Blurs

Project Nimbus Explained: The Cloud Computing Deal at the Centre of a Global Technology Debate Cloud computing projects rarely make international headlines. Most involve modernising government IT systems, improving cybersecurity or replacing ageing data centres.  Project Nimbus, however, has become one of the most closely scrutinised technology contracts in the world, placing cloud infrastructure, artificial intelligence and corporate ethics at the heart of a global debate. Signed in 2021, Project Nimbus is a US$1.2 billion cloud computing contract awarded to Google and Amazon Web Services (AWS) by the Israeli government.  While the project was originally presented as a major digital transformation initiative, it has since attracted widespread attention due to concerns over the potential military applications of cloud and AI technologies, particularly following the escalation of the Israel–Hamas conflict in October 2023. For the technology industry, Project Nimbus raises diffi...

Google and Meta don’t “magically” know your phone and laptop are yours.  They build that link by stitching together multiple overlapping signals until the probability becomes very high. Here we’ll break down how each company typically does it in practice. 1) The core idea: identity graphs Both Google and Meta Platforms build what’s called an identity graph. Instead of: “This is user X” They store: • devices • browsers • app sessions • login events • IP history • behavioural patterns ..and connect them with confidence scores like: “These devices likely belong to the same person or household” 2) The strongest link: login accounts Google If you’re signed into: • Gmail • YouTube • Chrome • Android …you’ve essentially provided a hard identity anchor. Once logged in: • laptop browser ID • phone app ID • tablet session ..all get tied to the same Google Account automatically. Even if you switch devices, the login connects them instantly. Meta Same idea with: • Facebook • Instagram • WhatsA...

What Are TPUs? Inside Google’s Custom AI Chips Powering the Future As artificial intelligence continues to dominate the tech landscape, one piece of hardware is rapidly gaining attention: the Tensor Processing Unit (TPU).  Developed by Google, TPUs are purpose-built chips designed specifically to accelerate machine learning workloads—and they’re reshaping how AI is trained and deployed at scale. What Is a TPU? A TPU (Tensor Processing Unit) is a custom application-specific integrated circuit (ASIC) built to handle the massive mathematical computations required for AI. Unlike traditional CPUs or GPUs, TPUs are optimized for: • Matrix multiplications (core to neural networks) • Parallel processing at extreme scale • High efficiency for AI training and inference In simple terms, TPUs are laser-focused on AI performance, cutting out unnecessary functions to deliver speed and efficiency. Why Did Google Create TPUs? As AI models grew more complex, relying on general-purpose chips became ...