Micron Technology

Following Micron's record fiscal Q4 and fiscal year 2026 results, Chairman and CEO Sanjay Mehrotra joined CNBC's Jim Cramer on Squawk on the Street to discuss the growing importance of memory and the strong demand environment driving Micron's momentum. "Memory is key in this era of AI or super intelligence. Micron has a comprehensive portfolio of technology, products, manufacturing capabilities, and we continue to address the fast growing opportunities ahead," Sanjay shared, highlighting Micron's technology leadership, expanding customer commitments and continued investment in the talent powering that success.   Watch the full interview here:

Today, Micron reported our Q4 FY26 financial results. Learn more: https://bit.ly/47sXaf2

Every second matters when a vehicle is making decisions in real time. While a human driver may need about 1.5 seconds to react, an AI-powered car can respond in a fraction of that time, provided the hardware inside the vehicle is fast enough. A car traveling at highway speeds must make critical safety decisions locally. The #AI has to quickly process data from cameras, LiDAR and radar all at the same time to understand its surroundings and respond accordingly. This makes the car's memory the ultimate reflex system. If the memory bottlenecks, the vehicle's reaction time slows down. That's why Micron engineers #automotive memory and storage to move massive amounts of data in milliseconds. On the road, speed matters.

Have you ever used an #AI eraser tool to wipe a photobomber out of an otherwise perfect shot? It happens in the blink of an eye, but behind the screen, your phone is actually crunching an unbelievable amount of data. When you circle someone to remove them, the AI isn't just cutting them out... it’s painting over them. To pull that off, the phone has to instantly scan millions of surrounding pixels, figure out the lighting and textures, and generate a brand-new, high-res background that looks natural. That means reading and writing huge chunks of data all at once. If your phone's memory can't keep up, the AI processor gets starved for data. You end up staring at a loading wheel, the phone gets hot, or the app crashes completely. This is why modern AI smartphones rely on ultra-fast mobile memory and storage to keep the studio running smoothly!

The promise of agentic AI is simple: less time spent managing tasks and more time focused on higher-value work. Behind that promise is a fundamental shift in how AI systems operate. As Micron SVP & GM, Core Data Center Business Unit, Jeremy Werner explains, agentic AI runs on two compute stacks, not one. The GPU reasons, generates and plans, while the CPU orchestrates actions and carries those plans out. Technologies like KV cache are essential to keeping AI context available and inference moving efficiently. But as context sizes grow and agentic workloads drive dramatically higher infrastructure traffic, memory requirements extend beyond reasoning alone. Learn more: https://bit.ly/3TOjj4g

From personalized food recommendations to real-time translations to smarter on-device capabilities, these AI experiences all share one growing challenge: getting the right data to the right place, at the right time. Which memory architecture do you think is best suited for sustained edge #AI inference as workloads become more demanding? A. Discrete memory architecture B. Shared memory systems C. UMA with GPU implemented D. NUMA architecture Drop your answer in the comments below. #EdgeComputing #IntelligenceAccelerated

AI doesn't process text word-by-word. Instead, it breaks language into smaller pieces called tokens. As a rule of thumb, 100 tokens is roughly 75 words. But once text is converted into tokens, how does AI know what comes next? A language model doesn't rely on experience the way humans do. Instead, it navigates an entirely different space: the embedding space. This is where words, concepts and relationships are encoded in ways that help AI recognize patterns and predict what comes next. It's also where the idea of AI "knowledge" becomes much more interesting. Discover how meaning is represented in this space and explore what "knowledge" really means for AI. 👉 https://bit.ly/4rsRgUy

Whether you're tracking a workout, navigating your commute or creating content on the go, AI helps turn everyday moments into smarter experiences. Behind every AI-powered insight is a massive flow of data that must be stored, moved and accessed efficiently. From training AI models in the cloud to delivering real-time intelligence on smartphones and wearables, Micron's memory and storage provide the consistent data foundation that helps AI scale across cloud, data center and edge environments. Because #AI doesn't stop at the screen. It moves with you. https://bit.ly/4hB89Jf #IntelligenceAccelerated

AI is raising expectations for personal computing. Users want intelligent experiences that are responsive, personal, and available wherever they work. Join Nirav Patel, CEO of Framework, and Chris Moore, VP of Marketing at Micron, as they discuss how AI is reshaping the future of PCs, from on-device AI and system design to the memory and storage innovations needed to deliver next-generation experiences at the edge. Watch the full conversation. #AI #AIPC #FutureOfComputing

Have you ever wondered how high-end robot vacuums manage to navigate around pet toys, shoes, and stray charging cables without getting tangled up?   It’s not just a camera; it’s #AI working at lightning speed. But for that AI to actually prevent a collision, it requires incredibly fast memory. Here is how it works:  When a robot vacuum rolls through your living room, its camera is capturing video at roughly 30 to 60 frames every single second. If the vacuum had to send that video feed to a cloud server to ask, "Is that a power cord?", the internet delay (latency) would mean the wheels would run over the cord before the AI could ever say "STOP!"   To prevent this, the AI has to run completely offline, right inside the vacuum itself.  That's where the memory comes in. The vacuum uses ultra-fast memory to catch, process, and analyze those 30+ frames per second quickly. The memory allows the neural processor to identify a hazard in a fraction of a millisecond.    The faster the memory, the faster the vacuum hits the brakes! 🛑