MAXVY Technologies Pvt Ltd’s Post

The technique of stacking multiple chips closely together is crucial to improving the performance of semiconductors.Great article with visualization of packages(HBM,CoWoS,InFo)

Ever wondered how tiny sand turns into the powerful chips inside our gadgets? 🤔 Join me, in this simple breakdown of how semiconductor chips are made! In this video, I'll walk you through the step-by-step process of turning sand into the tiny marvels that power our phones, computers, and more. From melting sand into silicon to crafting intricate circuits, you'll learn the basics of semiconductor manufacturing, also known as VLSI technology Whether you're a tech enthusiast or just curious about how things work, this video will make it easy to understand how semiconductor chips are created. Don't forget to like 👍, share with your friends, and subscribe to Science with Sage for more fun science stuff! #VLSI #Semiconductor #Chip #Manufacturing https://lnkd.in/gvA3pgBU

Outpacing #Samsung or the End of Race? #Kioxia Aims 1000-layer #NAND by 2027 🤩🤯💡 After ending production cuts amidst a recovery in the #memory industry, KIOXIA Group disclosed its plans on the 3D NAND roadmap. According to reports from PC Watch and Blocks & Files, Kioxia stated that achieving a 1,000-layer level by 2027 would be possible. According to the reports, the number of 3D NAND layers has generally increased from 24 in 2014 to 238 in 2022, representing a tenfold rise over eight years. #Kioxia stated that achieving a 1,000-layer level by 2027 would be possible at a rate of increase of 1.33 times per year. The #Japanese #memory #chipmaker seems to be more ambitious than Samsung regarding the battle of layers. In May, Samsung revealed its target to release advanced NAND #chips with over 1000 layers by 2030. According to Wccftech, the #SouthKorean memory giant plans to apply new #ferroelectric materials on the manufacturing of NAND to achieve this goal. Thanks again to TrendForce Corporation for the full article with more background and insights via the link below 💡🙏👇 https://lnkd.in/eiVgEP66 #semiconductorindustry #semiconductormanufacturing #semiconductor #technology #tech #innovation #mobile #datacenter #chip #ic #icdesign

#Micron Technology recently announced that its full suite of automotive-grade #memory and storage solutions has been validated on #Qualcomm Technologies Inc.'s #SnapdragonDigitalChassis™, a comprehensive cloud-connected platform designed to power data-rich, #intelligentautomotive services. Micron's #lowpower #LPDDR5X memory, #UniversalFlashStorage (#UFS) 3.1, Xccela™ #flashmemory and quad #serialperipheralinterface (#SPI) NOR flash have been pre-integrated for the latest generation of Snapdragon #automotivesolutions and #modules, including the Snapdragon Cockpit Platform, Snapdragon Ride™ Platform and Snapdragon Ride™ Flex System-on-Chip (#SoC), to meet the growing #AI workload demands of today and tomorrow. This collaboration between Micron and Qualcomm will help the #automotiveecosystem create the next generation of #intelligentvehicles powered by advanced #AItechnologies. #semiconductorelectronic #semiconductor #electronics #chip #lusengz #automotiveelectronics #electricvehicle #artificialintelligence

Way back when I was working at Intel we did a small skunkworks project (Wildcard) which was a system level module combining multiple die, passive devices and discrete logic onto a single FR4 substrate. It was a complete PC on a module - sans DDR memory. Back then we were anticipating heterogenous multiple die-on-module would be an important tool for forming these higher level systems. At the time we realized optimizing multiple system functions would always require different process nodes and processes and technologies. Rather than integrating flash, programmable logic, large logic functions (e.g. LAN) and CPUs into a single large, expensive die, we could keep die smaller and integrate functions at the last level packaging stage. We referred to this as more-than-Moore's scaling. It is interesting to see, 30+ years later, MCMs (Now called chiplets) are all the rage. You know it's hit mainstream when you read about the technology in the NY Times and Wall Street Journal. This IMEC presentation gives a good overview as to how more-than-Moore scaling addresses size, power and speed problems. https://lnkd.in/gB3r8Qyt The big challenge to sort out is yield -- driven by a robust known-good-die test process at temp (e.g. BIST and Hot chuck wafer test.) Our modest POC at Intel surprised us in that the system level yields we saw were actually greater than the yields we were getting with traditionally packaged parts

I told you we at Intel are on a mission. A mission to regain process technology leadership in our semiconductor industry. 2024 will be the year when we will ship our most advanced chips based on our Arrow Lake processor (Intel 20A) that will feature two revolutionary features: RibbonFET (gate-all-around) and PowerVia (backside power delivery). Read more about it here: #iamintel

It is a risk that can put Intel ahead of competitors. RibbonFET will provide better scaling and up to 15% improvement in power efficiency. RibboFET will appear in the new Intel 20A (angstrom) technology. A more significant game-changer will be the PowerVia, a unique way to deliver power from the backside of the wafer. 

https://lnkd.in/ggt3GaUF Strong demand for HBM memory will see the market more than double to US$4.976 billion in by 2025 from $2.005 billion in 2023, with most of that going to the Big 3 memory makers, SK Hynix, Samsung and Micron. > In Taiwan, several firms are offering HBM chip packaging, IP, or HBM-like alternatives, including Winbond’s CUBE, Etron’s VHM (Very High Bandwidth) and AP Memory’s MemoraiLink. #skhynix #micron #samsung #dram #artificialintelligence #llms #generativeai #nvidia #semiconductorindustry #semiconductor #technology #innovation #business  credit: ctee.com.tw

China’s top memory chip maker, Yangtze Memory Technologies Corporation (YMTC), is seeing a surge in demand for its flash memory chips, which have become a top choice for local clients, particularly those working in government and military-related projects, as Beijing strives to cut reliance on foreign semiconductor products. Demand for locally manufactured components, from graphics processing units to flash memory chips, has been surging, as China embarks on an ambitious drive to expand computing infrastructure to support generative artificial intelligence development. Leading industry players from YMTC to Huawei Technologies have become some of the biggest beneficiaries of this national endeavour. YMTC’s growing business momentum defied expectations that it might have to quit the 3D NAND flash memory market after being added to a US trade blacklist in late 2022.

https://lnkd.in/gmGHFm9z Samsung to use chip making tech favoured by SK Hynix as AI chip race heats up. > Samsung has started buying equipment & materials to manufacture HBM memory chips using a technique championed by rival SK Hynix called MR-MUF (Mass Reflow Molded Underfill). > Samsung’s HBM3 yields are 10-20%, compared to SK Hynix’s 60-70%. The move came after Micron announced an HBM3E deal with Nvidia, leaving Samsung the lone memory giant with no Nvidia deal. SK Hynix is Nvidia’s sole HBM3 supplier. #samsungelectronics #skhynix #artificialintelligence #dram #Semiconductorindustry #electronics #semiconductor #supplychain #engineering #manufacturing #technology #computerchips #business #innovation #semiconductors #chips #chipmaker #foundry credit: reuters, pulse news korea