One of the defining characteristics of SRAM-heavy architectures from Groq and its rival Cerebras is that they are very fast when running LLM inferencing workloads, routinely achieving generation rates exceeding 500 and even 1000 tokens a second. The faster Nvidia can generate tokens, the faster code assistants and AI agents can act. But this kind of speed also opens the door to what Huang describes as test-time scaling. The idea is that by letting "reasoning" models generate more "thinking" tokens, they can produce smarter, more accurate results. So, the faster you can generate tokens, the les
A closer look at Nvidia's Groq-powered LPX rack systemsIf you're not a hyperscaler, neocloud, model dev, LPX is probably not for you. The sheer number of LPUs required to serve large open models will likely put Nvidia's LPX platform out of reach for most enterprises. Speaking to press ahead of this week's keynote, Buck said Nvidia is focusing primarily on model builders and service providers that need to serve trillion-plus-parameter models with token rates exceeding 500 to 1,000 a second. Having said that, in a technical blog, Nvidia presented another use case for the LPUs as a speculative decode accelerator, something we suggested the company mi
A closer look at Nvidia's Groq-powered LPX rack systemsNvidia GTC Nvidia slaps $20B Groq tech into massive new LPX racks to speed AI response time GPUzilla's $20B acquihire paves to way to AI agents that halucinate faster than ever GTC Nvidia will use Groq's language processing units LPUs , a technology it paid $20 billion for, to boost the inference p… …
… Thomas Cornely, EVP of Product Management at Nutanix, said in a statement: "Nutanix Agentic AI extends our AHV hypervisor, Flow Virtual Networking, Nutanix Kubernetes Platform, and Nutanix Enterprise AI to deliver a cloud operating model to enterprise AI factories, enabling infrastructure and platf… …
… Tuesday's event was all about Arm’s newly announced AGI CPU products, which will free the company from the shackles of its IP licensing model by enabling the company to sell directly to end customers. Haas has high hopes for agentic AI to accelerate the British chip designer's datacenter business. …
… Last week, Amazon and Cerebras announced a collaboration to pair AWS' Trainium-3 accelerators with the latter's wafer-scale accelerators for many of the same reasons Nvidia built LPX. …
… It said its work would allow orgs to "scale AI initiatives" while "adhering to regional data sovereignty and compliance requirements." Dr Bastian Koller, Managing Director of the High Performance Computing Center at Stuttgart University and lead coordinator of HammerHAI said of the partnership: "Ha… …
… The Board said it is “concerning” that “Meta did not identify on its own initiative the clear engagement abuse signals from the page” and said the company’s mechanisms for flagging fake videos “are neither robust nor comprehensive enough to contend with the scale and velocity of AI-generated conten… …
… While Oracle and OpenAI may not be moving forward with the Abilene expansion, Altman and crew are going to need to find somewhere to perk 5 gigawatts worth of GPUs to claim the $30 billion carrot Nvidia has offered it as part of the $110 billion funding round announced last week in collaboration wi…
… The Department of Commerce is said to be drafting rules that would require chipmakers to obtain government approval before exporting AI chips to any country, effectively giving Washington a stranglehold on AI compute supply. …
… So far, Nvidia has announced the chips will be available in both an eight-way HGX platform or its NVL72 rack system, which as the name suggests, crams 72 Rubin SXM modules into a single system. …
… Announced at Nvidia's GTC 2026 keynote and due out this fall, DLSS 5 takes an application's existing 3D content, colors, and motion and uses AI to add photorealistic lighting. The AI understands what typical human elements such as skin, hair, and clothing should look like. …