AI

Towards the end of May, Huawei is reported by the Wall Street Journal to be intending to sample its Ascend 910D AI processor which is aimed at competing with Nvidia’s 2022-vintage Hopper processor H100. Mass production of the 910D is reported to be scheduled for Q1 2026 on SMIC’s 7nm process. It is said that the 910D can be competitive ...

Addressing the growing use of AI and high-performance computing, JEDEC has published a new high-bandwidth memory (HBM) DRAM standard: HBM4. It aims to improve the performance, efficiency, and scalability of the memory required to power the next generation of AI, HPC (high-performance computing), and graphics workloads. For example, the next generation AI training and inference systems. HBM4 With transfer speeds ...

Nvidia is working with its manufacturing partners to design and build factories that, for the first time, will produce Nvidia AI supercomputers entirely in the U.S. Together with leading manufacturing partners, the company has commissioned more than a million square feet of manufacturing space to build and test NVIDIA Blackwell chips in Arizona and AI supercomputers in Texas. Nvidia Blackwell ...

Indian start-up Ziroh Labs of Bengaluru has come up with a way to run AI programmes on relatively inexpensive CPUs instead of ultra-expensive GPUs. Called Kompact AI, the technology is aimed at bringing AI training and inference to SMEs which cannot afford GPUs. “Anything that can be computed on a GPU can also be calculated on a CPU; what really ...

STMicroelectronics has announced a series of AI-enabled and security-enabled industrial and IoT microcontrollers: STM32MP235. The parts have dual Arm Cortex-A35 cores (1.3 or 1.5GHz depending on version), a 400MHz Cortex-M33 and an 0.6Top/s neural network accelerator. On board is a 400MHz VeriSilicon GC8000UL 3D GPU capable of up to 66.5Mtriangle/s or 400Mpixel/s, Open GL ES 3.2.8 or Vulkan 1.2. This ...

Q1’25 saw global venture funding rise to $121 billion — the highest quarterly total since Q2’22 — driven by OpenAI’s $40 billion raise, which values the company at $300 billion, says CB zinsights. This ties OpenAI with ByteDance as the second-highest-valued private company globally (behind SpaceX at $350 billion. The OpenAI funding round — led by SoftBank and backed by ...

While investment in developing GenAI will increase 76% this year, expectations for GenAI’s capabilities are declining, says Gartner. This paradox will continue throughout this year and next. Spending on developing GenAI spending is expected to total $644 billion this year but, says Gartner vp John-David Lovelock: “Expectations for GenAI’s capabilities are declining due to high failure rates in initial proof-of-concept ...

Jensen Huang’s keynote at GTC 2025 aimed to stoke up the pace of AI development and adoption. “Almost the entire world got it wrong,” said Huang, “the amount of computation we need as a result of agentic AI, as a result of reasoning, is easily 100 times more than we thought we needed this time last year.” Announcing the latest ...

Following on from last year’s Embedded World announcement that Tria Technologies will be working with Qualcomm Technologies on high performance embedded compute modules. The embedded specialist unveiled more of the partnership roadmap in Nuremberg this year. There were five new Tria-branded modules, based on Qualcomm’s Dragonwing and Snapdragon technology at this year’s Embedded World conference. These are the Tria IQ-9075 ...

Synaptics announced an AI MCU at Embedded World today, based around Arm’s Ethos-U55 NPU (neural processing unit) and Cortex-M55 CPU, plus a low-power Cortex-M4 CPU and Synaptic’s in-house mini-NPU. “SR110 dynamically adjusts its compute power based on real-time system demands,” according to the company. “This context-aware computing enables ultra-low-power operation while maintaining high-performance AI capabilities when needed.” There are three ...