Home » Markets » AI (page 2)

AI

The support or use of AI (artificial intelligence) in electronics, including ML (machine learning), whether in software (supervised, unsupervised or reinforcement learning tools) or hardware (accelerators, GPUs, etc).

DAC 2024: embracing chiplets, 3D-IC and AI

dac-1-300x200.jpg

At the 61st DAC, it was clear that the software and silicon worlds are not colliding but meshing together, bringing opportunities for tools to address the challenges complex chip design presents. By Caroline Hayes. The electronic design automation (EDA) industry is benefiting from the evolution of the semiconductor industry in which the lines between systems and semiconductors are becoming meshed ...

AI back-end server network switch market to reach $80bn market in 2029

IMG_0173-300x200.jpeg

Spending on switches deployed in AI back-end networks used to connect accelerated servers is forecast to approach $80 billion over the next five years, nearly doubling the total data centre switch market opportunity, says Dell’Oro. Current data centre switch market spending is on front-end networks used primarily to connect general-purpose servers but AI workloads will require a new back-end infrastructure ...

ESA to launch AI satellite to show advances in Earth observation

Phsat-2_integrated_article-300x200.jpg

The European Space Agency’s Φsat-2 mission is aiming to push the boundaries of AI for Earth observation. Due to launch in the coming weeks, the Φsat-2 satellite (pronounced PhiSat-2) is equipped with a multispectral camera and AI computing to analyse and processes imagery in real-time. The promise is for smarter and more efficient ways of monitoring our planet AI capabilities ...

Efabless marks chip manufacturing milestone

efabless-chipIgnite-Board-300x200.jpg

At DAC 2024, Efabless celebrated a milestone of 40 commercial companies designed chips using its chipIgnite and Google-sponsored OpenMPW programme. Of these, some are ready for production volumes, the majority are at the prototype/proof of concept stage, said CEO, Michael Wishart. The platform allows start-ups and smaller companies to prototype designs and innovate without having to buy a license, he ...

Altair to acquire Metrics Design Automation

DAC24-logo-300x200.png

At DAC 2024, Altair showcased its 3D-IC design and SimLab multi-physics modelling and simulation tool. Today, the company has announced the acquisition of Metrics Design Automation, a Canadian company led by Joe Costello , former CEO of Cadence Design Systems (1988-1997). The company described DSim as a full feature System Verilog and VHDL RTL simulator. It enables designers to simulate ...

Space companies contract to ESA Zero Debris satellite commitment

Contract_signature_for_the_development_of_zero-debris_LEO_satellite_platforms_pillars-300x200.jpg

The European Space Agency (ESA) has reaffirmed its commitment to Zero Debris satellites, by the year 2030, with three European space companies signing contracts to develop large LEO satellite platforms conforming to its Zero Debris standards. The aim is to significantly limit the proliferation of debris – in Earth and Lunar orbits – by 2030, for all future missions, programmes ...

NoC design tool is cloud-based

DAC24-logo-1-300x200.png

DAC 2024: Believed to be the first cloud-based NoC (network on chip) design tool, the iNoCulator is available on an early access basis from SignatureIP. Today, there is a need to design multiple chipets and design scalable designs with chiplet-based systems, explained Purna Mohanty, CEO of SignatureIP. As the interconnect between compute, storage, memory and I/O blocks on a chip, ...

AI processor intellectual property aimed at TinyML

Ceva NeuPro-Nano block

Ceva has announced a neural network processing core aimed at SoCs needing to run TinyML models. There are two versions: NPN32 with 32 int8 MACs NPN64 with 64 int8 MACs “Both of which benefit from Ceva-NetSqueeze for direct processing of compressed model weights,” according to the company. “NPN32 is optimised for most TinyML workloads targeting voice, audio, object detection and ...

Managing power consumption in modelling the human brain

vicor-1-300x200.jpg

It has been widely reported that the electrical energy required to train large language generative AI models is daunting. Maury Wood of Vicor investigates. The average adult human brain consumes about 0.3kW hours per day. The typical modern artificial neural network ‘brain’ consumes vastly more. For example, it has been reported that training OpenAI’s GPT-3 model, with 175 billion parameters, ...

The AI roadmap is taking shape

nvidia-1-300x200.jpg

The Blackwell architecture took pole position at this year’s Nvidia GTC in California. Caroline Hayes looks at its role in Drive Thor in-vehicle computing in some of the latest models on our roads. Drive Thor targets the 2025 production models, in which more automated and assisted driving will be included. It is Nvidia’s latest in-vehicle computing platform and the successor ...