The SRO has announced a close partnership with the GTA for the 2027 #Suzuka1000km#, taking place September 5, confirming that GT300 teams will be eligible and run on their tyres with a BoP in place. Further detailes like entry incentives haven't been mentioned. #SUPERGT# #IGTC#

A quick peek at our NVIDIA GB300 NVL72 deployment process. NVIDIA #GB300# NVL72 brings next-generation rack-scale AI performance to the era of reasoning, with ultra-dense compute optimized for large-scale training and high-throughput inference. More to come. Stay tuned👐! #GPU# #AIInfrastructure# #neocloud#

$NBIS signs $1B+ compute agreement with Reflection AI, for GB300 access through 2029. Reflection also signed a multi-billion dollar agreement with $SPCX earlier. Interesting to say the least, seeing Nebius drop -5% off the news today. Also... counterparty to get this done kinda reminds me of OpenAI, where they might not have the funds to actually execute on these LTAs yet compared to $META or $MSFT. But generally positive long term developments, customer diversification was one of the core strengths of Nebius.

SpaceX has almost finished writing V1.0 of an in-house AI training stack in C that exact-maps to 220k GB300s with 800G NICs, making heavy use of pipeline parallelism and getting as close to bare metal as possible. The potential speed improvement vs JAX for large training runs is over an order of magnitude.

‼️ DeepSeek founder Liang Wenfeng says Nvidia is "digging its own grave" in China, according to a transcript of a closed-door investor meeting published this week by Tencent Tech. Liang continues "the Huawei 950 supernode - can fully substitute for NVIDIA's GB200 and GB300 in performance ... The price is definitely higher, but limited. Fifty percent or a hundred percent higher - a hundred percent higher doesn't matter; even two hundred percent higher doesn't matter." He puts DeepSeek's Huawei allocation at roughly 16,000 cards, about 4,000 Nvidia-equivalents by his own four-to-one ratio, against the roughly 200,000 Huawei cards he says a model at today's largest scale would need.

Pretty good chance that Grok 4.6 could surpass Fable 5 at solving complex problems and handling real-world tasks Elon Musk recently said SpaceXAI is closing the loop on solving real-world engineering problems across Tesla, SpaceX, Neuralink and The Boring Company The next Grok release is going to be a very big jump Additionally, SpaceXAI is developing its own C/C++ inference software from the ground up, mapped directly to GB300 hardware Elon said it could double output speed or potentially deliver an even greater improvement And remember, SpaceXAI has built one of the most highly optimized AI infrastructure on Earth That efficiency advantage could eventually reach users through faster responses, more generous usage limits and lower costs SpaceXAI is optimizing every layer of the AI stack for maximum intelligence, speed and efficiency: • Frontier models • Real-world engineering feedback • Highly optimized data centers • Custom training infrastructure • Ground-up inference software • Token and cost efficiency This could bring a huge jump in intelligence, real-world usefulness, agentic performance, speed and efficiency

Production AI demands infrastructure that holds up at scale. NVIDIA and @AWSCloud announce three advances that strengthen every layer of the AI infrastructure stack. 1️⃣ AWS EC2 G7 instances bring NVIDIA Blackwell GPUs to AWS, delivering up to 4.6x AI inference performance over G6 2️⃣ Amazon OpenSearch Serverless NextGen now uses GPU-accelerated vector indexing powered by NVIDIA cuVS as default, enabling vector indexing up to 10x faster at a quarter of the cost 3️⃣ AWS Achieves NVIDIA Exemplar Cloud Status for GB300 Training Performance 🔗