- 资讯公开站Digitimes
AI data center expansion is hitting new limits on power, land, and public consent
The global race to build data centers for artificial intelligence is entering a more difficult phase. Demand for computing capacity remains strong, but new projects increasingly face limits on land, electricity, water, a…
- 论文公开站arXiv
Agent in a Bottle: Can LLM Agents Turn Their Capabilities Into Cheap, Scalable Artifacts?
Large language models (LLMs) can solve many narrow tasks, but querying them separately for millions of related instances can be prohibitively expensive. Can LLM agents autonomously create cheaper solutions for such workl…
- 资讯公开站CNX Software
NuMaker-IoT-MA35D0-A2 (Chili Pro) ultra-compact evaluation board features MA35D0K Cortex-A35/M4 MPU
Nuvoton has added MA35D05KH67C, MA35D05KI67C, and MA35D05KJ67C, three new LQFP128 MPUs, to the NuMicro MA35D0 series, along with the NuMaker-IoT-MA35D0-A2 (Chili Pro) evaluation board. Designed for industrial AI, smart g…
- 资讯公开站Digitimes
Hua Yang, Zeiss launch COSMA-M for chip metrology
Hua Yang Precision Machinery has teamed up with Zeiss to launch COSMA-M, a high-precision measurement system aimed at the semiconductor metrology market as demand surges from AI, high-performance computing (HPC), and adv…
- 资讯公开站Digitimes
Beyond AI: Starcloud tests whether orbital compute can stay busy enough to pay
The excitement around massive orbital infrastructure mainly stems from its tremendous potential in offering compute nodes powered by near-continuous solar energy beyond the spatial limits of terrestrial infrastructure.
- 资讯公开站Semiconductor Engineering
Monolithic 3D Memristor-TFT Stack For Programmable Neuromorphic Computing (SNU, Yonsei)
Researchers at Seoul National University and Yonsei University published a technical paper titled “Monolithic three-dimensionally integrated memristor-thin-film transistor for electrically programmable multimode reservoi…
- 资讯公开站Semiconductor Engineering
Programmable Memory Controller Eases Adoption of New DRAM Techniques (ETH Zürich, CISPA, NYU)
Researchers at ETH Zürich, CISPA, and New York University published a technical paper titled “Terracotta: Enabling the Adoption of New DRAM Techniques via a Flexible DRAM Interface and Memory Controller.” Abstract Excerp…
- 资讯公开站EE Times
The AI Boom Has a Gigawatt Accounting Problem
- 资讯公开站Digitimes
AMD to expand Taiwan supply chain investment as AI demand drives capacity needs
AMD CEO Lisa Su said the company plans to further increase investment in Taiwan's semiconductor supply chain as demand for CPUs, GPUs, and AI computing continues to outpace available capacity.
- 资讯公开站Digitimes
China's 15th five-year plan elevates PCB role in AI supply chains
China's Ministry of Industry and Information Technology and the National Development and Reform Commission have released the 15th Five-Year Plan for the development of the electronic information manufacturing industry, a…
- 资讯公开站Digitimes
Taiwan bets on SiPh, CPO to tackle AI data center power bottlenecks
Taiwan's National Science and Technology Council (NSTC) has launched two research programs targeting silicon photonics (SiPh) and next-generation AI data center power systems, seeking to address energy and performance bo…
- 论文公开站arXiv
Paradee: Distilling Kokoro-82M into an 8M-Parameter Single-Voice Text-to-Speech Model
We distill Kokoro-82M, a widely used open text-to-speech model with 54 voices, into Paradee, an 8.07M-parameter model that speaks one of them. Paradee keeps Kokoro's architecture with much narrower layers, and each of it…
- 论文公开站arXiv
TasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts
We introduce TasteVal, a benchmark to evaluate the experimental research taste of frontier models. We define research taste as the ability to pick interesting problems to solve, design experiments, and interpret experime…
- 论文公开站arXiv
MemPilot: Orchestrating On-Demand Multimodal Memory Curation for LLM Agents
Memory has become integral to the LLM agent ecosystem, supporting information retention and reuse across interactions. However, most existing agent memory systems construct memory in a query-agnostic manner, which can in…
- 论文公开站arXiv
Towards Looped Models Done Right, Part II: Rethinking at Fixed Points
Every recurrence of a looped language model adds cost in training, decoding, prefill, and reinforcement learning (RL). The closer recurrent states get to fixed points, the less the path to them matters. This enables trun…
- 论文公开站arXiv
One Figure, Every Canvas: Editable Flowchart Relayout via Agentic Pipeline
Pipeline figures in ML papers must be repurposed across many canvases, including paper columns, 16:9 slides, portrait posters, 1:1 social teasers, 9:16 phone previews. Each format imposes a different aspect ratio on the …
- 资讯公开站Digitimes
How RobCo and the robotics unicorn wave are redefining industrial automation
Physical AI is shifting from a hardware niche into the next foundational computing platform, driving record investor capital into embodied intelligence systems.
- 资讯公开站Digitimes
3.2T optical upgrade triggers PCB redesign, shrinking supplier pool to 3
As AI computing clusters continue to expand, performance requirements for scale-out interconnects between racks are rising rapidly. This is pushing traditional pluggable optical transceivers from 800G to 1.6T and acceler…
- 资讯公开站Digitimes
AM Intelligence expands India AI infrastructure push with 20,000 more Nvidia Rubin GPUs
AM Intelligence (AMI) is widening its AI infrastructure push in India and Malaysia, adding to a regional buildout that could matter well beyond Asia. The latest orders expand access to frontier computing for cloud provid…
- 资讯公开站Digitimes
Behind the US$4.2 trillion revenue gap in AI infrastructure
AI data center plans and computing demand are expanding far faster than physical infrastructure can be built, creating a widening gap between planned and actually deployed compute. A review tracking construction progress…
- 资讯公开站Digitimes
Can satellites share power?— Star Catcher puts an orbital grid to the test
As in-orbit AI computing, Earth observation, and defense missions expand rapidly, satellites are demanding more power, turning energy supply into an emerging challenge for the growing space economy.
- 资讯公开站Digitimes
The death of the software seat: cheaper AI tokens turn enterprise computing into metered labor
The ferocious push toward artificial intelligence (AI) is gathering momentum, built upon increasing industry sentiment regarding demand for higher AI inference and compute-intensive models.
- 资讯公开站Digitimes
Marvell sees custom chips as a must for CSPs
Nvidia CEO Jensen Huang has said Marvell's optical technology could help make it the next trillion-dollar company, and Marvell took a step toward that goal on September 17 by expanding its partnership with GlobalFoundrie…
- 资讯公开站IEEE Spectrum
Smart Car Researcher Wants to Eliminate Stoplights
Christos Cassandras has spent more than 40 years studying how computers and machines make decisions. Lately, his work has focused on a problem nearly every driver knows well: sitting at a red light with no cross traffic …
- 资讯公开站Semiconductor Engineering
Row-Parallel DRAM Computing Cuts Data-Reorganization Overhead (Syracuse, FAU, TU Dresden)
Researchers at Syracuse University, Friedrich-Alexander-Universität Erlangen-Nürnberg, and TU Dresden published a technical paper titled “RAPID: Row-Parallel Arithmetic Processing in DRAM.” Abstract Excerpt: “Processing-…
