AMD Ryzen™ Embedded V3000 Processor Powers Cisco’s Latest N9300 Switches and 8000 Series Routers for AI Workloads
Cisco has integrated AMD’s Ryzen Embedded V3000 processor into its new N9300 series switches and 8000 series service‑provider routers. The V3000 provides 4‑8 server‑class x86 cores, up to 96 GB DDR5 memory, and high performance‑per‑watt for control‑plane tasks such as routing, telemetry, and AI workload orchestration. Paired with Cisco’s Silicon One G300 ASIC, the solution delivers 102.4 Tbps switching capacity and supports 400G‑1.6T interfaces. The partnership targets AI‑ready data centers, hyperscalers, and edge deployments that demand low latency, security and long product lifecycles.
The Overlooked Chips Powering the AI Boom
Power semiconductors, the unseen workhorses that convert high‑voltage electricity for AI chips, are poised for rapid growth as AI data centers consume ever more power. Omdia projects global revenue for power chips to reach $100 billion by 2029, up from $80 billion...
Meta Buys Tens of Millions of AWS Graviton Arm Cores in a CPU Land Grab
Meta announced a massive purchase of tens of millions of AWS Graviton Arm cores, bundled with the surrounding power, networking and data‑center infrastructure. The deal marks Meta’s continued commitment to Arm‑based AI compute after earlier pledges to be a launch...
Meta Signs Multibillion-Dollar Deal With Amazon to Use Its CPU Chips for AI
Meta Platforms has inked a multiyear, multibillion‑dollar agreement with Amazon Web Services to run tens of millions of AWS Graviton5 CPU cores for its AI agents and related workloads. The contract, lasting three to five years, makes Meta one of...
AI Euphoria Returns as Intel Earnings Sparkle
Intel’s latest earnings beat forecasts, reigniting market optimism about artificial‑intelligence hardware and underscoring the sector’s growth momentum. The upbeat results come as China’s DeepSeek released previews of its most powerful open‑source model, heightening the U.S.–China AI rivalry. Washington responded by...
Samsung Boosts GDDR6 Output for Tesla as Auto Memory Demand Rises
Samsung Electronics has tripled its monthly output of 8GB GDDR6 memory chips to satisfy Tesla’s growing in‑vehicle computing needs. Production was redirected in April to the Hwaseong plant, where capacity now supports Tesla’s infotainment and autonomous‑driving platforms. The ramp‑up follows...
Lam Research Posts Third Straight Record Quarter With $5.84 Billion in Revenue
Lam Research reported a record $5.84 billion in first‑quarter 2026 revenue, up 23.7% year‑over‑year and 9% sequentially, marking its third straight quarter of record sales. Operating profit rose 31% to $2.05 billion, with gross margin near 50% and EPS of $1.45, beating...
SK Hynix to Supply HBM4E Samples in Second Half Using 1c DRAM
SK hynix announced it will ship samples of its seventh‑generation high‑bandwidth memory, HBM4E, in the second half of 2024, using its advanced 1c DRAM process. The company targets mass production of HBM4E by 2025, aiming to catch up with Samsung,...
ABOV Semiconductor Launches A34G43A MCU for Premium Appliances
ABOV Semiconductor introduced the A34G43A MCU, the first AZPro high‑performance microcontroller, built on a 28‑nm eFlash process and featuring a 200 MHz Arm Cortex‑M4F core. The chip consolidates multiple functions—motor control, PWM, ADC, and AES‑128 security—into a single device, aiming at...
Investors Rotate to China’s Chipmakers as DeepSeek Intensifies AI Competition
Chinese semiconductor stocks rallied sharply after DeepSeek released its V4 foundation model, which it touts as competitive with OpenAI and Google DeepMind. SMIC rose 10% and Hua Hong surged 15% in Hong Kong, reflecting investor bets on domestic AI compute...
Kaspersky Discovers Vulnerability in Qualcomm Snapdragon Chips
Kaspersky ICS CERT uncovered a hardware‑level flaw in Qualcomm Snapdragon BootROM firmware that spans several MDM, MSM and SDX50 chip families. The vulnerability, catalogued as CVE‑2026‑25262, lets an attacker with brief physical access bypass secure‑boot, seize sensor data, and install persistent malware....
Intel Soars on Signs AI Boom for CPUs Is Here
Intel's shares surged over 24% to $83 after Q1 results showed unprecedented demand for its Xeon CPUs from AI service providers, prompting the company to sell chips it had previously written off. The strong AI inference market lifted not only...
STMicronelectronics Sends Industrial Chips Into Overdrive
STMicroelectronics reported a 23% year‑over‑year revenue jump to $3.1 billion in Q1, driven by double‑digit gains in analog, embedded processing and RF segments. While power and discrete fell 1.8%, the company’s balance sheet remains strong with nearly $2 billion in cash and...
SK Hynix to Deploy 2,000 Nvidia Blackwell GPUs at Cheongju Fab for AI Infrastructure
SK hynix is installing a dedicated AI system at its Cheongju fab, deploying 250 servers equipped with 2,000 Nvidia Blackwell GPUs. The infrastructure, costing tens of millions of dollars, will support digital‑twin simulations and proprietary AI agents that automate workflow...
What Hormuz Exposed About Our Semiconductor Supply Chain
The closure of the Strait of Hormuz in early March has halted Qatar’s Ras Laffan helium plant, sending spot helium prices soaring and exposing a critical dependency in the semiconductor supply chain. South Korea and Taiwan source roughly two‑thirds of their...
Samsung, SK Hynix Set for Close Race in Q1 Memory Operating Margins
Brokerages now expect Samsung Electronics' memory operating margin to reach the low‑70s percent range in Q1 2024, driven by a sharp rise in commodity DRAM and flash prices. SK hynix reported a record 72% margin last quarter, prompting analysts to...
M31 Achieves Successful Tapeout of eUSB2V2 on TSMC N2P Process
M31 Technology announced the tapeout of its eUSB2V2 interface IP on TSMC’s N2P 2nm‑class process. The IP delivers up to 4.8 Gbps high‑speed USB 2.0 performance while operating at low 1.2 V/0.9 V voltages and consuming roughly 50 mW. Co‑optimization of design, circuit, and...
Bolt Graphics Zeus GPU
Bolt Graphics announced the tape‑out of its Zeus GPU test chip, a compute‑focused accelerator built around an open‑source RISC‑V command processor. The company claims the quad‑silicon Zeus 4C can deliver up to 13× the ray‑tracing performance of Nvidia’s RTX 5090, while the...
AI Boom Is Hiding Korea’s Next Crisis
South Korea’s KOSPI surged to a record 6,400 in early 2026, propelled by AI‑driven demand for semiconductors from Samsung and SK Hynix. Despite the rally, the economy faces deep structural challenges, including chaebol concentration, soaring household debt, and a rigid...
Intel CEO Lip-Bu Tan: ‘This Is A Fundamentally Different Company Today’
Intel reported first‑quarter revenue of $13.6 billion, up 7 percent YoY, driven by a 22 percent surge in its data‑center business amid AI demand. CEO Lip‑Bu Tan credited organizational cuts, a refreshed leadership team, and a stronger balance sheet for the earnings beat...
Intel’s Data Center Revenue Grows 22 Percent Amid Pent-Up AI Demand
Intel reported a 22% year‑over‑year jump in data‑center revenue, reaching $5.1 billion in Q1 2026, which helped lift total quarterly revenue to $13.6 billion—well above analysts' $12.4 billion forecast. The strong performance sparked a more than 14% surge in the stock during after‑hours...
Why Chip Sovereignty Is No Longer About Chips—But Systems
Governments are still spending tens of billions on semiconductor fabs, but AI is reshaping where value is created. Industry leaders say the competitive edge now lies in integrated systems that combine GPUs, CPUs, accelerators, memory, storage and software, not just...
Intel Showcases Wildcat Lake Reference Laptop with Aluminum Chassis and Fanless Design
Intel unveiled a reference laptop built on its new Core Series 3 "Wildcat Lake" silicon, featuring a six‑core 2P+4LPE CPU, integrated Xe graphics, a 17 TOPS NPU and 16 GB soldered RAM in an aluminum chassis. The design offers both a 17 W PL1/22 W...
GPU Renters Are Playing a Silicon Lottery
Researchers from William & Mary, Jefferson Lab, and Silicon Data analyzed 3,500 GPUs across 11 cloud providers and found substantial performance variability even among identical models. The study, using the SiliconMark benchmark, showed up to a 34.5% difference in 16‑bit...
An AI Agent Just Designed a Complete RISC-V CPU From Scratch in 12 Hours
Verkor.io’s agentic AI system, Design Conductor, generated a complete RISC‑V CPU core—named VerCore—in just 12 hours, moving from a 219‑word specification to a GDSII layout ready for fabrication. The five‑stage, in‑order design runs at 1.48 GHz and achieved a CoreMark score...
Hiring for Precision: Why Thermal Stability Expertise Is the New Must-Have in Semiconductor Talent
A recent incident at an Arizona semiconductor fab saw a $2.3 million yield loss when aluminum handling arms warped 42 μm during thermal cycling, exposing the critical role of thermal stability. The root cause was a mismatch in coefficient of thermal expansion...
Cool-Running Solutions Meet AI Core, AI Factory, and Industrial Power Challenges
At APEC 2026, Alpha and Omega Semiconductor (AOS) unveiled a suite of power‑management products aimed at AI‑core, AI‑factory and industrial‑power applications. The lineup includes the 16‑phase AOZ73216QI controller for high‑performance GPUs, SiC and GaN MOSFETs for ultra‑low‑loss DC‑DC conversion, and integrated...
Quantum Chips Could Scale Faster with New Spin-Qubit Readout that Reduces Sensors and Wiring
Researchers at Quantum Motion and UCL unveiled a radio‑frequency electron‑cascade readout that amplifies spin‑qubit signals, boosting signal‑to‑noise ratio by over 35 dB. The technique reads two‑electron spin states in roughly 7.6 µs, a hundred‑fold speed gain versus prior dispersive methods. By eliminating...
SpaceX May Build Its Own GPUs to Reduce Chip Supply Risks
SpaceX disclosed in its S‑1 filing that it may manufacture its own GPUs to mitigate chip supply constraints and rising compute costs. The proposal is listed as a substantial capital expenditure and ties into the Terafab complex, a joint venture...
Adisyn Secures $10 Million Backing Following Recent Graphene Breakthroughs
Adisyn announced an institutional placement that will raise AU $14 million (approximately US $10 million) after achieving two pivotal graphene technology milestones. The round is anchored by Regal Funds Management, which oversees more than $20 billion, and Meitav, Israel’s largest investment house with about $190 billion...
The Chip Sector Is on a Historic Tear, Fueled by some Unsuspecting Stocks
The PHLX Semiconductor Index logged its 17th straight daily gain, extending a record‑long rally that has lifted the sector over 40% this month. Texas Instruments led the surge, jumping nearly 19% and pulling analog‑chip peers On Semiconductor and Analog Devices...
Analyst View: The Rise of AndTek — Why Google and MediaTek Are the New WinTel
Analysts predict a decisive shift from the traditional Windows‑Intel (WinTel) model to an Android‑MediaTek partnership, dubbed “AndTek,” that could dominate the PC market. Windows is seen as a legacy‑heavy platform while Intel struggles with advanced node transitions, opening space for...
Only One European Company in Global EMS Top 30
The latest in4ma and EMSNOW ranking shows that only one European firm, Germany’s Zollner Elektronik, made the global EMS/ODM Top 30, while Asian companies dominate more than half of the list. The Top 100 providers generate about $820 billion, representing 84% of the...
AFE ICs Accelerate Industrial Image Scanning
Cirrus Logic introduced the CS82L4x family of analog front‑end ICs for CIS and CCD sensors used in scanners and industrial imaging platforms. The series—CS82L41, CS82L44, and CS82L46—offers one, four, or six channels with a 24 Msps per‑channel SAR ADC and 16‑bit...
Samsung and Kingston Trigger New SSD Price Hike Above Ten Percent
Samsung Electronics and Kingston Technology announced SSD price hikes exceeding 10 percent, marking the latest wave of upward pressure in the storage market. Samsung, the world’s largest NAND flash producer, has already applied the new pricing, while Kingston is extending...
Introspect Technology Adds New ATE-on-Bench Solution for LPDDR6, DDR6, and HBM Interfaces
Introspect Technology has begun shipping its M5504 High‑Speed Digital Test System, the latest addition to the M Series. The benchtop instrument combines Bit Error Rate Testing, Automatic Test Equipment and System‑Level testing to validate LPDDR6, DDR6 and HBM interfaces under...
AI Now Gobbling up Power and Management Chips for Servers
The chip shortage is now hitting power‑management ICs and Baseboard Management Controllers as AI‑focused servers commandeer capacity. TrendForce cut its 2026 server‑shipment growth forecast from 20% to 13% because lead times for these components have stretched to 35‑40 weeks for...
Here’s How Our TPUs Power Increasingly Demanding AI Workloads.
Google’s Tensor Processing Units (TPUs) are custom‑designed chips that have evolved over more than a decade to handle massive AI workloads. The newest generation delivers 121 exaflops of compute power and doubles the interconnect bandwidth of its predecessor. By focusing...
Musk Bets Tesla's AI Future on Intel Node that Isn't Finished Yet
Elon Musk announced that Tesla will build its next‑generation AI chips using Intel’s not‑yet‑finished 14A process as part of a new "Terafab" manufacturing push. The strategy aims to secure a proprietary silicon supply for autonomous‑driving workloads, sidestepping potential shortages from...
Shares of Apple Supplier STMicroelectronics Jump After Strong Quarter
STMicroelectronics reported first‑quarter revenue of $3.10 billion, a 23% year‑over‑year increase that topped both its own guidance and analyst expectations. The European chipmaker highlighted accelerating AI‑related demand as major tech firms expand data‑center capacity. Shares in Paris jumped as much as...
Cadence Expands TSMC Collaboration for AI Chip Design
Cadence announced an expanded partnership with TSMC to deliver AI‑ready IP, signoff‑ready design infrastructure, and certified end‑to‑end EDA flows for the foundry’s N3, N2, A16 and A14 processes. The collaboration adds DDR5, PCIe 6.0, LPDDR6/5X, and HBM4E IP to Cadence’s portfolio...
Why 6.4 Gbps DDR5 Designs Fail and How to Avoid It
Designing DDR5 PHYs that exceed 6.4 Gbps is no longer a simulation exercise; real‑world PCB loss and timing margins make first‑pass silicon success essential. Most failures stem from the physical layer’s inability to handle signal‑integrity and power‑integrity stresses rather than logical...
Google to Launch Inference-Focused AI Chip Amid Rising Demand for Faster Deployments
Google announced the launch of its eighth‑generation Tensor Processing Unit (TPU) engineered specifically for inference, the phase where AI models generate outputs. The new chip aims to cut latency and cost for AI agents that handle complex, multi‑step requests. Google...
Tesla Plans $3 Billion Chip R&D Fab in Texas, Eyes Terafab Project
Tesla will invest about $3 billion to build a semiconductor research fab at its Giga Texas campus, producing a few thousand wafers per month for design testing. The pilot line serves as a stepping stone to a larger “Terafab” effort, with...
Semiconductor Revenue to Surge 2026: AI, DRAM and NAND Drive 62.7% Growth Amid Supply Crunch
Omdia has lifted its 2026 semiconductor market forecast to a 62.7% rise, projecting revenue near $1.35 trillion. The surge is driven by explosive demand for memory chips, with DRAM expected to double and NAND to grow up to four‑fold as AI...
GigaDevice GD32F5HC Series MCU Grand Launch, Advancing High‑performance Innovation for HMI and IoT Edge Solutions
GigaDevice has launched the GD32F5HC series, a new line of 32‑bit microcontrollers built on an Arm Cortex‑M33 core that runs at 200 MHz. The devices combine high‑frequency compute, up to 2 MB Flash and 320 KB SRAM, and extensive hardware security including TrustZone...
ST Reports Revenue Growth, Improved Margins for 1Q 2026
STMicroelectronics posted first‑quarter 2026 net revenues of $3.1 billion, up 23% year‑over‑year, and improved its gross margin to 33.8% under U.S. GAAP. Operating income rose to $70 million and net income to $37 million, while non‑U.S. GAAP figures showed a $122 million profit. Growth...
TSMC Starts Work On Arizona Packaging Plant
Taiwan Semiconductor Manufacturing Co. (TSMC) has broken ground on a new chip‑packaging plant in Arizona, aiming to have it operational by 2029. The facility will add CoWoS and 3D‑IC advanced packaging capabilities, addressing a current bottleneck that forces many high‑end...
How I Doubled My GPU Efficiency without Buying a Single New Card
A global retailer’s LLM search service was over‑provisioned, running 48 H100 GPUs at 55% average utilization because the inference workload alternates between compute‑heavy prefill and memory‑bound decode phases. Profiling revealed 92% compute usage during a 200 ms prefill, then a drop...
CSA Catapult Issues Annual Report for 2024–25
The UK’s Compound Semiconductor Applications (CSA) Catapult released its 2024‑25 annual report, highlighting a series of strategic milestones. It opened the nation’s first open‑access advanced packaging facility with $3.1 million funding and joined the £5.5 million REACT sustainable electronics centre in Scotland....