Chips and Semiconductors

Processors, GPU units, on-device artificial intelligence, chip manufacturing, and industry news.

74 Articles

Chips and Semiconductors

Nvidia Revenue Surges to $96.2 Billion in the Second Quarter of Fiscal 2027

Nvidia reported revenue of $96.2 billion in the second quarter of its fiscal 2027, an annual increase of 106%, driven by data center revenue that reached $89 billion. The company expects revenue of $108 billion in the following quarter, without assuming any revenue from data center computing sales in China.

2026-08-27

Chips and Semiconductors

Amazon Doubles Its Order from Nvidia to Two Million Additional GPU Chips for AWS Data Centers

Amazon and Nvidia announced the expansion of their partnership to add another two million GPU units to AWS data centers during 2027 and 2028, along with the integration of Nvidia networking, processor, software, and robotics technologies. The decision comes despite Amazon continuing to develop its own chips, driven by rising demand from startups, enterprises, AI laboratories, and governments.

2026-08-26

Chips and Semiconductors

Apple Unveils M5 Ultra and M6 Chips to Enhance Local AI Processing

Apple announced the M5 Ultra and M6 chips, along with updated Mac Mini and Mac Studio models. The M5 Ultra raises the ceiling for processing and memory capabilities for professional workloads, while the M6 targets users and developers running AI models on-device.

2026-08-25

Chips and Semiconductors

Three Research Paths Redefining the Transistor, Memory, and Semiconductor Materials

Recent research from MIT and American and Korean universities presents three different directions in hardware: bacterial circuits for environmental monitoring, an antiferromagnetic device that harnesses forgetting to process time series, and a high-entropy material combining semiconducting transport with low thermal conductivity.

2026-08-25

Chips and Semiconductors

IBM Reveals Dual-Architecture Main Processor Running Arm and Z Software on the Same Cores

IBM announced a new main processor designed to run Arm and IBM Z workloads on the same cores, along with support for a new Spyre accelerator for artificial intelligence. The processor will appear in the successor to the z17, expected to launch around 2028, while Arm support is currently limited to Linux and all performance and availability promises have not yet been demonstrated in practice.

2026-08-24

Chips and Semiconductors

Intel Unveils Three Architectures to Scale Agentic AI from Data Centers to the Edge

At Hot Chips 2026, Intel unveiled the Diamond Rapids, Crescent Island, and Wildcat Lake architectures to distribute agentic AI workloads across general-purpose processing, model inference, and client and edge devices. The designs rely on Intel 18A, 3D packaging technologies, and open chiplet interconnects.

2026-08-24

Chips and Semiconductors

Chip Industry Week: Price Increases, AI Funding, and Advanced Packaging Expansion

During the week ending August 21, 2026, the semiconductor sector saw coordinated movements in manufacturing prices, funding for AI chip companies, and research into memory, packaging, and optical interconnects. New investments in the United States, along with developments in quantum computing and robotics, also stood out.

2026-08-21

Chips and Semiconductors

CFET Transistors Open a New Race Over Interconnects and Materials

The semiconductor industry is preparing to transition to three-dimensional CFET transistors, which place nFET and pFET transistors on top of one another to increase density and performance. However, their success will depend on interconnect precision, defect control, and warpage. Intel, TSMC, Samsung, and IBM are pursuing different manufacturing approaches, with growing reliance on multiphysics simulation and virtual manufacturing.

2026-08-20

Chips and Semiconductors

Interposer Design Faces Thermal Expansion and Electromigration in 2.5D Packages

As the wiring density required for AI workloads increases, 2.5D packages are emerging as a less complex option than 3D, but they remain vulnerable to thermal deformation, delamination, and electromigration. Recent studies are examining copper microstructure and deep-learning models to predict failure points and reduce simulation costs.

2026-08-20

Chips and Semiconductors

How Will the Custom HBM Memory Market Take Shape with the HBM4 Generation?

HBM4 enables customization of the base logic die within the memory package, opening the door to cHBM designs tailored to specific workloads, particularly for hyperscalers. However, the implementation model will vary from project to project, while die design and the availability of engineering teams will remain the primary bottlenecks.

2026-08-20

Chips and Semiconductors

Why Will AI Computing in the Future Not Rely on a Single Type of Chip?

Data centers are moving toward heterogeneous clusters that combine CPUs, GPUs, NPUs, custom accelerators, and optical interconnects, rather than relying on a single GPU for all tasks. This shift makes software, network management, and power key factors in reducing token costs and improving hardware utilization.

2026-08-19

Chips and Semiconductors

Results from 80 Companies Reveal the Broadening Wave of Demand for AI Chips

A compilation of results from 80 semiconductor companies showed that AI-related demand growth extended from data center processors to memory, advanced packaging, testing and manufacturing equipment. Eleven companies recorded growth exceeding 100%, while results for phones and consumer electronics were more mixed.

2026-08-18

Chips and Semiconductors

Key Developments in the Chip Industry: Expansions in Advanced Packaging and Shifts in Data Center Architecture

Semiconductor Engineering tracked major expansions in packaging and testing capacity, architectural shifts in artificial intelligence data centers, and developments in memory, quantum computing, hardware security, and autonomous vehicles over a single week. Key moves included ASE’s investment of approximately $3.1 billion, Lam Research’s $3 billion plan, and initiatives to standardize optical architectures and power sources for data centers.

2026-08-14

Chips and Semiconductors

Vertical Integration Takes Hold in System and Chip Design

A Semiconductor Engineering analysis argues that the traditional separation between hardware and software teams is no longer suitable for modern systems, particularly as AI workloads accelerate and power and thermal constraints intensify. This is driving companies to design more specialized hardware in parallel with software, while challenges involving tools, cost and standardization remain.

2026-08-13

Chips and Semiconductors

Why NPUs Need Redesigning as Edge AI Shifts to LLMs

Sharad Chol of Expedera analyzes how the transition of edge devices from traditional vision networks to LLM and VLM models is shifting the nature of the bottleneck from computational capacity to memory traffic. He outlines the role of packet-based processing in reducing external data transfers and improving model execution inside vehicles and embedded devices.

2026-08-13

Chips and Semiconductors

Synopsys Develops 112-Gigabit PHY for Linear Optical Links in AI Systems

Synopsys says the new PHY, compatible with the OIF-CEI-112G-LINEAR-PAM4 specification, enables signal processing to be moved to the host SerDes, potentially reducing power consumption, latency, and heat in high-speed optical links. The solution targets AI and hyperscale data center architectures as the industry moves toward 224 gigabits per lane and beyond.

2026-08-13

Chips and Semiconductors

Intel Launches SEPP Program to Expand Education Pathways to Semiconductor Jobs in Four U.S. States

Intel launched its company-funded Semiconductor Education Pathways Program (SEPP) to support educational and career pathways extending from K–12 education to universities, community colleges, and jobs in the semiconductor sector. The program focuses on Arizona, New Mexico, Ohio, and Oregon by funding students, educators, institutions, and community partners.

2026-08-06

Chips and Semiconductors

Intel Expands Advanced Packaging Capabilities in the United States to Support Artificial Intelligence Semiconductors

Intel Foundry uses Foveros, EMIB, and EMIB-T technologies to connect small, specialized chiplets within larger packages, aiming to overcome photomask size limits and improve the performance and efficiency of artificial intelligence systems. The company says its facilities in New Mexico enable packages to be expanded to eight times the current industry standard, and to more than 12 times that standard by 2028.

2026-07-29