惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

J
Java Code Geeks
F
Fortinet All Blogs
Martin Fowler
Martin Fowler
M
MIT News - Artificial intelligence
G
Google Developers Blog
P
Proofpoint News Feed
Recent Announcements
Recent Announcements
MyScale Blog
MyScale Blog
D
DataBreaches.Net
Stack Overflow Blog
Stack Overflow Blog
月光博客
月光博客
爱范儿
爱范儿
罗磊的独立博客
腾讯CDC
Hugging Face - Blog
Hugging Face - Blog
博客园 - 叶小钗
Vercel News
Vercel News
酷 壳 – CoolShell
酷 壳 – CoolShell
B
Blog
C
Check Point Blog
美团技术团队
宝玉的分享
宝玉的分享
Microsoft Security Blog
Microsoft Security Blog
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻

Hacker News

GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis Bonsai 1-bit WebGPU - a Hugging Face Space by webml-community Moving a large-scale metrics pipeline from StatsD to OpenTelemetry / Prometheus GitHub - Nightmare-Eclipse/RedSun: The Red Sun vulnerability repository GitHub - SethPyle376/hiraeth: Local AWS emulator focused on fast integration testing, with SQS support, SQLite-backed state, and a debug-friendly web UI. GitHub - macOS26/Agent: Any AI, replaces Claude Code, Cursor, OpenClaw. Over 18 LLM providers (Claude, OpenAI, Gemini, Ollama, Zai, HF, Qwen) wired into a native Mac app that writes code, builds Xcode projects, bumps versions, manages git, automates Safari, use AppleScript, JS or Accessibility, extend Agent! w/ MCP Servers, run tasks from your iPhone via Messages. YouTube now lets you turn off Shorts I Made a Terminal Pager Burgers | マクドナルド公式 Commands — HackerNews CLI documentation ChatGPT for Excel PiCore - Raspberry Pi Port of Tiny Core Linux Live Nation illegally monopolized ticketing market, jury finds Google Broke Its Promise to Me. Now ICE Has My Data. Founding Engineer at Adaptional | Y Combinator CRISPR takes important step toward silencing Down syndrome’s extra chromosome GitHub - saffron-health/libretto: The AI toolkit for building reliable browser automations US v. Heppner (S.D.N.Y. 2026) no attorney-client privilege for AI chats [pdf] Retrofitting JIT Compilers into C Interpreters IPv6 – Google The Accursèd Alphabetical Clock Cybersecurity Looks Like Proof of Work Now Fragments: April 14 Cal.com Goes Closed Source: Why AI Security Is Forcing Our Decision | Cal.com - Scheduling Software for Online Bookings Laravel raised money and now injects ads directly into your agent When moving fast, talking is the first thing to break Too much Discussion of the XOR swap trick – Heather Cafe Introduction to Spherical Harmonics for Graphics Programmers The Grand Line
India’s AI deal with the UAE challenges U.S. cloud dominance
Indranil Ghosh · 2026-06-15 · via Hacker News

India is partnering with the United Arab Emirates to ease the grip of Amazon, Microsoft, and Google on artificial intelligence computing.

G42, backed by Abu Dhabi sovereign wealth fund Mubadala, signed an agreement on May 15 to deploy an AI supercomputer in India, comprising 64 systems made by U.S. chipmaker Cerebras. A G42 unit will handle installation, operations, and maintenance, with Cerebras providing technical support, G42 said in response to questions from Rest of World.

Any government that wants to use AI today typically rents computing power from Amazon, Microsoft, or Google. India already has at least $45 billion in commitments from the three companies. The country’s $1.25 billion national AI program runs entirely on Nvidia processors, with 34,000 available to researchers and businesses, and a target of 100,000 by the year’s end.

The G42 deal adds a second path, where India will have machines on its own soil, under its own rules, run by a non-U.S. partner.

This is an example of India’s pragmatic approach to AI sovereignty, using the power of its scale.”Cameron Kerry, former acting secretary at the U.S. Department of Commerce

“This is an example of India’s pragmatic approach to AI sovereignty, using the power of its scale to adapt what’s available from other countries, whether AI leaders like China and the U.S. or others, to adapt to its own needs,” Cameron Kerry, former acting secretary at the U.S. Department of Commerce, told Rest of World.

Kerry, now a fellow at the Brookings Institution, co-authored a report in February, arguing that no country has full control over every element of AI and that governments must assemble capabilities from multiple partners.

G42 has been working on what it calls the Intelligence Grid, a global network of AI facilities that it builds, owns, and operates for governments. India is the first country to sign up. Discussions are on with other governments, G42 said, without naming the countries.

India’s autonomous scientific society, the Centre for Development of Advanced Computing, will work with G42’s Core42 unit, with all data remaining under Indian governance rules. G42 declined to disclose the financial terms of the India deal or confirm who will own the hardware after installation.

India already has substantial AI infrastructure commitments from U.S. companies: Microsoft plans to invest $17.5 billion over four years, Google has pledged $15 billion, and Amazon Web Services has earmarked $12.7 billion, all built around Nvidia processors and the companies’ own cloud platforms. 

G42 faces a steep challenge entering India, Chris Miller, professor at the Fletcher School at Tufts University and an expert on global semiconductor competition, told Rest of World. Amazon, Microsoft, and Google offer integrated packages of hardware, software, developer tools, and customer support that any newcomer has to match.

“A key question will be whether they can offer competitive software and other services to their data-center customers,” Miller said.

Cerebras chips

Cerebras, founded in 2016, makes AI chips and the supercomputer systems they power. Its AI chip is the largest in the world — a single piece of silicon the size of a dinner plate that accomplishes what Nvidia needs thousands of smaller processors wired together to do. 

The choice of Cerebras appears driven by practical considerations, Kerry said.

Nvidia’s most advanced processors are designed for training large AI models from scratch. Cerebras is built for speed in running AI applications, which matches India’s focus on deploying AI across healthcare, agriculture, and public services.

G42 became Cerebras’s largest customer in 2021, and from 2023, the two companies built three supercomputer facilities in California, Texas, and Minnesota, called Condor Galaxy. The Condor Galaxy partnership gave G42 hands-on experience in deploying and maintaining Cerebras hardware across multiple locations and regulatory environments, G42 said.

Cerebras counts G42 and the UAE’s Mohamed bin Zayed University of Artificial Intelligence as its two largest customers. It went public on the Nasdaq on May 14 and raised $5.55 billion in the biggest U.S. tech stock offering since Uber in 2019. Together, the two UAE buyers accounted for 86% of Cerebras’ 2025 revenue, according to its SEC filing.

At home in the UAE, G42 works with the same U.S. companies to which it offers an alternative abroad. Amazon operates a full cloud region in the UAE, and Microsoft has committed $15.2 billion to UAE data center expansion through 2029, working through G42’s subsidiary Khazna.

The deal may give India less control than it appears, Kerry said. India’s laws allow personal data to be sent to most countries and restrict transfers to a short list of blacklisted nations, according to the Digital Personal Data Protection Act, 2023. Whether India’s government has imposed stricter rules for this specific deal is unknown, and G42 declined to share those details.

As more governments seek to own their AI machines rather than rent them, Amazon, Microsoft, and Google will need to find ways to serve those customers or risk losing them to partners like G42, said Kerry. “The better they respond, the better for the U.S. position in the world regardless of what the U.S. government does,” he said.

Additional reporting by Divsha Bhat in Dubai.