Tech News by InClicks

▲ Pentagon cuts Anthropic, picks Google Gemini The Pentagon AI supply-chain decision is the formal process that determines which AI vendors are allowed inside federal systems. On April 28, 2026, the Pentagon confirmed it had blacklisted Anthropic and chosen Google Gemini for classified networks. For daily chatbot users, this Pentagon AI decision now ripples into how each major model handles sensitive policy questions. The Single Clause That Cost Anthropic the Contract Anthropic, the maker of the Claude chatbot, bans use of its AI for US civilian surveillance and lethal autonomous weapons under its acceptable-use policy. The Pentagon asked Anthropic to remove those carve-outs. Anthropic refused. Roughly two months ago , the DoD then designated Anthropic as a supply-chain risk - a label typically used for foreign-adversary-controlled entities - effectively blocking federal procurement. ▲ Gemini cleared for classified networks How Google Gemini Walked Into the Open Seat On...

▲ Cambridge brain-inspired chip unveiled A brain-inspired chip is a semiconductor designed to mimic how the human brain stores and processes information at the same physical spot. Researchers at the University of Cambridge just demonstrated a working version that could cut AI energy use by up to 70% . For anyone using ChatGPT, Claude, or Gemini, this matters because data center power costs flow straight into the price you pay every month. The Office Analogy: Why GPUs Burn So Much Power Today's AI chips work like a clumsy office. The library (HBM, the memory) sits in one room while the office (the GPU, where calculations happen) sits in another. Every single computation requires hauling books back and forth between the two rooms. Engineers call this the von Neumann bottleneck , and it's where most of the energy goes. Cambridge's new chip flips the model. Storage and compute live on the same desk - exactly how the brain's synapses work. No shuttling, no waste. T...

▲ JAL humanoid robot pilot at Haneda The JAL Haneda humanoid robot pilot is Japan's first airline-led trial of human-shaped robots performing actual ground operations at a major international airport. Japan Airlines (JAL), through its subsidiary JAL Grand Service, is teaming up with GMO AI&Robotics, the robotics arm of GMO Internet Group. The pilot starts in May 2026 and runs for roughly 3 years in phases. For travelers and workers in the US, this is more than a Japan story. It is a preview of how labor-stretched aviation systems may turn to humanoid robots within the next 1-2 years. Why Japan Chose Humanoid Robots Now Japan's working-age population shrinks every year, and inbound tourism keeps hitting record highs. The result is a widening gap between airport workloads and available staff. JAL says fixed conveyors and arms cannot adapt to the chaotic, variable nature of ramp work. Human-shaped robots, by contrast, slot into existing facilities without major re...

▲ Google's $40B Anthropic investment Google's Anthropic investment is Alphabet's commitment to fund the AI startup with up to $40 billion in staged tranches. It's the largest single investment in an AI startup ever, and it directly affects the price and quality of AI services like ChatGPT and Claude that you may already use. Bloomberg, CNBC, TechCrunch, and Axios all reported the deal on April 24, 2026. How the Anthropic deal is structured The deal is staged. Google puts in $10 billion in cash upfront, valuing Anthropic at roughly $350 billion . The remaining $30 billion ships in tranches as Anthropic hits performance milestones. The milestone structure caps Google's risk while pressuring Anthropic to keep growing fast. Google already held about a 14% stake in Anthropic and had previously invested over $3 billion. Once this deal fully clears, Google becomes Anthropic's largest single investor. ▲ $10B upfront plus $30B in tranches Why Google bet on ...

▲ NVIDIA $5T market cap, 14x in 4 years NVIDIA's $5 trillion market cap is the total value of all NVIDIA shares trading above $5,000,000,000,000 in aggregate market capitalization. According to CNBC, NVIDIA closed at $208.27 on April 24, 2026 , jumping 4.3% in a single session and pushing the chipmaker above a $5 trillion market cap for the first time ever. If you hold an S&P 500 index fund or a target-date 401(k), Friday's close meaningfully moved your retirement balance because of one stock. NVIDIA's $5 Trillion Market Cap Milestone The market cap level is striking, but the speed is the real story. NVIDIA is now worth roughly $1 trillion more than the second most valuable US company, Alphabet (Google's parent), and the gap to Microsoft, Apple, and Amazon is even wider. Friday's close was also NVIDIA's first record high since October 2025, reclaiming the leadership it briefly lost during the late-2025 pullback. The immediate catalyst was a posi...

▲ TSMC A13 2029 mass production roadmap The TSMC A13 process is the world's #1 foundry's next-generation 1.3nm-class chip manufacturing technology, targeted for mass production in 2029 . It was unveiled together with A12 at TSMC's annual North America Technology Symposium in Santa Clara, California on April 22, 2026. The announcement is widely interpreted as a direct counter to Intel's 18A node, which begins mass production in late 2025. Why TSMC Showed Its 2029 Cards Now TSMC holds annual technology symposiums in the US and Taiwan to disclose its next-generation process roadmaps. However, unveiling A13 and A12 simultaneously at its 2026 North America Symposium was unusual. Companies typically introduce only the next single node. The reason is Intel. Intel begins 18A mass production in late 2025 , marking its first credible attempt in roughly a decade to catch TSMC at the leading edge. Intel made massive investments in next-generation High-NA EUV (high-numeric...

▲ GPT-5.5 auto-upgrade hook card GPT-5.5 is OpenAI's newest general-purpose LLM (large language model - the AI engine that reads and writes text), released on April 23, 2026. If you pay for ChatGPT, your default model is quietly rotating to GPT-5.5 right now. What makes GPT-5.5 unusual is the cadence: it arrived just 6 weeks after GPT-5.4. Why OpenAI Shipped GPT-5.5 in Just 6 Weeks GPT-5.5 reads less like a replacement and more like a fast-follow of GPT-5.4. OpenAI president Greg Brockman called it "a new class of intelligence" in the press briefing, emphasizing the same latency at higher intelligence per token. The timing is not accidental either. Anthropic's Claude has been eating into OpenAI's enterprise share in the public narrative, and Fortune reports that BNY Mellon's CIO Leigh-Ann Russell is benchmarking GPT-5.5 alongside Anthropic's latest. When large customers can A/B test frontier labs in real time, shipping faster is the cheapest way...

▲ Google TPU 8 launch cracks Nvidia monopoly era Google TPU 8 is Google's 8th-generation custom AI accelerator, unveiled on April 22, 2026, at Google Cloud Next 2026, and split into two variants: the training-focused TPU 8t and the inference-focused TPU 8i. The real significance of the launch is not the spec sheet but the customer list behind it. For anyone using AI services every day, this announcement signals that ChatGPT, Claude, and Gemini could soon become cheaper and faster . Why Nvidia Owned 90% of the AI Chip Market For years, roughly 90% of AI accelerator spending has flowed to Nvidia, according to major US tech outlets. GPUs (Graphics Processing Units, chips built for massive parallel computation) proved perfect for training and running AI models, and Nvidia's CUDA software stack locked developers in. The result: OpenAI, Meta, and Anthropic spent years queuing for scarce Nvidia hardware at premium prices, and the appetite for a credible alternative kept growi...

▲ Anthropic Mythos Preview enters UK banks Anthropic Mythos Preview is a new premium-tier AI model that Anthropic released on April 7, 2026 and deemed too powerful to make broadly available to the public. If you use a bank app, this matters directly: starting next week, Mythos begins operating inside UK financial institutions, changing how your account is defended before any of your personal habits change. What Is Anthropic Mythos Preview Mythos Preview sits above Claude Opus, Sonnet, and Haiku as a new premium tier focused on cybersecurity. The model autonomously discovered thousands of zero-day vulnerabilities (flaws not yet publicly known) across Windows, macOS, Chrome, and Safari. On benchmarks, Mythos scores 93.9% on SWE-bench Verified (vs Opus 4.6's 80.8% ) and 97.6% on USAMO 2026 (vs Opus 4.6's 42.3% ). Most strikingly, the model reads code, tests it, and produces functional exploit code end-to-end without human prompts - a level of autonomy that prompted Anth...

▲ NVIDIA and Siemens deployed a humanoid robot inside a real factory Physical AI (artificial intelligence that leaves the screen and performs real physical labor through a robotic body) is the quiet shift reshaping industrial work. On April 16, 2026, Siemens, NVIDIA, and UK startup Humanoid officially announced that this technology is no longer a future concept. For anyone working in logistics or repetitive factory tasks, the change happening right now will directly touch your job. How the Siemens-NVIDIA humanoid alliance started This announcement is the first real-world result of the "AI Factory" strategic partnership that Siemens and NVIDIA signed at CES 2026 in January. Siemens brought its digital twin technology (a virtual replica of a real factory used for testing and monitoring) and decades of factory operations expertise. NVIDIA contributed its AI platform, which acts as the robot's brain. UK startup Humanoid completed the three-way alliance by providing th...

▲ Claude Opus 4.7 official launch, April 16, 2026 Claude Opus 4.7 is Anthropic's newest flagship AI model, officially released on April 16, 2026, with sharper coding, stronger reasoning, and the first real high-resolution vision support in the Claude family. For developers and teams running AI workloads, the upgrade changes both what the model can do and how much it costs to run at scale. Why Claude Opus 4.7 matters right now The release arrives roughly five months after Opus 4.6. Through March, OpenAI's GPT-5.4 held the lead on several coding benchmarks. Anthropic's April 16 shipment of Claude Opus 4.7 flips that ordering: the company reclaims the top slot among generally available models. An internal model called Mythos remains in limited preview and outperforms 4.7, but for builders using public APIs, Opus 4.7 is the new ceiling. ▲ SWE-bench 87.6% and GPQA 94.2% highlights How big is the jump over Opus 4.6? SWE-bench Verified moved from 80.8% to 87.6% , a +...

▲ OpenAI GPT-5.4-Cyber cybersecurity AI model launch GPT-5.4-Cyber is a specialized AI model built by OpenAI exclusively for defensive cybersecurity, launched on April 14, 2026. If you use any app, website, or digital service, this matters to you - AI is now actively protecting the software you rely on every day. The model is fine-tuned from GPT-5.4 with enhanced capabilities for security professionals, marking a major shift in how cyber threats are detected and neutralized. Why Did OpenAI Build a Cybersecurity AI? Cyberattacks are becoming more sophisticated every year, and attackers are increasingly using AI to automate and scale their operations. Meanwhile, the cybersecurity industry faces a chronic talent shortage - there simply aren't enough skilled defenders. OpenAI's approach is not to limit AI capabilities, but to gate access through identity verification. Rather than making the model weaker, they ensure only verified defenders can use its full power. This philo...