AI Daily Briefing

Anthropic’s Texas compute push is nearing a $15 billion Nexus financing backstopped by Google, CNBC’s Ashley Capoot and Hugh Son report, while Amazon, Meta, AMD and Georgia utilities show the same constraint: AI capacity is now a balance-sheet and power-grid fight.

In this episode:
Top stories:
1. Nexus Data Centers in advanced talks to secure $15B for Google-backed Anthropic data center — CNBC
https://www.cnbc.com/2026/07/30/nexus-data-centers-in-advanced-talks-to-secure-15b-for-google-backed-anthropic-data-center.html
2. Amazon Boosts CAPEX Plan to $220B Amid AI Capacity Constraints - Converge Digest — Converge Digest
https://convergedigest.com/amazon-boosts-capex-plan-to-220b-amid-ai-capacity-constraints/
3. Meta's Future Data Center Leases Top a Quarter-Trillion Dollars - Business Insider — Business Insider
https://www.businessinsider.com/meta-future-ai-data-center-leases-quarter-trillion-dollars-2026-7
4. AMD Secures Up to 2.5 GW of Capacity Through Core Scientific Partnership - Converge Digest — Converge Digest
https://convergedigest.com/amd-secures-up-to-2-5-gw-of-capacity-through-core-scientific-partnership/
5. Southern Company’s data center pipeline keeps filling up — Atlanta Journal-Constitution
https://www.ajc.com/business/2026/07/southern-companys-data-center-pipeline-keeps-filling-up/

Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Meta’s Mark Zuckerberg is weighing whether to sell spare AI compute as capex burns cash, CNBC’s Jonathan Vanian reports; meanwhile Brookfield, NextEra, NAVER and Nebius are sketching gigawatt-scale campuses that turn the infrastructure race into a power-market story.

In this episode:
Top stories:
1. Zuckerberg Meta's AI capacity dilemma: What to sell vs. what to keep — CNBC
https://www.cnbc.com/2026/07/29/zuckerberg-metas-ai-capacity-dilemma-what-to-sell-vs-what-to-keep.html
2. $100 Billion AI Data Center Campus Planned at DOE Paducah Site with 4.6 GW Dedicated Power Buildout — Construction Review Online
https://constructionreviewonline.com/100-billion-ai-data-center-campus-planned-at-doe-paducah-site-with-4-6-gw-dedicated-power-buildout/
3. NAVER Targets 1 GW Sovereign AI Infrastructure in Korea - Converge Digest — Converge Digest
https://convergedigest.com/naver-targets-1-gw-sovereign-ai-infrastructure-in-korea/
4. Nebius Unveils Plans for 1.2GW AI Data Center Campus in Pennsylvania - DataMagz — DataMagz
https://datamagz.com/nebius-unveils-plans-for-1-2gw-ai-data-center-campus-in-pennsylvania/
5. [2607.25991v1] Phantom Evidence: How and Why Generative AI Manufactures False Positives in Science — Arxiv
https://arxiv.org/abs/2607.25991v1

Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

Nvidia signed leases worth up to $50 billion for Hut 8’s Texas AI campus, deepening its move from GPU supplier into infrastructure operator as AI compute demand keeps swallowing power, land, and capital.

In this episode:
Top stories:
1. Nvidia signs leases worth up to $50 billion for massive Texas data center housing hundreds of thousands of GPUs — CryptoBriefing
https://cryptobriefing.com/nvidia-texas-data-center-hut8-gpu-leases/
2. The Observer Was the Bet - Bret kerr — Bret Kerr
https://bretkerr.substack.com/p/the-observer-was-the-bet
3. Why RL Training Collapses on Long-Horizon Agents - DEV Community — DEV Community
https://dev.to/cxw_orkas/why-rl-training-collapses-on-long-horizon-agents-229f
4. [2607.24131v1] MAPLE: Efficient and Diverse Multi-Alpha Generation for Portfolio Construction — Arxiv
https://arxiv.org/abs/2607.24131v1

Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

OpenAI and Nvidia’s Wall Street Journal-reported talks over a $250 billion Ohio data-center guarantee anchor a wave of AI-infrastructure financing, from Blackwell GPU clusters to SK Telecom’s new AI data-center arm and India/Japan buildouts. The stakes are power, leases, and balance sheets.

In this episode:
Top stories:
1. Nvidia plans $250bn push to bolster OpenAI’s infrastructure ambitions | Technology News | Al Jazeera — Al Jazeera
https://www.aljazeera.com/economy/2026/7/27/nvidia-plans-250bn-push-to-bolster-openais-infrastructure-ambitions
2. Axe Compute wins $1.5B AI infrastructure contract | AGPU 8-K Filing — Stock Titan
https://www.stocktitan.net/sec-filings/AGPU/8-k-axe-compute-inc-reports-material-event-a6704d207de2.html
3. SK Telecom launches dedicated AI infra company — RCR Wireless News
https://www.rcrwireless.com/20260727/carriers/sk-telecom-ai-infra
4. CtrlS secures Rs 5,500 crore bank funding for Hyderabad data centre project - Business News | The Financial Express — The Financial Express
https://www.financialexpress.com/business/news/ctrls-secures-rs-5500-crore-bank-funding-for-hyderabad-data-centre-project/4302982/
5. Japan Funds Noetra’s 140MW Physical AI Data Centre | SendTech Times — SendTech Times
https://stechtimes.com/en/article/japan-funds-noetras-140mw-physical-ai-data-centre-mrxe4zu6

Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com

NVIDIA’s Korea AI factory push leads a day of execution checks: NAVER’s Brookfield-backed 200-megawatt plan, Moonshot’s still-pending Kimi K3 weights, and open releases from Open Dreamer and Meituan’s LongCat-2.0.

In this episode:
Top stories:
1. NAVER and NVIDIA Triple Korea's AI Factory to 200 Megawatts in $10 Billion Deal With Brookfield — TechTimes
https://www.techtimes.com/articles/321630/20260726/naver-nvidia-triple-koreas-ai-factory-200-megawatts-10-billion-deal-brookfield.htm
2. Kimi K3's Open Weights Are Still a Countdown, Not a Download - DEV Community — DEV Community
https://dev.to/breachprotocol/kimi-k3s-open-weights-are-still-a-countdown-not-a-download-4h40
3. Meet Open Dreamer: A JAX/Flax Reproduction of the Dreamer 4 World Model — The News 92
https://thenews92.com/meet-open-dreamer-a-jax-flax-reproduction-of-the-dreamer-4-world-model-pipeline-with-the-full-training-recipe-published/
4. Meituan LongCat-2.0: 1.6T Model for Agentic Coding Released | AIToolly — AIToolly
https://aitoolly.com/ai-news/article/2026-07-26-meituan-open-sources-longcat-20-a-16t-parameter-model-optimized-for-agentic-coding-and-domestic-gpu

Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email ai-daily-briefing@lanternpodcasts.com