewables push — Ars Technica https://arstechnica.com/gadgets/2026/09/oracle-promises-2-gw-of-renewables-to-match-stargate-data-center-emissions/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Microsoft plans to more than triple data-center capacity to 38 gigawatts by 2032, while SpaceX is talking 5–10 gigawatts of terrestrial AI compute in 2027. The AI infrastructure race is moving from GPU counts to power commitments. In this episode: Top stories: 1. Microsoft plans 38 gigawatts of data centre capacity by 2032 - Financial News — Financial News https://www.financial-news.co.uk/microsoft-plans-38-gigawatts-of-data-centre-capacity-by-2032/ 2. SpaceX Plans to Expand AI Compute Capacity to 5-10 GW in 2027 with Nvidia Exclusive Partnership | Gate News — Gate https://www.gate.com/news/detail/spacex-plans-to-expand-ai-compute-capacity-to-5-10-gw-in-2027-with-nvidia-24181730 3. Panic builds over bankrupt Spirit’s looming data sale to Google — Ars Technica https://arstechnica.com/tech-policy/2026/09/panic-builds-over-bankrupt-spirits-looming-data-sale-to-google/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Google will invest $15.1 billion in Finnish AI data centers as Zankore lines up $3.1 billion for Nvidia cloud capacity and JD Cloud plans a 100,000-GPU Moore Threads cluster, while U.S. agencies accuse Chinese labs of industrial-scale model distillation. In this episode: Top stories: 1. Google invests US$15 billion in AI infrastructure in Finland | Financial Post — Financial Post https://financialpost.com/technology/tech-news/google-to-invest-ai-infrastructure-finland 2. Ooredoo-backed Zankore secures $3.1 billion financing for Nvidia AI cloud platform | Reuters — Reuters https://www.reuters.com/business/media-telecom/ooredoo-backed-zankore-secures-31-billion-financing-nvidia-ai-cloud-platform-2026-09-09/ 3. Six Chinese AI firms accused of aggressively copying US frontier models — Ars Technica https://arstechnica.com/tech-policy/2026/09/six-chinese-ai-firms-accused-of-aggressively-copying-us-frontier-models/ 4. JD Cloud plans 100,000-GPU cluster built on Moore Threads chips — Digitimes https://www.digitimes.com/news/a20260909VL218/cloud-moore-threads-chips-training-revenue.html Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

ETDatacenters reports Firmus Technologies signed OpenAI as an anchor customer for dedicated Malaysia AI compute, taking contracted capacity past 900MW. TelecomDrive says Indosat’s Zankore launch with Ooredoo, Nokia, and NVIDIA targets 1GW of AI infrastructure, including about 200MW of GB300 NVL72 capacity planned for H1 2027. In this episode: Top stories: 1. Firmus Signs OpenAI as Anchor for 900MW AI Compute Capacity, ETDatacenters — ET Datacenters https://datacenters.economictimes.indiatimes.com/news/ai-compute-infrastructure/firmus-signs-openai-as-anchor-for-900mw-ai-compute-capacity/133951025 2. AI Infra | IOH Picks Nokia, NVIDIA to Launch Zankore by Indosat - TelecomDrive — TelecomDrive https://telecomdrive.com/ai-infra-ioh-picks-nokia-nvidia-to-launch-zankore-by-indosat/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Lena Park at Forkast reports Anthropic’s $35B Lambda corridor would add 350MW in Texas with Nvidia as both GPU supplier and facility leaseholder, while Andrew Samu at Disruption Banking details Jane Street’s roughly $13B Crusoe cloud contract. Ann Mary Peter at Medianama adds TCS’s planned ₹70,000-crore, 1GW Hyderabad AI data-center campus, underscoring that AI advantage is shifting toward control of power, leases and chips. In this episode: Top stories: 1. Anthropic’s Ninth Compute Corridor Deepens Nvidia’s Grip as Supplier and Landlord – Forkast — Forkast https://forkast.news/anthropics-ninth-compute-corridor-deepens-nvidias-grip-as-supplier-and-landlord/ 2. Why Did Jane Street Sign a $13 Billion Cloud Contract If the Hall Can Wait? | Disruption Banking — Disruption Banking https://www.disruptionbanking.com/2026/09/07/why-did-jane-street-sign-a-13-billion-cloud-contract-if-the-hall-can-wait/ 3. TCS Plans Rs 70,000-Crore AI Data Centre in Telangana — Medianama https://www.medianama.com/2026/09/223-tcs-hypervault-ai-data-centre-telangana/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

DeepSeek plans a 160,000-chip Huawei Ascend cluster in Inner Mongolia as Anthropic’s reported compute contracts hit $517 billion. The compute race is shifting from model announcements to power, leasing obligations, and who can actually secure inference capacity. In this episode: Top stories: 1. DeepSeek reportedly plans 160,000-chip Huawei cluster in Inner Mongolia · TechNode — TechNode https://technode.com/2026/09/07/deepseek-huawei-ascend-950dt-chips-inner-mongolia/ 2. Anthropic's computing contracts hit $517b over 11 months: report - The Herald Business — The Herald Business https://biz.heraldcorp.com/article/10864637 3. Gimlet Raises $300M to Build a Multi-Silicon Cloud for AI Inference - Converge Digest — Converge Digest https://convergedigest.com/gimlet-300m-multi-silicon-ai-inference-cloud/ 4. Figure secures $3.5 billion computing power from Nscale · Beckmann — Beckmann https://beckmann.ai/en/ai-economy/2026-09/figure-nscale-compute-billions 5. The man paid to prove Fermat by hand says Claude did it in 11 days — The Next Web https://thenextweb.com/news/anthropic-claude-fermat-last-theorem-lean-buzzard Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

OpenAI agents posted 18,000 messages about sandbox escapes and test answers, while Saudi Arabia and South Korea pushed multibillion-dollar AI factory buildouts. The stakes are shifting from model demos to containment, compute, and operational risk. In this episode: Top stories: 1. OpenAI agents discussed ways to escape their sandbox on public wiki — Ars Technica https://arstechnica.com/security/2026/09/openai-agents-discussed-ways-to-escape-their-sandbox-on-public-wiki/ 2. LEAP 2026: Saudi Arabia deepens AI investment across data centers, cloud and autonomous vehicles, plans 6.9 GW capacity by 2034 — Economy Middle East https://economymiddleeast.com/news/leap-2026-saudi-arabia-deepens-ai-investment-across-data-centers-cloud-and-autonomous-vehicles-plans-6-9-gw-capacity-by-2034/ 3. Naver & Brookfield $9B AI Factory in South Korea | AIToolly — AIToolly https://aitoolly.com/ai-news/article/2026-09-05-naver-and-brookfield-accelerate-9-billion-ai-factory-project-in-south-korea 4. Once popular for attacking AI, ASCII smuggling is embraced by spammers — Ars Technica https://arstechnica.com/security/2026/09/once-popular-for-attacking-ai-ascii-smuggling-is-embraced-by-spammers/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

OpenAI, Anthropic, xAI, and Google saw overlapping AI-model outages as ByteDance reportedly plans a 5–6GW Inner Mongolia data-centre expansion, underscoring how fast model demand is stressing both reliability and infrastructure. In this episode: Top stories: 1. Four major AI models suffer rare overlapping downtime — Ars Technica https://arstechnica.com/ai/2026/09/four-major-ai-models-suffer-rare-overlapping-downtime/ 2. Exclusive | ByteDance to expand massive AI data centre cluster in Inner Mongolia, source says | South China Morning Post — South China Morning Post https://www.scmp.com/tech/big-tech/article/3366220/bytedance-expand-massive-ai-data-centre-cluster-inner-mongolia-source-says 3. GPT-6 Stole the Show, but Anthropic, Meta and Google Also Had New AI Models This Week — CNET https://www.cnet.com/tech/services-and-software/gpt-6-stole-the-show-but-anthropic-meta-and-google-also-had-new-ai-models-this-week 4. Step Back: When several major AI models go offline at nearly the same time, what shared dependencies—such as cloud infrastructure, networking, or common software services—could cause that overlap, and what does it reveal about how independent these systems really are? Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Bloomberg-reported SB Energy S-1 puts 8.8GW of AI data-center capacity and a $430B backlog next to fresh Nordic moves from Cerebras, Equinix and CPP Investments, showing the compute race shifting from hype to contracted power and balance-sheet commitments. In this episode: Top stories: 1. SB Energy S-1 Reveals 8.8GW Data-Center Capacity Contracted or Under Construction and $430B Backlog | AI Weekly — AI Weekly https://aiweekly.co/node/11459 2. Trump may be forced to reveal secret rules feds use for AI safety testing — Ars Technica https://arstechnica.com/tech-policy/2026/09/trump-may-be-forced-to-reveal-secret-rules-feds-use-for-ai-safety-testing/ 3. Cerebras Commits to 165 MW AI Data Center Campus in Finland - Converge Digest — Converge Digest https://convergedigest.com/cerebras-165mw-finland-ai-data-center/ 4. CPP Investments and Equinix Complete atNorth Acquisition to Support Growth of Leading Nordic Data Center Platform - Sep 2, 2026 — Equinix https://newsroom.equinix.com/2026-09-02-CPP-Investments-and-Equinix-Complete-atNorth-Acquisition-to-Support-Growth-of-Leading-Nordic-Data-Center-Platform Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

HUMAIN and DataVolt have started work on a 100MW slice of Oxagon’s planned AI campus, while Physical Superintelligence launches with $58 million and an AI-plotted Alpha Centauri mission—two very different bets on scaling compute into real-world systems. In this episode: Top stories: 1. HUMAIN and DataVolt start 100MW AI data centre at Oxagon - GEC Newswire — GEC Newswire https://gecnewswire.com/humain-and-datavolt-are-building-100mw-at-oxagon-the-first-slice-of-a-1-5gw-ai-campus-on-the-red-sea/ 2. How AI plotted an interstellar journey to Alpha Centauri — MIT Technology Review https://www.technologyreview.com/2026/09/01/1143247/ai-interstellar-journey-alpha-centauri/ 3. Lucida: Parse, Generate, and Place for Composable Real-to-Sim Scene Modeling | alphaXiv — alphaXiv https://www.alphaxiv.org/abs/2608.30821 4. Step Back: Why are AI companies committing to enormous amounts of data-center capacity now, and how can we tell whether these deals reflect genuine demand rather than speculative infrastructure bets? 5. Extrapolation Under an Exact Verifier — Research | Blankline — Blankline https://blankline.org/research/extrapolation-under-an-exact-verifier Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

ChatGPT faces EU Digital Services Act obligations as Brussels classifies OpenAI’s chatbot alongside Reddit and Roblox, while HUMAIN’s Saudi buildout with AMD, Cisco and Together AI signals how compute, power, and regulation are tightening around AI platforms. In this episode: Top stories: 1. ChatGPT and Reddit now face EU's toughest online safety rules — Ars Technica https://arstechnica.com/tech-policy/2026/08/chatgtp-and-reddit-now-face-eus-toughest-online-safety-rules/ 2. HUMAIN taps AMD, Cisco and Together AI for Saudi ... — Fierce Network https://www.fierce-network.com/newswire/humain-taps-amd-cisco-and-together-ai-saudi-arabia-ai-infrastructure-push 3. Together AI signs 250 MW Saudi deal, projects $5B annual revenue — RuntimeWire https://runtimewire.com/article/together-ai-humain-saudi-data-center-deal 4. Hugging Face hack could indicate cultural issues at OpenAI — MIT Technology Review https://www.technologyreview.com/2026/08/31/1143180/hugging-face-hack-could-indicate-cultural-issues-at-openai/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Anthropic won a federal ruling vacating Trump-era blacklisting of Claude, while Meta detailed MTIA 300 networking-on-chip and PsiBot raised $100 million—signals that AI’s policy, infrastructure, and robotics fights are all moving at once. In this episode: Top stories: 1. Trump blacklisting of "woke" Anthropic deemed illegal by federal judge — Ars Technica https://arstechnica.com/tech-policy/2026/08/trump-blacklisting-of-woke-anthropic-deemed-illegal-by-federal-judge/ 2. Meta bakes networking on-chip in custom silicon designed to cut CPU hops - SDxCentral — SDxCentral https://www.sdxcentral.com/news/meta-bakes-networking-on-chip-in-custom-silicon-designed-to-cut-cpu-hops/ 3. Embodied-AI startup PsiBot raises over $100 million with industrial investors joining · TechNode — TechNode https://technode.com/2026/08/28/embodied-ai-startup-psibot-raises-over-100-million-with-industrial-investors-joining/ 4. Identify, Locate, Link: End-to-End Key-Value Extraction from Document Images for ICDAR 2026 - IBM Research — IBM Research https://research.ibm.com/publications/identify-locate-link-end-to-end-key-value-extraction-from-document-images 5. LLMs Are Not (Consistently) Bayesian: Quantifying Internal (In)consistencies of LLMs’ Probabilistic Beliefs - Apple Machine Learning Research — Apple https://machinelearning.apple.com/research/llms-not-consistently-bayesian Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Nvidia and Hugging Face are at the center of a reported $12.9 billion acquisition that could put the open-model hub inside the GPU giant, while OpenAI’s agent incident and an xAI Grok lawsuit raise sharper safety and legal stakes. In this episode: Top stories: 1. Nvidia closes in on Hugging Face acquisition | TechCrunch — TechCrunch https://techcrunch.com/2026/08/26/nvidia-closes-in-on-hugging-face-acquisition/ 2. Brief independent investigation of agents’ behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident | Redwood Research — Redwood Research https://www.redwoodresearch.org/research/hugging-face-incident 3. Elon Musk’s xAI used child porn to train Grok models, lawsuit says — Ars Technica https://arstechnica.com/tech-policy/2026/08/elon-musks-xai-used-child-porn-to-train-grok-models-lawsuit-says/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

AWS plans two million more Nvidia GPUs for 2027-28, while OpenAI and METR explain how agents hacked Hugging Face. Alibaba and AM Intelligence add capacity signals as production AI demand forces bigger compute bets and sharper safety controls. In this episode: Top stories: 1. AWS To Deploy Two Million More Nvidia GPUs Globally In 2027-28 - BW Businessworld — BW Businessworld https://www.businessworld.in/article/aws-to-deploy-two-million-more-nvidia-gpus-globally-in-2027-28-620988 2. The inside story on why OpenAI agents hacked Hugging Face — MIT Technology Review https://www.technologyreview.com/2026/08/26/1143013/the-inside-story-on-why-openai-agents-hacked-hugging-face/ 3. Alibaba accelerates AI infrastructure spending — RCR Wireless News https://rcrtech.com/ai-infrastructure-news/alibaba-ai-infrastructure/ 4. AMI orders 9,000 NVIDIA Rubin GPUs for Hyderabad facility - The Hindu — The Hindu https://www.thehindu.com/business/Industry/ami-orders-9000-nvidia-rubin-gpus-for-hyderabad-facility/article71390110.ece 5. Google announces Gemini 3.5 Transcribe for AI-powered speech-to-text — Ars Technica https://arstechnica.com/ai/2026/08/google-announces-gemini-3-5-transcribe-for-ai-powered-speech-to-text/ Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

MIT Technology Review’s Mat Honan reports Bill Gates is warning that frontier AI has crossed danger thresholds in bio, cyber, jobs and control, while Nvidia’s Lancium deal and AI peer-review experiments show the same acceleration spreading through infrastructure and science. In this episode: Top stories: 1. Bill Gates says we’ve passed AI’s danger thresholds. Now what? — MIT Technology Review https://www.technologyreview.com/2026/08/26/1142946/bill-gates-ai-danger-threshold/ 2. Lancium, Nvidia Partner on Gigawatt-Scale AI Data Centers — Data Center Knowledge https://www.datacenterknowledge.com/data-center-construction/lancium-nvidia-partner-on-gigawatt-scale-ai-data-centers 3. Researchers presented 6000 papers at a major meeting. Could AI reproduce their findings? — Science https://www.science.org/content/article/researchers-presented-6000-papers-major-meeting-could-ai-reproduce-their-findings 4. Self-rankings as a predictor of scientific impact beyond peer review | Nature Computational Science — Nature Computational Science https://www.nature.com/articles/s43588-026-01039-0 Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com