🗞️In this edition

• Amazon breaks Microsoft's OpenAI Cloud monopoly

• Microsoft invests $15.2B in UAE AI infrastructure

• DeepSeek + 20 Chinese companies make Time's best inventions

• Workflow Wednesday #43 ‘AI in Action’

• In other AI news –

  • Studio Ghibli and others accuse OpenAI of Sora 2 copyright violations

  • AI browser Dia adds Arc-inspired features with built-in memory

  • Experts say most AI safety tests are unreliable

  • 4 must-try AI tools

OpenAI signs $38 billion AWS deal breaking Microsoft exclusivity

What's happening:

OpenAI struck a $38B deal with Amazon Web Services giving the AI company access to "hundreds of thousands" of Nvidia GPUs to power its models.

The seven-year partnership comes as Microsoft loosens its grip on OpenAI, dropping its status as exclusive cloud provider and losing first right of refusal to host AI workloads.

OpenAI will "immediately" start using AWS compute to train AI models, according to the press release. "All capacity targeted to be deployed before the end of 2026, and the ability to expand further into 2027 and beyond."

Last week, OpenAI completed its for-profit restructuring and announced a new Microsoft deal giving the company rights to OpenAI's technology until it reaches AGI. Under the agreement, OpenAI can work with third parties to develop some AI products and release certain open weight models.

The new deal includes OpenAI's commitment to purchase $250B of Microsoft's Azure services, dwarfing the $38B it will pay Amazon. OpenAI also has a reported $300B contract with Oracle, while Amazon continues pouring billions into Anthropic.

Why this is important:

OpenAI's spreading $588B across three cloud providers: $250B Microsoft, $300B Oracle, $38B Amazon.

Microsoft losing exclusivity is the story. They invested $13B in OpenAI and built custom infrastructure. Now they're one of three major providers instead of the sole partner.

The timing matters. This comes one week after OpenAI's restructuring gave Microsoft extended IP rights through 2032 in exchange for dropping exclusivity requirements. Microsoft took $135B stake but lost monopoly on hosting OpenAI's workloads.

Hundreds of thousands of Nvidia GPUs deployed by end of 2026 represents massive compute capacity. OpenAI's securing infrastructure for the next training runs while model sizes continue scaling.

Amazon's investment in Anthropic while hosting OpenAI creates interesting dynamics. They're backing both leading AI labs simultaneously.

Our personal take on it at OpenTools:

Microsoft's paying the price for loosening control.

They invested early, built custom infrastructure, and were OpenAI's exclusive cloud partner. Now they're getting $250B over seven years but sharing OpenAI with AWS and Oracle.

The $250B Azure commitment is largest, but losing exclusivity means Microsoft can't bank on hosting all future OpenAI workloads. If GPT-6 training happens primarily on AWS or Oracle, Microsoft's infrastructure investment becomes stranded capital.

OpenAI's diversifying cloud providers the same way Anthropic diversified between AWS and Google. No single vendor lock-in. Better negotiating leverage. Redundancy if one relationship sours.

The $588B total across three providers is staggering. That's more than most countries spend on infrastructure. And it's just one AI company's cloud commitments.

Amazon hosting OpenAI while funding Anthropic is hedging bets. If OpenAI wins, AWS wins. If Anthropic wins, Amazon's equity wins. Smart strategy.

Oracle getting $300B is surprising. They're third in cloud market share but securing massive AI workload. Suggests Oracle's offering pricing or terms AWS and Azure won't match.

Microsoft's extended IP rights through 2032 matter more than hosting exclusivity long-term. Rights to the models matter more than where they're trained. But losing hosting monopoly still stings.

Read More

Microsoft invests $15.2 billion in UAE AI infrastructure

What's happening:

Microsoft will invest $15.2B in the United Arab Emirates over four years, the company announced Monday at the first annual Abu Dhabi Global AI Summit. The investment includes first-ever shipments of the most advanced Nvidia GPUs to the UAE.

The US granted Microsoft a license to export Nvidia chips to the UAE, positioning the country as both a proving ground for export-control diplomacy and regional anchor of American AI influence.

The deal allows Microsoft to expand into the Middle East, a key region in the global fight for AI dominance. In May, President Trump struck a deal with UAE president Sheikh Mohamed bin Zayed Al Nahyan to build an AI data center campus in Abu Dhabi. The project was delayed due to US export controls restricting sale of powerful Nvidia chips needed for advanced AI systems.

Microsoft became the first company to receive a license from the US Commerce Department to ship chips to the UAE in September. Critics say the deal undermines the logic of export restrictions to China by introducing possible back channels through a Chinese ally.

Microsoft said it performed substantial work to meet strong cybersecurity and national security conditions required by licenses, enabling the firm to accumulate equivalent of 21,500 Nvidia A100 GPUs in the UAE, based on combination of A100, H100, and H200 chips.

Why this is important:

Microsoft's the first company getting approval to export advanced Nvidia chips to the UAE despite China concerns.

Critics argue UAE's close relationship with China creates backdoor risk. Export controls block China from accessing these chips directly, but UAE could theoretically provide access. Commerce Department approval suggests they accepted Microsoft's security guarantees.

21,500 A100-equivalent GPUs represents significant compute capacity in a region that previously lacked access to frontier AI infrastructure.

Our personal take on it at OpenTools:

Export controls are political theater if Microsoft can ship advanced chips to UAE with Commerce Department blessing.

The entire justification for blocking China is preventing access to frontier AI capabilities. UAE maintains close economic and diplomatic ties with China. If chips flow to UAE, there's plausible pathway to China through partnerships or leaks.

Commerce Department approval suggests they accepted Microsoft's security measures. But cybersecurity is only effective until it isn't. One breach, one insider, one policy shift in UAE, and those chips or the intelligence derived from them could reach China.

UAE's becoming a Middle East AI hub by default. Saudi Arabia, Qatar, and other Gulf states will watch this closely. Expect similar deals across the region.

Read More

Over 20 Chinese companies geatured on Time's best inventions list

What's happening:

Over 20 Chinese companies earned spots on Time magazine's "Best Inventions of 2025" list, with innovations in AI recognized for the first time alongside advances in robotics and consumer electronics.

Products from DeepSeek, Unitree Robotics, and Huawei Technologies were featured on the annual list selected by the magazine's editors.

This year's edition covers 300 innovations across 36 categories, the most ever. It marks the first time Chinese advances in AI have been highlighted since the sector was introduced in 2020.

"The abrupt appearance of DeepSeek's R1 advanced reasoning model at the start of the year was akin to the 'shot heard 'round the world' in AI circles," according to Time.

Released in January, the high-performance, low-cost model from the Hangzhou-based startup rivals some of the most advanced in the US. Its rapid ascent has forced a rethink of China's position in the global AI race.

Seven innovations from the US and Netherlands rounded out the AI category, including Anthropic's Claude Sonnet 4 launched in May, Nvidia's compact DGX Spark supercomputer, and DocuMine, a document analysis tool from Amsterdam-based DataSnipper.

Chinese products returned to the robotics category after two-year absence, claiming three of four spots. That included Unitree Robotics' R1, marketed as "born for sport" for its dynamic movements and one of the world's most affordable humanoids.

Notably, Beijing has vowed to prioritize technological innovation in its coming five-year plan, accelerating development of AI and robotics as crucial emerging industries.

Why this is important:

This is the first time Chinese AI advances made Time's list since the category launched in 2020.

DeepSeek's R1 forcing a "rethink of China's position in the global AI race" is significant validation. The model matched frontier performance at fraction of the cost, challenging assumptions about US AI dominance.

China claiming three of four robotics spots shows where they're pulling ahead. Unitree's R1 being "one of the world's most affordable humanoids" while ready for mass production is competitive advantage. US robotics companies like Boston Dynamics focus on performance. Chinese companies are optimizing for cost and manufacturability.

Our personal take on it at OpenTools:

DeepSeek's recognition validates what the technical community already knew: Chinese AI caught up faster than expected.

R1's "shot heard 'round the world" framing is accurate. The model proved you don't need tens of billions in funding and unrestricted access to Nvidia chips to build frontier AI. That undermines the entire Western strategy of limiting compute access.

China dominating robotics is underreported story. Three of four spots on Time's list. Unitree shipping affordable humanoids at scale. That's not future capability. That's present advantage.

The cost optimization matters more than performance optimization long-term. Boston Dynamics' robots are technically impressive and prohibitively expensive. Unitree's are good enough and affordable. Which strategy wins mass adoption?

Beijing's five-year plan prioritizing AI and robotics means competition intensifies. State backing plus private innovation creates formidable combination.

Time's recognition matters because it's mainstream validation, not just technical community acknowledgment. That shifts narrative from "China's behind" to "China's competitive."

Read More

This Week in Workflow Wednesday #43: AI in Action – Real-World Workflow Transformations

This week, I’ll show you how to use ProWritingAid to take customer-facing text and actually see where readers might lose attention. It’s like running a usability test—on your writing.

Workflow #1: Transform Customer Communication with AI-Powered Writing (ProWritingAid, free trial).

Step 1: Sign up, upload your text, or paste it straight into the editor.
Step 2: Head to the top bar and click the Summary Report (4th from the left). You’ll see a breakdown across grammar, readability, sticky sentences, engagement score—basically the hotspots where readers…We dive into this ProWritingAid workflow and 2 more real-world AI transformations in this week’s Workflow Wednesday.

📢 What else is happening?

• Studio Ghibli, Bandai Namco, Square Enix demand OpenAI stop using their content to train AI – Japanese trade association CODA says Sora 2’s opt-out policy may have violated copyright law.

• Dia’s AI browser starts adding Arc’s ‘greatest hits’ to its feature set – The new AI browser will bring “Arc’s greatest hits” to Dia, including things like the sidebar mode, and combine that with AI-native features like memory and agents.

• Experts find flaws in hundreds of tests that check AI safety and effectiveness – Scientists say almost all have weaknesses in at least one area that can ‘undermine validity of resulting claims’.

👩🏼‍🚒Discover mind-blowing AI tools

1. VideoMyListing - An AI-powered tool that creates engaging videos to promote listings on various platforms like Airbnb, Etsy, and more

2. Kansei.app - An AI language learning platform that offers personalized language practice through immersive conversations with lifelike personas

3. Avaturn - Allows developers to create realistic and customizable 3D avatars for their metaverse, game, or app

4. ReachInbox - An email outreach tool that helps businesses maximize their outreach potential