AI news, scored for the stocks it moves.
DatacenterDynamicsInfrastructure
The reality of AI infrastructure has arrived: the power-hungry chips of today, such as the NVIDIA Blackwell, have rendered traditional air cooling obsolete
ServeTheHomeInfrastructure
NVIDIA launched the DGX Spark 64GB at $4,999, while the 128GB model increased to around $6,950. The 64GB version is available from multiple OEMs starting Oct. 23.
Tom's HardwareOther
OpenAI's Jalapeño inference ASIC, co-developed with Broadcom, was discussed in an interview with hardware VP Richard Ho, highlighting its AI-assisted design and efficiency improvements over Nvidia. Synopsys launched its Autopilot platform for AI-driven chip design, aiming for general availability by year-end.
Tom's HardwareInfrastructure
AI agents used 5x more tokens than humans as of August, with 7.3 trillion vs. 1.4 trillion, according to OpenRouter data. Over 85% of agent tokens came from cached prompts, driving demand for high-bandwidth memory.
Tom's HardwareFunding
Amazon announced a $1 billion community program over five years to support data center regions, focusing on education, energy efficiency, and local priorities. The initiative includes free degree access for 300,000 students and energy grants to cut electricity costs by 20-40%.
CNBCInfrastructure
America’s fight over data centers is going global. Communities across Europe and Asia are pushing back over the rising costs of the AI infrastructure boom.
Financial TimesProduct launch
CME Group plans to introduce a new commodity to its roster that will allow the hedging of fluctuations in the price of chip rental
DigiTimesInfrastructure
Biren Technology is preparing its next-generation GPU for mass production, advancing a domestic AI accelerator roadmap built around broader low-precision support and locally available manufacturing.
FintolkInfrastructure
Copper prices hit a record $14,600 per metric ton due to AI demand and supply constraints. Hyperscale data centers, led by Microsoft, Google, and Amazon, require up to 50,000 metric tons of copper each.
The ElecPartnership
LS Cable & System partnered with HARTING to develop AI data center power solutions, signing an MOU in Singapore on September 29. The collaboration combines LS Cable's cables and busducts with HARTING's connectors to supply power to server racks.
CoreWeave BlogProduct launch
CoreWeave hosted a conference with over 5,000 AI pioneers, showcasing new AI platform capabilities. Cognition was the first to run production workloads on NVIDIA Vera Rubin NVL72, with 11,000 concurrent AI environments supported per rack.
The InformationFunding
AI cloud company Lambda said it has secured its first delayed draw term loan at over $1 billion to purchase more than 30,000 Nvidia graphics processing units. The fixed-rate facility carries an interest rate of 6.78%. Privately held Lambda said in an announcement late on Thursday the facility ...
DatacenterDynamicsPartnership
Will see Cerebras' wafer-scale compute coming to the AI cloud platform
The InformationInfrastructure
Debt financing for AI data centers has grown into a distinct category, with at least 20 high-yield bond deals in the past 12 months. Key players like Nvidia, Microsoft, and Google are connected to these projects through leases or credit support.
DatacenterDynamicsInfrastructure
Facility estimated to cost €24 million and open in 2030
DatacenterDynamicsInfrastructure
The conversation around AI infrastructure often centers on one question: Where will all the power come from?
Tom's HardwarePolicy & safety
A California tech CEO was arrested for smuggling $300 million in Nvidia AI servers to China, violating U.S. export controls. Greg Lui, owner of Earthmade Computer Inc., faces up to 20 years in prison for conspiracy, smuggling, and money laundering.
Tom's HardwarePartnership
Amazon and Synopsys signed a multi-year, over-$1 billion deal to collaborate on AI chip design, with Amazon licensing Synopsys' IP and EDA tools. Synopsys will optimize its solutions for Amazon's Trainium and Graviton chips and adopt AWS services.
BloombergFunding
Broadcom’s Wall Street syndicate is starting to gather $60 billion of fresh AI chip financing to benefit Anthropic PBC and other companies, according to people with knowledge of the matter. Carmen Reinicke has more on "Bloomberg Open Interest." (Source: Bloomberg)
DatacenterDynamicsInfrastructure
GPUs could be coming to a parking lot near you
CoreWeave BlogInfrastructure
CoreWeave announced NVIDIA Vera Rubin NVL72 is in limited availability, with Cognition achieving 4.8x throughput increase for SWE-2 inference. The platform features 72 Rubin GPUs and 1,400 TB/s HBM4 memory bandwidth.
CoreWeave BlogProduct launch
CoreWeave opens ARENA Pass to new enterprise teams, offering access to CoreWeave Forge for testing AI workloads. You.com achieved p99 latency below 1.3 seconds with 1,200 concurrent sessions.
CoreWeave BlogInfrastructure
NVIDIA Vera CPU, featuring 88 custom cores, launches on CoreWeave, offering 3x faster agent sandbox startup times than x86 CPUs. Each rack includes 128 Vera CPUs and supports 11,000+ concurrent environments.
CoreWeave BlogInfrastructure
CoreWeave earned a third consecutive Platinum rating in SemiAnalysis' ClusterMAX 3.0 evaluation, the only provider to achieve this across three major assessments. The rating highlights its reliability, with 10,000 GPUs provisioned weekly and advanced systems for straggler detection and node validation.
CoreWeave BlogProduct launch
CoreWeave announced five initiatives to create an integrated AI development stack, including partnerships with NVIDIA and tools like Weights & Biases and marimo. The goal is to enable faster discovery by connecting knowledge across disciplines.
NVIDIA BlogProduct launch
NVIDIA DGX Spark 64GB, available from Acer, ASUS, Dell, Gigabyte, HP, and MSI on Oct. 23, offers 64GB of unified memory for local AI development. It supports up to 100-billion-parameter models and can cluster with another unit for 128GB, boosting performance by 1.7x in tests.
Mistral AI NewsInfrastructure
Mistral AI introduced regional inference endpoints and a priority tier with SLA-backed service levels, enabling customers to process data in Europe or the U.S. and ensuring reliability for mission-critical workloads. The company also expanded support for third-party open models like Z.ai's GLM-5.2 and formed a coalition to secure up to 1 GW of European AI compute capacity by 2030.
Mistral AI NewsPartnership
Mistral and HUMAIN announced a strategic collaboration to advance sovereign AI in Saudi Arabia and the Middle East, focusing on cybersecurity, voice, and Arabic language models. The partnership involves hundreds of millions of Euros and explores HUMAIN’s data center infrastructure for local compute needs.
Tom's HardwareProduct launch
Nvidia launched a 64GB version of the DGX Spark, priced at $4,999, to offer a more affordable option for local AI users. The new model uses the GB10 SoC and retains clustering capabilities via ConnectX 7 NIC.
DatacenterDynamicsInfrastructure
And why there's constraints to overcome in order to meet AI workload demands