Tag
General Compute is deploying Cerebras systems alongside NVIDIA GPUs to offer 20x faster AI inference, funded by $400M in debt, to address compute availability for developers.
Elon Musk comments on the future of compute in space, referencing a claim that Google is launching TPUs on a SpaceX rocket to test AI data centers in orbit.
Google's Project Suncatcher explores using space solar energy to power AI data centers, aiming to provide efficient and scalable compute for AI by leveraging orbital solar panels and satellite networks.
AMD has surpassed $1 trillion in market capitalization for the first time, highlighting the expansive potential of the AI compute market to support multiple trillion-dollar chip companies.
Chris McGuire discusses the Huawei Chairman's admission of constraints in China's AI chip production, highlighting long-standing concerns about Huawei's limited capacity for domestic and international AI compute.
Crusoe raised $3.9 billion in a Series F round to build massive data centers and modular AI factories, increasing its valuation to $30.9 billion and expanding AI infrastructure capacity.
Jensen Huang identifies a third major area of AI compute demand: giant data centers built specifically to evaluate and safety-test frontier AI models, separate from training and inference.
Elon Musk discusses a consensus estimate that 15GW of AI compute produced in 2027 cannot be turned on due to infrastructure challenges such as power, transformers, cooling, and networking.
At Hot Chips 2026, Samsung presented their Processing-in-Memory (PIM) technology, which integrates MAC units within LPDDR5X chips to exploit high internal bandwidth for enhanced compute efficiency in AI workloads.
Anthropic has secured a $45 billion deal with Nscale for AI compute power using Nvidia's Vera Rubin chips, continuing its aggressive scaling of infrastructure to compete in the AI industry.
Dylan Patel predicts that Anthropic and OpenAI will control most of the world's AI compute by 2028 due to superior monetization and economies of scale, raising concerns about centralization and potential sovereign debt crises.
Arm announces Apple's launch of new M6 and M5 Ultra chips built on Arm architecture, representing a significant advancement in AI compute and performance.
Apple launches new Mac Studios with M5 Ultra chip, claimed to be the most powerful chip ever with up to 4.3x faster AI compute, starting at $5,499 and launching with macOS 27 Golden Gate.
Apple announces new Mac Mini models with M6 and M5 Pro chips, offering upgraded performance, faster storage, and higher starting prices of $899 and $1,699 respectively.
Apple announced its new M6 chip, the first 2nm chip with enhanced performance and AI compute, along with the M5 Ultra for high-end tasks. The chips will power new Mac Mini and Mac Studio models, available for preorder soon.
Xiaomi has shown an AI Cube Prototype with custom chips and advanced specs, aiming to compete with GB10 by running large AI models locally with high performance and memory bandwidth.
Starcloud has raised $250 million at a $2.3 billion valuation to expand AI compute capacity in orbit, with investors including NVIDIA and Cisco.
The article speculates on the potential impact of concentrating the global compute used for millions of AI prompts onto a single scientific problem, such as curing cancer, and questions whether this could lead to deeper scientific intelligence.
Silicon Data has raised $30 million in Series A funding to develop a reference price for GPU rental and launch compute futures trading on CME, aiming to help Wall Street hedge AI compute costs.
Elon Musk agrees that the space economy will be crucial for scaling AI compute and handling most internet traffic.