Rabimba Karanjai @[email protected]

@AnishA_Moonka And how exactly do you get access to it? :) By being the only few companies who can meet with PMO with a big plan but nothing else but connections?

If you're an AI startup in India, renting processing power from the government to train your model costs about $0.7 per hour. The same hardware on Amazon Web Services costs $3.7. On Microsoft Azure, $6.6. The Indian government is subsidizing AI infrastructure at rates that would make most Western startups do a double-take. I read all 26 pages of the white paper this tweet links to. The numbers inside are wild. The IndiaAI Mission has a budget of about $1.2 billion over five years, approved in March 2024. Almost half of that, roughly $500 million, goes straight to building the processing power AI companies need to train their models. The original plan was to deploy 10,000 processors. By December 2025, they had 38,000 running. 3.8x what they promised. A government open call in January 2025 pulled 506 proposals. The four startups picked first were Sarvam AI, Soket AI, Gnani AI, and Gan AI. Eight more were added by September. India now has 12 separate teams building AI models, ranging from tiny ones for basic chatbots to massive ones rivaling those from the US and China. They cover language, voice, vision, medical diagnosis, material science, and even brain-computer interfaces. The one I keep coming back to is Sarvam AI. They raised $41 million from Lightspeed, Peak XV, and Khosla Ventures. In May 2025, they released a model built on top of a French AI system (Mistral Small) and customized for Indian languages. It got roasted online. Critics said it was a foreign model in Indian clothing. So they went back and built Sarvam-105B completely from scratch, using Indian hardware under the government mission. It outperformed China's DeepSeek-R1 on certain tests, even though it was a model six times larger. Both were released for anyone to download and use in March 2026. There's something else buried in the paper I haven't seen another country try at this scale. India is building a copyright system specifically for AI training data. Under a December 2025 government proposal, AI companies can train their models on any copyrighted content they can legally access, books, articles, music, anything. Creators cannot say no. But the moment an AI product makes money, royalties are collected by a centralized government body and distributed back to creators. Singapore allows AI companies to use content without payment. China requires strict consent before training. India is trying a middle path, and publishers are already calling it forced participation. Stanford's AI Vibrancy Index, which measures a country's overall AI strength across research, talent, infrastructure, and investment, ranked India third globally in 2025. Up from seventh in 2023. But the actual scores tell you how far the gap still is: US at 79, China at 37, India at 22. And India's $1.2 billion budget sits next to China's $47.5 billion semiconductor fund and Saudi Arabia's $100 billion Project Transcendence. India is currently spending 40x less than the frontrunners. This white paper is the most detailed public bet yet that smart infrastructure design can close that gap.

Ever wished you could see dense academic papers? 🤯 Google just announced Nano Banana 2 (Gemini 3.1 Flash Image)! It brings insane speed, 4K fidelity & precise subject control to visualize complex science. Check out my early access deep dive: rkrants.blogspot.com/2026/02/visual… #AI #Gemini

@b1ackd0g If its third party, Sui makes the perfect candidate to be that third party.

It's crazy how ahead of its time Libra was. We starting working on it in 2018!

Something I helped built for @Google is live now :) I unfortunately wasn't able to attend the field test day due to @UH_NSM awarding me uh.edu/nsm/news-event… on the same time. But I am incredibly proud of what we all built for @googledevs as @GoogleDevExpert.

I was there as promised ✍️🇳🇬 The future of Blockchain with Life AI will be 100X of scalability @LifeNetwork_AI @rabimba , @EntropianDeFi , @tuan_lifeai

NSM Celebrates 677 Graduates at Spring 2025 Commencement uh.edu/nsm/news-event…

@FrancoisChauba1 @agupta This should also focus on the students who doesnt even have any, and were able to scrape by just applying for cloud compute and basic 4080 and still get papers out as well as awards.

Last night, @agupta and I hosted a great dinner with 14 professors at #NeurIPS2025 from leading academic labs across the US, and many cited compute in academia as "abhorrent". Out of curiosity I just pulled these stats. This is insane. To do meaningful AI research today you need at least 1 GPU/student. Likely 8+ to be honest. The best university (Princeton) is at 0.8 GPUs/student. Stanford is at 0.14 GPUs/student. Marlowe (Stanford's "super cluster") has only 248 H100s for the whole CS Dept to use. Every frontier lab has >100k. This needs to be fixed.

@UHouston @UH_NSM made a story about our recent award at @AIwareConf for our work suported by @SuiNetwork NSM Computer Scientists Earn Distinguished Paper Award for AI Tool That Detects and Repairs Code in Real Time uh.edu/nsm/news-event…

You can see the full project demo and team roster in update here: linkedin.com/posts/ajmirwan… #GoogleCloud #GDE #Gemini #AI

⚙️ The Engineering:We fine-tuned Gemma using QLoRA to understand telemetry tokens. We mapped physics (G-force, yaw) to language to create a "Trustable" racing agent. Read the implementation guide: rkrants.blogspot.com/2025/12/fine-t…

How do you coach a driver at 150mph when a 500ms latency means hitting the wall? 🏎️ For #AISprintH2, I architected a "Split-Brain" AI using Gemini (Strategy) and Gemma (Reflex) to solve high-velocity, safety-critical coaching. Here is the breakdown of how we built it 🧵👇

🏆 Award Winner: "Securing Smart Contract Languages with a Unified Agentic Framework..." Also presenting: 📄 "HPCAgentTester" (AIWare) 📄 "Unlocking On-Chain Intelligence" (BRAINS) Links: brains.dnac.org/2025/ 2025.aiwareconf.org/track/aiware-2… #AI #Blockchain #Research

Excited to announce 3 papers being presented today & tomorrow at #BRAINS2025 (Zurich) and #AIWare2025 (Seoul)! 🌍 I’m not there, but my collaborators are demoing our work. Humbled to share that one of our papers won the ACM SIGSOFT Distinguished Paper Award! 🏆👇

Run massive AI models in VS Code—on free Colab T4 GPUs. 🚀 Forget server bills and local crashes: fine-tune, experiment, even mix quantum and classical ML, all inside your familiar dev setup. #ColabSprint 👇 rkrants.blogspot.com/2025/11/colab-…

Building AI systems that go beyond notebooks, using the new Google Colab VS Code extension. 🚀 Learn how to turn Colab experiments into production-ready workflows inside VS Code, for free. By a @GoogleDevsNA Expert, powered by @googlecolab 👇 @rabimba/the-google-colab-vs-code-extension-beyond-notebooks-building-production-ai-systems-with-free-a46b4a4a41e5" target="_blank" rel="nofollow noopener">medium.com/@rabimba/the-g…

Humbled and honored by this @UHouston! Huge thanks to my professor, collaborators, the @SuiFoundation, and the @GoogleDevsNA for their incredible support in my research journey. uh.edu/nsm/news-event… #AI #Research #GDE #Sui #GoCoogs

Interested to know more about #AI Red Teaming and how you can protect your AI Infra and Models? I'll be speaking at @kc_dc tomorrow on exactly that with some "questionable" demos *wink* Drop by if that's of interest to you #kcdc2025 #gemma @googledevs

@vxunderground Well most people there are sponsored by their companies.

The only thing "BLACKHAT" about the "BLACKHAT" conference is the criminal costs associated with the event. $3,000 for entry? These people have lost the plot.

The only way you can stop @Apple and Siri to not send your data, in plainly , not being part of private compute. And by data, your precise location Everytime you use your iPhone, any app you use including your message, documents, communication and text. Disable Dictation