AI Infrastructure news and analysis on Web Pulse: 35805 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
A Bitmain Antminer Z15 Pro mining Zcash is generating an estimated $727.30 in gross revenue per megawatt-hour of electricity, about 3.3 times the estimated industry-average HPC yield of $222.73 per MWh, according to The …
Andreessen Horowitz (a16z) published a slide deck by Anish A. arguing that intelligence is the primitive and applications are the diffusion layer, emphasizing the importance of the application layer in AI. The deck outli…
Arduino has opened pre-orders for the VENTUNO Q board, a dual-brain edge AI platform featuring a Qualcomm Dragonwing IQ8 processor with up to 40 dense TOPS of AI performance and an STM32H5 real-time microcontroller, enab…
OpenAI announced the upcoming release of o3 Mini, a lightweight AI reasoning model, on September 12, 2026. The model delivers near-state-of-the-art performance with 3x faster inference and 70% less energy consumption tha…
OpenAI's new inference chip, Jalapeño, outperforms Nvidia's Blackwell and other competitors in performance per watt across almost all scenarios, according to benchmarks run by SemiAnalysis with OpenAI engineers. The chip…
Raymond James upgraded Advanced Micro Devices (AMD) stock to strong buy from outperform, citing surging demand for CPUs powering agentic AI workloads. Analyst Simon Leopold raised his outlook, noting CPUs are underapprec…
NVIDIA Corp. reports Q2 FY27 earnings on August 26, with CEO Jensen Huang guiding to $91 billion in revenue, plus or minus 2%, after Q1 Data Center revenue hit $75.246 billion, up 92% year over year. Shares trade at $208…
OpenAI unveiled its Jalapeño AI accelerator at the Hot Chips conference on Tuesday, claiming the custom chip, developed with Broadcom, delivers 1.5x to 1.9x higher peak throughput and 1.7x to 3.6x lower latency than Nvid…
A developer found a scale-selection rule for NVFP4 gradient quantization that reduces mean squared error by 14% on real gradient tensors compared to the state-of-the-art MS-EDEN estimator, but training runs at 2.8B param…
OpenAI unveiled its custom AI inference chip, Jalapeño, at the Hot Chips conference, reporting 1.5 to 4 times the performance-per-watt of leading AI chips, including Nvidia's Blackwell, based on SemiAnalysis's InferenceX…
OpenAI announced that its new AI chip, Jalapeño, completes tasks more efficiently and returns responses faster than other AI systems, according to a blog post published on Tuesday. OpenAI hardware vice president Richard …
Apple's M6 Mac mini starts at $899 with 16GB of unified memory, up from the previous M4 model's price, while the Mac Studio with M5 Ultra starts at $5,499, featuring Apple's first four-die processor and 2nm silicon. Both…
Microsoft Research released BitNet b1.58 2B4T, the first open-source native 1-bit LLM at the 2-billion parameter scale, which uses ternary weights (-1, 0, +1) requiring about 1.58 bits per weight, reducing storage to rou…
Netra Kernel, an open-source ahead-of-time GPU kernel compiler for AMD, achieved 78,498.66 output tokens per second mean throughput serving Qwen3.6-35B-A3B-FP8 on eight AMD MI350X GPUs, within 0.32% of the locked baselin…
Lancium announced a strategic collaboration and investment from NVIDIA to accelerate deployment of large-scale accelerated computing infrastructure across its data center portfolio, targeting gigawatt AI factories across…
Broadcom Inc. and Micron Technology Inc. reported blockbuster quarterly results driven by the shift from AI training to inference and agentic workloads, with Broadcom's fiscal Q2 revenue reaching $22.2 billion (up 48% ye…
Cerebras Systems reported Q1 FY27 core revenue of $209.87 million, up 103% year over year, while NVIDIA reported $81.61 billion in quarterly revenue with Data Center revenue of $75.246 billion, up 92%. Cerebras CEO Andre…
ASUS announced the ROG Xbox Ally X20 gaming handheld will launch at $1,299.99, a $300 premium over the standard ROG Xbox Ally X, with pre-orders starting soon on ASUS's online store and Best Buy. The X20 features a 7.4-i…
Apple announced the new Mac Studio on August 25th, featuring the M5 Max and M5 Ultra chips, with prices starting at $2,499 and $5,499, and deliveries beginning September 22nd. The M5 Ultra, built from two M5 Max packages…
Google Cloud has partnered with Subsquid's $SQD Network to contribute blockchain datasets to Google Cloud Web3 Blockchain Analytics, allowing users to access $SQD-validated onchain data through BigQuery. The integration,…