News

Image on Unsplash Chinese tech companies including Alibaba and ByteDance are conducting training for their most recent artificial intelligence models in data centers across Southeast Asia now. The reason for this is because The United States has imposed strict export restrictions on advanced semiconductors to prevent China from developing cutting-edge AI technology. But these restrictions apply within China. They do not apply in Southeast Asia. By moving their AI training ope

Image on Unsplash On November 24, Trump signed an executive order that instructed the US Department of Energy to create a platform where private AI companies can access federal scientific datasets to train their models. The order is called the Genesis Mission, and it is framed as part of a race for technology dominance against China. The Department of Energy has 60 days to create a list of 20 scientific and technology challenges that the country cares about, including nuclear

Image on Unsplash Meta is in talks to use Google's custom-designed AI chips, known as tensor processing units or TPUs, in its data centers by 2027. They may also rent TPUs from Google's cloud division as early as next year. Till now, Nvidia has dominated the AI chip market. Almost every major tech company building AI infrastructure has relied on Nvidia's GPUs. This year, Meta alone is spending between 70 and 72 billion dollars on AI infrastructure. What changed is that compan

Image on Unsplash Russia’s top AI executive recently compared artificial intelligence development to the nuclear race. Alexander Vedyakhin, who oversees Sberbank’s AI transformation, said only seven countries currently possess their own AI technologies, and nations without sovereign models risk being “left behind”. Vedyakhin emphasized that Russia must develop at least two or three indigenous large language models rather than relying on “retrained foreign models,” especially

Image on Unsplash Trump is preparing an executive order to prevent states from regulating AI, according to a CNN report on November 20. The order would create an “AI Litigation Task Force” in the Justice Department tasked with challenging state-level AI regulations and replacing them with more lenient federal rules. The draft instructs the attorney general to contest laws around deepfakes, algorithmic bias, and AI safety that states have already passed. The administration wan

Image on Unsplash Sundar Pichai, Google’s CEO, just warned that no company would be immune if the AI bubble bursts, including Google itself. In a BBC interview, he acknowledged that the moment feels extraordinary but said it contains many irrational elements. Alphabet’s stock is up 46% this year, powered almost entirely by investor enthusiasm around AI capabilities and deployment. However, the way the AI industry is moving seems too good to be true as of now and soon there wi

Google just rolled out Gemini 3, and it’s scoring higher on reasoning benchmarks than GPT 5 Pro. The model hit a score of 37.4 on the Humanity’s Last Exam benchmark, designed to measure high-level reasoning and general expertise, compared to GPT 5 Pro’s 31.64.  Google released it just weeks after OpenAI launched GPT 5.1 and Anthropic released Sonnet 4.5, continuing a pattern of increasingly rapid model releases across the industry. What’s interesting is that the frontier has

Image on Unsplash Google DeepMind released WeatherNext 2, an AI weather model that can generate forecasts eight times faster than previous systems and produce hundreds of possible weather outcomes in under a minute. Traditional physics-based supercomputer models take hours to run the same scenarios. WeatherNext 2 does it with a new approach called a Functional Generative Network, which generates the full range of possible forecasts in a single pass.  The result is not just sp

Image on Unsplash Jeff Bezos stepped back into an operational CEO role for the first time since leaving Amazon in 2021. He is co-leading Project Prometheus, a newly announced AI startup backed by $6.2 billion in funding, partly from Bezos himself.  This might be a signal that the next frontier in AI is not digital agents or chatbots, but systems that design and engineer physical things including rockets, cars, chips, and factories. Project Prometheus has already assembled nea

Image on Unsplash OpenAI released GPT-5.1 this week, and the update signals a shift in how frontier models are being deployed nowadays. Instead of a simple capability jump on benchmarks, the new version introduces a “reasoning effort knob” that lets developers dial down the amount of computation spent on a given task, with early reports showing two to three times faster responses than GPT-5 on standard workloads at similar accuracy. The good implication is that frontier model

Image on Unsplash India's AI infrastructure is about to expand significantly. Reliance Industries announced on November 13 a second 1 GW AI data centre facility in Andhra Pradesh, operating as a twin to its existing gigawatt-scale centre in Jamnagar. The announcement, signed in the presence of the Andhra Pradesh Chief Minister, marks one of the largest private AI infrastructure commitments in Asia. The facility is designed to host the world's most advanced GPUs, TPUs, and AI

India released its AI Governance Guidelines this week, and the approach is about as far from Europe’s AI Act as you can get (MeitY, 2025). The Ministry of Electronics and Information Technology published the framework on November 5, outlining a hands-off regulatory model where sector-specific regulators handle enforcement rather than a single, binding AI law. The guidelines operate on principles like trust, innovation over restraint, and fairness that set the direction withou

Image on Unsplash For years, the frontier AI playbook was predictable. You’d train a model, license it, and watch others build on top of it. Microsoft’s latest move changes that. The company has stood up a super intelligence unit under Mustafa Suleyman, Microsoft’s AI CEO, to train advanced models in‑house, and reworked its OpenAI deal so it can pursue independent R&D without waiting on a partner’s roadmap. Suleyman plainly stated that Microsoft “needs to be self‑sufficient i

Image on Unsplash India’s AI story has long been told on borrowed infrastructure. But this week, Google moved some of the horsepower home. The company expanded local AI compute on its Hypercomputer stack with Trillium TPUs, so enterprises and public-sector teams can train and serve Gemini models inside India’s borders. This means lower latency and fewer compliance acrobatics.​ The second shoe is even more interesting. IIT Madras’ AI4Bharat  has opened the Indic LLM‑Arena – a

Image on Unsplash Google has added a new “Vibe Coding” experience to its AI Studio platform. The goal: let users go from a descriptive idea to a working AI-powered app in minutes, without needing to manage all the underlying code, APIs or models. Key features include: Prompt-to-app workflow:  Users describe their desired app in natural language (for example, “a photo-editing app that turns selfies into line-drawings”), and AI Studio, powered by Gemini models, wires up the rig

Google Labs has launched Pomelli , an AI-powered marketing tool developed with Google DeepMind, now in public beta across the United States, Canada, Australia, and New Zealand. The platform is designed to help small and medium-sized businesses create professional marketing materials quickly and efficiently. Pomelli analyses a company’s website to build a “Business DNA” , identifying tone, color palette, fonts, and visual style. Based on this data, it generates campaign ideas

Microsoft has rolled out its Copilot Fall Release , marking a major leap toward creating AI that feels more personal, social, and genuinely human-centred. Announced by Mustafa Suleyman, the update introduces 12 new features  that turn Copilot from a simple assistant into a helpful companion that supports creativity, learning, and well-being. The company says the goal is simple: technology should work for people, not the other way around.  The new version of Copilot builds on

NVIDIA and General Atomics, with help from major super-computing centres, have built a high-fidelity digital twin of a fusion reactor, designed to simulate in seconds what used to take weeks. The digital twin uses three large AI surrogate models trained on decades of fusion data and runs on powerful GPU infrastructure (including NVIDIA’s omni-verse, CUDA libraries and leading supercomputers). This isn’t just for show: the system synchronises with the real-world reactor (the D

Image on Unsplash Google Research has rolled out new updates to its Earth AI platform, making its advanced geospatial models and reasoning tools more widely available. The system is built to let users analyse satellite imagery, population dynamics and environment data, and now the company is enabling users to “ask questions” about the world in more natural ways. Key points: A new “Geospatial Reasoning” framework powered by AI (including the Gemini model) allows the platform t

Image Source: Unsplash Google has introduced significant updates to its AI video generation tools, Veo 3.1 and Flow, aiming to provide creators with enhanced control and realism. Key Updates: Veo 3.1 Enhancements:  The latest version offers improved audio quality, better adherence to prompts, and enhanced realism in video outputs. Flow Tool Upgrades:  Flow now includes features like "Ingredients to Video" and "Frames to Video," allowing users to generate videos from multiple