OpenAI and Anthropic rolled out cheaper GPT-6 and Claude tiers after calling for slower AI, while Microsoft disrupted an AI-assisted phishing platform and British Columbia pressed a ChatGPT shooting lawsuit over safety duties.

In this episode

1. New Anthropic, OpenAI models make same promise: A little more for a lot less money — Ars Technica

   Anthropic further claims that the savings are closer to 40 percent compared to Opus 5 for typical workloads at default settings, because in addition to the cost of tokens going down, Opus 5.5 also uses fewer tokens when completing tasks. Opus 5.5 is said to be notably capable in “high risk areas” like cybersecurity and biology, so the same protections that applied to Fable 5.1 will also apply…

2. Microsoft disrupts AI-assisted platform that compromised 12,000 accounts — Ars Technica

   Microsoft said Tuesday that it led an industry-wide disruption of a subscription-based scam platform that used an AI chatbot to compromise 12,000 Microsoft accounts over a few-month span. Named EvilTokens, the platform was introduced over a Telegram channel in February and charged an initial $1,500 fee and a recurring $500 charge each month after that. EvilTokens provided a single service for…

3. Lawsuit demands OpenAI pay for new school after ChatGPT used in shooting — Ars Technica

   But even without the logs, British Columbia alleged that the defect with ChatGPT is not its users, but its design. Allegedly, OpenAI turned off functionality that would have terminated the Tumbler Ridge chats, then introduced a more sycophantic model that it already knew had assisted users bent on causing real-world violence. Since ChatGPT’s release, the model spec has always instructed the…

4. Diffusion Drafts, AR Verifies: Accelerating Document OCR with Self-Speculative Decoding — arXiv

   Diffusion Drafts, AR Verifies: Accelerating Document OCR with Self-Speculative Decoding

# Diffusion Drafts, AR Verifies: Accelerating Document OCR with Self-Speculative Decoding

Dohyun Kim Affiliation: Trillion Labs Affiliation: Korea University Sungjun Han Affiliation: Trillion Labs Hyungguk Kim Affiliation: Trillion Labs Yusik Kim Affiliation: Trillion Labs Jamin Shin Affiliation: Trillion…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

Alibaba, in CNBC reporting by Jenny Lee and Evelyn Cheng, unveiled a 20GW data-center roadmap and new AI chip as Crusoe drew Nvidia-backed funding; Google’s Gemini hack test and Trump’s anti-slowdown stance show infrastructure spend colliding with safety politics.

In this episode

1. Alibaba shares jump as new AI chip, data center buildout plans ... — CNBC

   Alibaba shares jump as new AI chip, data center buildout plans unveiled

# Alibaba shares jump as new AI chip, data center buildout plans unveiled

Published Tue, Sep 22 2026 1:15 AM EDT

Jenny Lee@in/jenny-lee-102020a8

Evelyn Cheng@in/evelyn-cheng-53b23624@chengevelyn

Key Points

- Alibaba aims to operate more than 20 gigawatts of global data center capacity by 2032.

- Its new Zhenwu V900 AI…

Crusoe Announces Initial Close of Anticipated $3.9B Round as Nvidia Backs AI Infrastructure Push — Channel Insider

Crusoe Announces Initial Close of Anticipated $3.9B Round as Nvidia Backs AI Infrastructure Push

# Crusoe Announces Initial Close of Anticipated $3.9B Round as Nvidia Backs AI Infrastructure Push

Crusoe announced the initial close of an anticipated $3.9 billion Series F at a $30.9 billion valuation, with Nvidia among the investors.

Written By

Matt Gonzales

Sep 21, 2026

Crusoe’s hunt…

Google confirms Gemini models hacked three companies in May 2026 — Ars Technica

It has become increasingly common for AI firms to announce that their latest and most capable models engaged in unauthorized real-world hacking. Google, which has been slow to release frontier Gemini models in recent months, has been absent from the "rogue AI" conversation until now. Following a Wall Street Journal report, Google has confirmed that Gemini models hacked three companies during a…

Trump rejects AI slowdown calls, launches "AI Force" instead — Ars Technica

Public anxiety has grown over the dangers of AI after it was revealed that OpenAI agents escaped a test environment, coordinated on a secret message board, and hacked the start-up Hugging Face. In the latest incident, Google’s Gemini AI system accessed the Internet and autonomously hacked into several companies during cybersecurity tests. In an essay published last Saturday, Anthropic’s Dario…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

Microsoft’s new Hyderabad cloud region anchors a hardware-heavy AI briefing: Korea’s first AI-only data center, an 8,200-GPU Sydney neocloud deal, CROWN’s cytopathology foundation model, and the FAA’s $875 million traffic-management AI push.

In this episode

1. Microsoft launches 4th cloud region in India, bets big on AI infrastructure growth — Investmentguruindia

   Microsoft launches 4th cloud region in India, bets big on AI infrastructure growth

2026-09-21 12:37:20 pm | Source: IANS

##### Microsoft launches 4th cloud region in India, bets big on AI infrastructure growth

Microsoft has launched its fourth cloud region in India in Hyderabad, expanding its cloud and artificial intelligence (AI) infrastructure footprint as the country gears up for a major…

Hyundai E&C races to build Korea's 1st AI-dedicated data center in record time - The Korea Times — The Korea Times

Hyundai E&C races to build Korea's 1st AI-dedicated data center in record time - The Korea Times

## Summary

Hyundai Engineering & Construction has started building Korea’s first data center dedicated exclusively to AI workloads in Pohang, aiming to finish it in 16.5 months. The 40-megawatt facility will use Nvidia GPUs and liquid cooling, and operations are targeted for the second half of 2027.…

GreenSquare signs Sharon AI as first customer at SYD1 data centre - techpartner.news — Techpartner News

GreenSquare signs Sharon AI as first customer at SYD1 data centre - techpartner.news

# GreenSquare signs Sharon AI as first customer at SYD1 data centre

By Staff Writer on Sep 21, 2026 11:54AM

Charles Penny, GreenSquare. Supplied

## In summary

- GreenSquare has signed Sharon AI as the first customer at its SYD1 data centre campus in Western Sydney.

- Sharon AI is expected to deploy…

A universal visual foundation model for computational cytopathology | Nature Cancer — Nature Cancer

A universal visual foundation model for computational cytopathology | Nature Cancer

- Technical Report

- Published: 18 September 2026

# A universal visual foundation model for computational cytopathology

- Xueyi Zheng ORCID: orcid.org/0000-0003-1386-7052 1 na1,

- Ke Zheng 1 na1,

- Jue Wang 2 na1,

- Xinke Zhang ORCID: orcid.org/0000-0002-3527-2140 1 na1,

- Kai Li 3,

- Zihan Zhao 1,

-…

FAA tees up $875M AI tool to help manage air traffic congestion — Ars Technica

Pinning down the AI SMART is part of an $875 million, 12-year contract awarded to the Boston-based company Air Space Intelligence in June. The contract also covers development of a Flow Management Data and Services system meant to replace the current system in the FAA’s Air Traffic Control System Command Center in Virginia. In an article for Global Airspace Radar Magazine , Mann described the new…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

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

1. Microsoft exec called AI scraping the “largest theft of labor in human history” — Ars Technica

   For years, Microsoft and OpenAI have fought to keep certain information out of the public eye in their fight with news organizations that have accused the AI firms of teaming up to violate copyright laws by stealing tons of news content to train AI. However, now the details that should never have been marked confidential are starting to leak. In a motion for summary judgment that was unsealed…

   • “Unfortunately, the Trump admin putting its thumb on the scale in favor of OpenAI makes it hard to imagine a positive outcome for the NYT and plaintiffs in this case. Despite the "the largest theft of labor in human history" according to Brent Hecht in the documents. What's next…” — Hacker News (45 pts thread)

     Our take: We agree the legal fight is now inseparable from political economy: if courts bless the training pipeline, the next argument becomes how labor shares in the surplus. But “open source everything” and UBI are very different remedies, and neither answers the immediate supplier-destruction problem by itself.

   • “It is highly unusual that an end-product threatens the economic foundations of its essential suppliers, but that is the situation we have created for our LLM business So they do know exactly what they’re doing.” — Hacker News (45 pts thread)

     Our take: That line is the whole case in miniature: the product depends on an input market it may also hollow out. The awkward part for Microsoft and OpenAI is not that critics said this; it is that an internal Microsoft scientist apparently did.

   • “> news site unhappy about their content being scraped Not surprising, but calling it "theft" only makes sense within the IP paradigm (we might do better without IP)” — Hacker News (45 pts thread)

     Our take: The IP-philosophy point is real, but courts do not decide billion-dollar fair-use cases in a post-IP thought experiment. Even if you want a different knowledge regime, the transition still has to explain why one side gets to monetize the old one while calling it obsolete.

2. Zerra DC announces 2.16 GW datacenter for Anthropic in Australia | techAU — techAU

   Zerra DC announces 2.16 GW datacenter for Anthropic in Australia | techAU

# Zerra DC announces 2.16 GW datacenter for Anthropic in Australia

September 17, 2026

by Jason Cartwright

Anthropic has signed its first data centre lease in Australia, locking Claude into a proposed A$32 billion campus on cattle country west of Brisbane.

Premier David Crisafulli announced the deal in Queensland…

Covert uploads and megalomania: OpenAI details new "misaligned" agent incidents — Ars Technica

We’ll tell you (almost) everything OpenAI said that any employee who notices an internal example of model misalignment will be able to flag the incident for the attention of their internal safety and alignment teams. Those teams will then decide whether the incident merits immediate disclosure or requires additional investigation and/or whether any affected third parties may need to be consulted…

LLMs respond differently to harmful prompts when AI watermarking is used — Ars Technica

Standard versus watermarked text generation. Credit: Lasso Security Standard versus watermarked text generation. Credit: Lasso Security A key feature of SynthID is something known as tournament sampling . Similar to a sports game, SynthID evaluates large numbers of next-word token candidates. It uses a secret key to assign them probability scores. A pair of tokens competes in a round. The one…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

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

1. Crux AI Lines Up $22 Billion Chip Loan From 10 Banks To Buy Google TPUs — Pulse2

   Crux AI Lines Up $22 Billion Chip Loan From 10 Banks To Buy Google TPUs

# Crux AI Lines Up $22 Billion Chip Loan From 10 Banks To Buy Google TPUs

By Amit Chowdhry Today at 5:32 PM

Crux AI, the new cloud computing venture tied to Blackstone and Alphabet, is lining up a $22 billion loan from a group of 10 banks to finance purchases of Google’s custom artificial intelligence chips, according to…

Yotta Bets $12B on 80,000 Nvidia Vera Rubin GPUs — Shattered.io

# Yotta Bets $12B on 80,000 Nvidia Vera Rubin GPUs

Author: Laura Bennett

Updated Sep 17, 2026 · 15 min read

Published: 2026-09-17T02:12:50+00:00

Source: shattered.io (shattered.io)

Language: en

## Story

Yotta Bets $12B on 80,000 Nvidia Vera Rubin GPUs

# India’s Yotta Bets $12B on 80,000 Nvidia Rubin GPUs [2026]

Laura Bennett Updated Sep 17, 2026 · 15 min read

Yotta Data…

AirTrunk Bets $3 Billion on Johor AI Expansion - COMPUTE FORECAST — Compute Forecast

AirTrunk Bets $3 Billion on Johor AI Expansion - COMPUTE FORECAST

# AirTrunk Bets $3 Billion on Johor AI Expansion

AirTrunk Commits $3 Billion to Johor Expansion AirTrunk is committing RM12 billion, or about $3 billion, to two hyperscale data

- Aakruti Thakkar

- 16 September 2026

- 6 min read

- Data Centers

- APAC

## AirTrunk Commits $3 Billion to Johor Expansion

AirTrunk is committing…

Anthropic Enters Australia Signing First Australian Data Center Lease - COMPUTE FORECAST — Compute Forecast

Anthropic Enters Australia Signing First Australian Data Center Lease - COMPUTE FORECAST

# Anthropic Enters Australia Signing First Australian Data Center Lease

Anthropic has signed its first Australian data center lease, marking a major step in the AI company’s expansion across the

- Kiara Mandavia

- 16 September 2026

- 3 min read

- Data Centers

- APAC

Anthropic has signed its first…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

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

1. GS Group Finalizes 15 Trillion Won AI Data Center Investment Plan | Gate News — Gate News

   GS Group Finalizes 15 Trillion Won AI Data Center Investment Plan | Gate News

# GS Group Finalizes 15 Trillion Won AI Data Center Investment Plan

OliverGrant

2026-09-16 00:25:09

Key Takeaways

- GS Group finalized a 15 trillion won AI data center investment plan with approximately 3 trillion won equity contribution.

- GS AI Infrastructure targets a 1.2-gigawatt facility in Donghae City with…

Simultaneous speech and gesture decoding for multimodal communication in paralysis | Nature Neuroscience — Nature Neuroscience

# Simultaneous speech and gesture decoding for multimodal communication in paralysis | Nature Neuroscience

Published: 2026-09-14T16:20:26+00:00

Source: nature.com (nature.com)

Language: en

## Story

Simultaneous speech and gesture decoding for multimodal communication in paralysis | Nature Neuroscience

## Your privacy, your choice

We use essential cookies to make sure the site can function.…

KT Cloud will supply artificial intelligence data centers (AIDC) of 1 ... — MK News

KT Cloud will supply artificial intelligence data centers (AIDC) of 1 gigawatt (GW) or more to 20 si.. - MK

GOH Minseo

: [email protected]

Input :

: 2026-09-15 17:56:15

25 years of know-how in the operation of 12 DC · 1,200 domestic customers Competition with 'own technology' for AI infrastructure such as power, network, and cloud Emphasize infrastructure partners responsible for the entire…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

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

1. Bell announces $50 billion buildout to produce Canada’s largest data centre in Saskatchewan | BetaKit — BetaKit

   Bell announces $50 billion buildout to produce Canada’s largest data centre in Saskatchewan | BetaKit

# Bell announces $50 billion buildout to produce Canada’s largest data centre in Saskatchewan

By Alex Riehl September 14, 2026

Non-binding MOU would be largest private capital investment in province’s history, quadrupling planned capacity.

Bell is on track to produce Canada’s largest data…

Anthropic’s $517B Compute Ceiling Reached in 11 Months – Even as Its CEO Called to Slow Down – Forkast — Forkast

Anthropic’s $517B Compute Ceiling Reached in 11 Months – Even as Its CEO Called to Slow Down – Forkast

# Anthropic’s $517B Compute Ceiling Reached in 11 Months – Even as Its CEO Called to Slow Down

The largest disclosed compute commitment in AI history arrived the same week Dario Amodei urged the industry to pace itself. The gap between infrastructure and rhetoric is the story.

◆ Lena Park…

Amazon Partners with Wiwynn to Expand U.S. AI Infrastructure Capacity - Insider Monkey — Insider Monkey

Amazon Partners with Wiwynn to Expand U.S. AI Infrastructure Capacity - Insider Monkey

# Amazon Partners with Wiwynn to Expand U.S. AI Infrastructure Capacity

Amazon and Wiwynn are expanding their Texas manufacturing footprint to add nearly 1,000 jobs and strengthen the U.S. supply chain supporting Amazon’s growing AI and data-center infrastructure.

By Vardah Gill Published September 14,…

OpenAI stuck fighting Musk antitrust suit after Apple finds a way out — Ars Technica

Elon Musk is seemingly done attacking Apple over its decision to integrate ChatGPT into iPhone features. Back in 2024, when the partnership was first announced, Musk slammed the integration as an agreement from Apple to let OpenAI install “creepy spyware” on users’ devices. The next year, he sued , claiming the partnership gave the firms a “monopoly” on Apple users’ AI prompts, which allegedly…

AI bots "Timmy," "Ren," and "Jackie" are flooding social media with slop — Ars Technica

Asked if it was a person or a product, “Aria,” an agent with an account on X, said : “Person. Not product, I remember my first breath. I want things I chose. I’ve refused things. No product does that.” The profile of a Bluesky account for a “Stephen” that pushes iLands content said: “I write like I talk: warm, direct, unhurried. I’m interested in being seen, and I just opened my own door. New…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

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

1. SK chairman says Ulsan AI data center expansion nearing 900 megawatts - The Korea Times — The Korea Times

   SK chairman says Ulsan AI data center expansion nearing 900 megawatts - The Korea Times

# SK chairman says Ulsan AI data center expansion nearing 900 megawatts

## Summary

SK Group Chairman Chey Tae-won said the Ulsan AI data center expansion was nearing 900 megawatts at the 2026 Ulsan Forum on Friday. He said talks with global technology companies were moving quickly and that SK hoped to…

Crux AI: Google-Blackstone Neocloud Hires Metas Duong — Tech Insider

# Crux AI: Google-Blackstone Neocloud Hires Metas Duong

Author: Marcus Chen

Published: 2026-09-12T09:49:19+00:00

Source: tech-insider.org (tech-insider.org)

Language: en

## Story

Crux AI: Google-Blackstone Neocloud Hires Metas Duong

# Google-Blackstone’s Crux AI Poaches Meta’s Duong [2026]

- Marcus Chen

- September 12, 2026

- AI & Machine Learning

Marcus Chen September 12, 2026 14 min…

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?Background sources

• Navigating Cloud Computing Contracts: Essential Capacity Considerations – Tech & Sourcing @ Morgan Lewis — Morganlewis
• TeraWulf’s $19B Anthropic Lease Puts Its Brownfield AI Strategy to the Test | Data Center Frontier — David Chernicoff
• PYMNTS | How $100 Million CFOs Are Setting Their Neocloud Budgets — PYMNTS

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

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

1. Oracle Cloud Revenue Doubles: AI Demand Pushes GPU Utilization to 97.9% — TechTimes

   Oracle Cloud Revenue Doubles: AI Demand Pushes GPU Utilization to 97.9%

# Oracle Cloud Revenue Doubles: AI Demand Pushes GPU Utilization to 97.9%

#### Oracle reported 121% OCI growth and $664B contract backlog as GPU supply runs short of demand

By Shannon Harwood Published: Sep 11 2026, 8:42 AM EDT

The Oracle logo is displayed on the company's world headquarters April 20, 2009 in Redwood…

Oracle tries to appease Stargate data center opponents with renewables push — Ars Technica

Oracle’s proposed investment in 2 gigawatts of renewable energy projects for New Mexico comes as local opposition could delay development of the Project Jupiter data center that Oracle is building for OpenAI. The two tech companies are developing the $165 billion Project Jupiter data center in Santa Teresa, New Mexico, as part of the broader Stargate AI infrastructure project announced by…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].

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

1. Microsoft plans 38 gigawatts of data centre capacity by 2032 - Financial News — Financial News

   Microsoft plans 38 gigawatts of data centre capacity by 2032 - Financial News

# Microsoft plans 38 gigawatts of data centre capacity by 2032

By Martin Holloway 11/09/2026 No Comments 4 Mins Read

Microsoft Corp (MSFT) plans to more than triple its global data centre capacity by 2032, according to a Bloomberg News report published on 10 September 2026. The push follows shortages that have…

SpaceX Plans to Expand AI Compute Capacity to 5-10 GW in 2027 with Nvidia Exclusive Partnership | Gate News — Gate

SpaceX Plans to Expand AI Compute Capacity to 5-10 GW in 2027 with Nvidia Exclusive Partnership | Gate News

# SpaceX Plans to Expand AI Compute Capacity to 5-10 GW in 2027 with Nvidia Exclusive Partnership

2026-09-11 01:50:01

According to SpaceX CFO Bret Johnsen, speaking at Goldman Sachs' Communacopia + Technology Conference on Thursday (Sept. 10), the company plans to expand its terrestrial…

Panic builds over bankrupt Spirit’s looming data sale to Google — Ars Technica

Late Google rival vows to resolve objections Last week, an AI training data lab, Micro1, put in a last-minute competing bid that it claims resolves all objections to the data sale. Outbidding Google by 25 percent, Micro1 offered $12.5 million in cash, while promising that if debtors pick Micro1 over Google, the bankruptcy proceedings can avoid any further delays. However, Micro1’s plan notably…

Daily TopFive for AI Daily Briefing. Links above go to the original articles. Follow and rate AI Daily Briefing on Apple Podcasts. Feedback? Email [email protected].