Meta’s Post

Super cool !

$100 billion in CAPEX and $7.5 billion annually in OPEX to power 3 million H100/H200 GPUs. "In the next 5-10 years, the cost that you pay for services delivered relatively freely today will go up by an order of magnitude or more." It is worth remembering that cloud servers depreciate at 20 to 25% per annum, while Mark Zuckerberg estimated the amount of computing power required to train Meta’s next large language model would be “almost ten times more” than the previous version, even while conceding that it would be “years” before some of its AI features, made any money “by themselves”. #ai #machinelearning #sustainability #economic #technology

🔥 Google Cloud and NVIDIA just announced a deeper partnership to level up the cloud 🤝 This isn't just a handshake deal - it means serious upgrades for: {Cloud Gamers}: Get ready for buttery-smooth performance, even on the most demanding games! {#AI Enthusiasts}: Unleash the power to build and train next-generation AI models that will blow your mind. The official details are still under wraps, but one thing's for sure, this team-up is a game-changer for anyone who wants to push the boundaries of what's possible in the cloud. #googlecloud #nvidia #cloudgaming #AI #futureoftech

Last-night read: Google's latest "Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters." Woke up today with Meta launched Llama 3.2 — and it's got me thinking… The race is on for edge computing. AI will truly be democratized when LLMs run locally on mobile devices — accessible to everyone, anywhere, without relying on the cloud or expensive hardware. It's no longer just about bigger models — it’s about smarter compute.

This hits my guiding principle - if your going to accelerate, accelerate a lot, not a little. Azure is proud to announce the exclusive launch of the NMads MA35D Virtual Machine (VM) – a groundbreaking addition to our portfolio specifically designed to redefine media processing and video transcoding in the cloud. This Microsoft-exclusive VM harnesses the power of AMD’s Alveo MA35D accelerator to deliver unparalleled media processing capabilities, offering enterprises the scalability, efficiency, and performance they need to push their workloads to the next level. Learn more about the AMD MA35D here: https://lnkd.in/gpAGpbuc https://lnkd.in/ggKspRAa #amd #ma35D #alveo

🚀 This is how Google Cloud is revolutionizing sports broadcasting! FOX Sports is using AI to analyse millions of videos and images instantly, giving fans access to real-time highlights and iconic moments. 🤯 Cricket fans in India, get ready for a revolution in sports viewing! 🏏 With Google Cloud's AI, imagine reliving iconic moments like Sachin's centuries or Dhoni's World Cup six instantly. StratAspire is bringing this cutting-edge tech to India. Stay tuned for more! #StratAspire #GoogleCloudIndia #AI #Cricket #TheNewWayToCloud

Part of Alibaba Cloud's Qwen (Tongyi Qianwen) model series, Qwen-Audio is an LLM that accepts diverse audio files from human speech to natural sounds as input, allowing users to understand audio in new ways. #AudioLLM #TongyiQianwen

🚀 $NVDA is soaring high! Here's why: 💯The semiconductor industry is booming! 💡$NVDA has a strong competitive edge in AI, gaming and cloud 📈 The chart shows a breakout from a long consolidation pattern 🛑 We re-entered on January 8 after an earlier stop-out 💰 We sold some at double our risk, and we're holding the rest for more 🤖 $NVDA is the infrastructure supporting the #AIspring NVIDIA, #semiconductor #AI #cloud

NVIDIA GTC 2024 AI and Cloud partnership rundown At NVIDIA GTC 2024 a plethora of key partnership renewals and expansions were announced, here’s a rundown of the key AI and Cloud partnerships announced on day one. Before you read this, we strongly recommend you check out the key product announcements to gain an understanding of the new offerings from NVIDIA. 👇 https://bit.ly/3VqmQV0 With the new NVIDIA Blackwell GPU platform on offer, Amazon Web Services (AWS) has been quick to snap up plenty for itself and strengthen its longstanding partnership with NVIDIA. Read more 👇 https://bit.ly/3vjw76w #ElectronicSpecifier #Engineering #Nvidia #AI #Cloud Google Cloud Microsoft Oracle NVIDIA

Another day, another exciting listing on BingX. I’m personally Thrilled that $NOS will be live on BingX at 10:00:00 (UTC+0). For those unfamiliar, $NOS is behind Nosana, a decentralized marketplace for cloud computing, and it’s honestly one of the most innovative platforms I’ve seen for AI workloads. It makes accessing powerful GPUs more affordable by allowing users to rent from idle hardware across the globe. What’s even better is that GPU owners—whether gamers, miners, or businesses—can easily earn passive income by becoming Nosana Nodes and putting their unused hardware to work. I love how this project brings a fresh, practical solution for both AI users and hardware owners