MediaTek Dimensity 9600 Pro: 2nm Chip with AI, Gaming, Imaging, Security

MediaTek Launches Dimensity 9600 Pro With 2nm Design and Agentic AI. Built using a 2-nanometre process, the Dimensity 9600 Pro introduces what MediaTek calls a Native AI Architecture, bringing the CPU, GPU, neural processing unit and image signal processor closer together. The company is positioning the chip for a new generation of smartphones where AI agents can remain active in the background, understand context and coordinate tasks across applications. For more MartechAI updates | Visit – www.martechai.com | Read the full story. Link in comments. Follow MartechAI's WhatsApp Channel: https://lnkd.in/dWmjVfCx Follow MartechAI on Instagram: https://lnkd.in/gTGXpBqY #MartechAI #MediaTek #Dimensity9600Pro #AgenticAI #OnDeviceAI #AIHardware

MediaTek Dimensity 9600 Pro Brings Exciting 2nm Power to Phones MediaTek unveils the Dimensity 9600 Pro with a 2nm design, LPDDR6, UFS 5.0, stronger AI, ray tracing and 4K 240fps video support....

Meet the A20 Pro: Apple’s most ambitious chip yet, built from the ground up to power the iPhone with speed, efficiency, and AI in mind. With its 2nm process, reorganised DRAM layout, stronger GPU and neural engine, and custom thermal design, it’s engineered for both today’s tasks and tomorrow’s demands. What sets it apart is full control over the entire stack—hardware, software, cooling, packaging—allowing stable, sustained performance under load. Vertical integration at this level raises the bar for competitors. Keep an eye on how such holistic design shapes future device performance, especially as AI and graphics push chips harder than ever before. #Apple #A20Pro #SoC #AI #Silicon #Smartphone https://lnkd.in/dJjyFpEy

We often talk about the explosive demand for GPUs and AI accelerators. But there is another component quietly becoming a major supply-chain story: MEMORY. AI data centres need enormous quantities of high-bandwidth memory (HBM). As memory manufacturers prioritize HBM and AI-related demand, conventional DRAM and NAND supply is getting tighter. And the impact is spreading far beyond AI servers. Smartphones PCs & laptops Gaming Automotive electronics Industrial electronics For procurement and supply-chain, this creates an interesting shift: AI demand is now influencing the availability and economics of components far outside the AI ecosystem. This is a classic reminder that semiconductor supply chains are deeply interconnected. A capacity decision made for AI today can influence the cost, availability and allocation of memory for thousands of other products tomorrow. AI is becoming the demand driver! Memory could be the next supply-chain battleground. #AI #Semiconductors #Memory #HBM #DRAM #NAND #Procurement #SupplyChain #Electronics #StrategicSourcing

𝐅𝐫𝐨𝐦 𝐭𝐡𝐞 𝐍𝐨𝐤𝐢𝐚 𝟑𝟑𝟏𝟎 𝐭𝐨 𝐭𝐡𝐞 𝐢𝐏𝐡𝐨𝐧𝐞 𝟏𝟕 𝐏𝐫𝐨: 𝐚 𝐭𝐞𝐚𝐫𝐝𝐨𝐰𝐧 𝐨𝐟 𝐡𝐨𝐰 𝐟𝐚𝐫 𝐬𝐞𝐦𝐢𝐜𝐨𝐧𝐝𝐮𝐜𝐭𝐨𝐫𝐬 𝐡𝐚𝐯𝐞 𝐜𝐨𝐦𝐞. Source: By Andrew Chan Yik Hong Credit to Gidion V. Simbo and Behind the Chip Substack for the visual. 📱 #Nokia 3310 Built around ~180nm technology, with relatively simple silicon designed for one clear purpose: calls and messages. 📱 #iPhone 17 Pro Built on a 3nm-class process, with tens of billions of #transistors packed into an incredibly small footprint. Today’s #smartphone silicon doesn’t just process information. It enables computational photography, #AI inference, advanced graphics, connectivity across multiple networks, security, and increasingly sophisticated on-device intelligence. The difference between these two devices is more than 20+ years of product evolution. It is a story of semiconductor scaling, #transistor density, advanced packaging, #architecture, and relentless #innovation. From making a phone call to running #AI models in your pocket, the smartphone is perhaps one of the clearest demonstrations of what #semiconductor technology makes possible. And the pace of innovation shows no sign of slowing down. #Semiconductors #Technology #AI #iPhone17Pro #Nokia3310 #ChipDesign #AdvancedPackaging #MooresLaw #Innovation #ArtificialIntelligence

Apple’s release of the M6 chip gives developers and hardware watchers their best look yet at what to expect from the forthcoming A20 Pro in the iPhone 18 series. Key takeaways include the move to TSMC’s 2nm process node, a dual 16-core Neural Engine delivering up to 2× compute, refined GPU architecture with enhanced geometry rates and ray tracing, and modest memory bandwidth improvements via LPDDR5X and WMCM packaging. For performance and efficiency, the A20 Pro looks set to offer up to 18% faster compute and around 30% better power efficiency compared to its predecessor. Yet despite these upgrades, certain features—like AV1 encoding—are probably off the table this cycle, while transformative jumps beyond graphics and neural compute are unlikely. Why this matters: it signals Apple is doubling down on architectural improvements and power efficiency rather than chasing sheer spec leaps. What to watch: how these advances stack up in real-world AI tasks, whether lower-tier models get similar neural benefits, and whether media support catches up in future silicon roadmaps. #Apple #ChipNews #M6 #A20 #iPhone18Pro #NeuralEngine #2nm https://lnkd.in/d-qxKzst

As Southeast Asia ramps up advanced packaging, OSAT production lines face a high-stakes shift: How do you protect a $30,000+ AI package when packaging sizes surge from 15×15 mm to 150×150 mm? With multi-die architectures and HBM stacks, a minor assembly defect means massive scrap loss. Traditional rework tools simply cannot manage the thermal warpage and precision needed for ultra-large footprints. At iST, we help OSATs across Asia solve this with: Precision large-scale BGA rework & de-soldering Thermal warpage control & reballing IPC-9701 reliability validation Protecting yield on high-value lines is our mission. Let's connect right here on Linkedin to discuss! Full engineering insights in the comments. #AdvancedPackaging #OSAT #SoutheastAsiaSemiconductor #BGARework #YieldOptimization #SEMICONTaiwan2026 #iST

【⚠️ 𝐏𝐫𝐞-𝐒𝐞𝐦𝐢𝐜𝐨𝐧 𝐓𝐚𝐢𝐰𝐚𝐧 𝐒𝐩𝐨𝐭𝐥𝐢𝐠𝐡𝐭! 𝐒𝐜𝐫𝐚𝐩 𝐂𝐫𝐢𝐬𝐢𝐬 𝐟𝐨𝐫 𝐡𝐢𝐠𝐡-𝐯𝐚𝐥𝐮𝐞 𝐀𝐈 𝐂𝐡𝐢𝐩𝐬】From a "tiny fingernail" to a "giant lunchbox"—what should we do when massive AI chips misalign and fail?! 💡 With Semicon Taiwan just around the corner, this year's hottest topics are undoubtedly 2.5D/3D advanced packaging and the limits of AI computing power! Back in the day, consumer electronics chips (like smartphone processors) were only around 15X15 mm, roughly the size of a fingernail. But with the boom of Generative AI, to pack in more computing cores and HBM stacks, flagship packages from Nvidia, Amd, and Intel easily exceed 90X90 mm and even reach 150X150 mm—expanding into the footprint of a square lunchbox, representing a massive 100x increase in surface area! Beyond their immense size, a single chip can easily cost upwards of $30,000+! However, engineers face a severe dilemma: "Standard pick-and-place equipment simply can't handle chips this large. If a placement error occurs or testing fails, how can we de-solder and rework these high-value chips without causing any damage?" #AdvancedPackaging #CoWoS #EMIB #BGARework #LargeScaleChips #ReliabilityVerification #iST #semicontw2026

AI PCs are moving toward a new stage, with more AI processing taking place directly on personal devices. Reuters reports that NVIDIA’s new RTX Spark platform is set to power a new generation of Windows PCs launching this October. For WOWPC, this shift is worth watching closely. As a PC retailer, we see AI capabilities increasingly becoming part of how customers evaluate and choose their devices—not simply an additional feature. We will continue monitoring how on-device AI shapes PC products and consumer expectations. Source: Reuters, September 3, 2026 https://lnkd.in/ettHRwTM #WOWPC #AIPC #PCIndustry #Technology #ComputerHardware

A new VisionCamera demonstration combines depth estimation from DINOv2 with Apple’s Vision hand-detection framework to create depth-aware light injection in real time. The system processes 448×336 frames in less than 25 milliseconds, then uses zero-copy rendering through react-native-webgpu to apply lighting and shadows based on the scene. The result is an interactive effect where a user can move an invisible light source around the camera view and see the lighting respond to the environment. The combination is notable because it brings multiple computer vision tasks into a single real-time mobile graphics pipeline. Depth estimation provides spatial information, while hand detection allows the system to understand where the user is interacting. For developers, this points toward more responsive augmented reality and spatial computing experiences that can run directly on mobile hardware. The underlying technologies also show how computer vision, GPU rendering and cross-platform development frameworks are increasingly converging into a single application stack. As mobile GPUs and on-device AI continue to improve, real-time effects that once required more powerful hardware could become increasingly practical on smartphones. What are your thoughts on this? 🤔💬 Credit: @mrousavy | X Curious about where AI is headed? 🤖 Follow Future AI to keep your edge over the competition #AI #ComputerVision #AugmentedReality #ReactNative #WebGPU