Today's Semiconductors Pulse
SMIC clears final hurdle for $5.97B takeover of SMNC
China’s Semiconductor Manufacturing International Corp (SMIC) secured regulatory approval to acquire the remaining 49 % of its Beijing‑based unit SMNC, valuing the stake at 40.6 billion yuan (≈$5.97 bn). The transaction will be funded by issuing 547.2 million A‑shares at ¥74.20 ($10.91) each, with a 12‑month lock‑up for sellers, making SMNC a wholly‑owned subsidiary.
Techvalley Supplies HBM Inspection Equipment to Samsung Electronics and SK Hynix
Techvalley has delivered custom Teraton 3D X‑ray CT systems to Samsung Electronics and SK Hynix for advanced packaging research, including high‑bandwidth memory (HBM) and through‑glass vias. The scanners achieve sub‑micron resolution—0.5 µm on the Teraton 7 and 0.9 µm on the Teraton 5—and complete a full 3‑D inspection in three to five minutes, far faster than conventional equipment that can take up to two hours. Standard units cost about ₩1 billion (≈$0.75 M), with custom versions priced higher. Techvalley posted 2023 revenue of roughly ₩26 billion (≈$19.5 M) and expects growth from future production‑line orders.
QCOM Margins Squeezed; Hold Off New Longs Till June
QCOM: handset volumes fall; memory-price shock squeezes margins. Auto/IoT gain offsets but Apple modem exit and licensing risk persist. Trade: avoid new longs until June analyst day. — Viktor Kopylov, PhD, CFA More insights: t.me/si14Kopylov
Lattice Materials Breaks Ground on Montana Silicon, Germanium Plant
Lattice Materials, a U.S. silicon and germanium producer, broke ground on an 80,000‑square‑foot plant in Bozeman, Montana. The project receives $18.5 million from the U.S. Department of Defense, aiming to double the company’s footprint and add the largest optical boule growth...
GreenBoost Memory Orchestrator For NVIDIA GPUs Introduces GreenBoost-Proton For Gaming
GreenBoost, an open‑source memory‑tiering solution for NVIDIA GPUs, now offers GreenBoost‑Proton, a Vulkan‑based layer that expands reported VRAM for Linux gaming. After legal pressure forced the original “nvidia_greenboost” repository offline, the developer relaunched the code on GitLab without NVIDIA branding....
Snap Taps Qualcomm for Next-Gen AI Glasses Ahead of Mass Launch
Snap has inked a multi‑year deal with Qualcomm to supply its Snapdragon XR chips for the upcoming Specs smart‑glasses line. The glasses, slated for a consumer launch later this year, will feature on‑device AI that supports voice and gesture control,...
BluGlass Completes Upsized AUS$8m Two-Tranche Placement
BluGlass Ltd secured an upsized AUS$8 million (≈US$5.3 million) two‑tranche placement at AUS$0.24 per share, adding a free‑attaching option exercisable at AUS$0.38 until May 2028. The raise includes a AUS$2.3 million cornerstone from the board, with executive chair Omer Granit committing AUS$2 million (≈US$1.3 million). Proceeds...
Taiwan’s Helium Imports Shift to the U.S. as Geopolitical Risk Reshapes Semiconductor Supply Chains
Taiwan is rapidly shifting its helium imports from Qatar to the United States, a move evident in 2026 trade data. Helium, a critical gas for semiconductor fabrication, cannot be easily substituted, making supply security paramount. Imports from Qatar have plunged...
Supermicro Introduces Compact, Energy-Efficient Systems to Accelerate Adoption of Intelligent Edge AI
Supermicro unveiled a new family of edge‑optimized servers built on AMD’s EPYC 4005 Zen 5 processors. The lineup includes a mini‑1U box, a short‑depth 1U rackmount, and a slim tower, each delivering up to 16 cores, DDR5 memory, PCIe Gen 5 and optional...
NoC Matters: Designing the Backbone of Next-Gen AI SoCs
The article argues that network‑on‑chip (NoC) design has become the cornerstone of modern AI‑centric System‑on‑Chips, dictating performance, power efficiency, and scalability. As heterogeneous accelerators proliferate, data movement dominates system behavior, making NoC topology, buffering, and QoS policies critical. Designers must...
Kepler Communications Deploys 40‑GPU Orbital Cluster, Adds Sophia Space as First Software Partner
Canada’s Kepler Communications has put the biggest orbital GPU compute cluster into service – 40 Nvidia Orin edge processors on ten satellites – and announced Sophia Space as its newest customer to run software on six of those GPUs. The...
VFabTech Launches Consulting Platform to Tackle AI Chip Capacity Bottleneck
VFabTech debuted its semiconductor engineering and consulting platform, aiming to streamline fab development, cleanroom planning, and workforce training for AI‑driven chip makers. Founder Binh Vu says the service addresses the complex, capital‑heavy challenges that have slowed capacity growth, positioning the...
AMD: Memory, Not Compute, Is the Next Bottleneck in AI Data Centers
AMD’s latest blog highlights memory—not compute—as the emerging bottleneck in AI data centers, emphasizing that data movement now drives performance and power limits. The company promotes LPDDR5X, a low‑voltage mobile memory now adapted for servers, claiming superior performance‑per‑watt compared with...
Infineon Holds Top Spot in Automotive Semiconductors for Sixth Consecutive Year as Microcontroller Share Rises to 36%
Infineon Technologies retained its position as the world’s largest automotive semiconductor supplier for the sixth year running, according to TechInsights’ 2025 market‑share report. The overall automotive semiconductor market expanded to $74.4 billion in 2025, up from $69.9 billion in 2024, while Infineon’s...
NVIDIA Hiring More LLVM Engineers To Work On CUDA Tile
NVIDIA announced it is hiring additional LLVM compiler engineers to advance its CUDA Tile programming model. CUDA Tile, unveiled last year, provides a virtual ISA for tile‑based parallelism and ships an open‑sourced intermediate representation built on LLVM's MLIR. The new...
Memory Shortage Pushes Supply Normalization to 2027
Summary of key points from a memory tracker I'm running weekly: Supply normalization now looks pushed to late 2027. New capacity is delayed, DRAM bit supply growth is only ~16% YoY, and that is still below historical norms. Demand Destruction: Memory inflation...
New Memory ETF Fills Gap in Pure HBM
Tema filing for a Memory ETF.. No doubt everyone seeing $DRAM's shocking insta-success and looking to copy. I went looking through other AI ETFs there's really nothing close to DRAM's HBM pure play-ness outside of the South Korea ETF...
Semiconductors Lead Rally: Broadcom, Marvell Breakout
Semiconductor leaders Broadcom and Marvell are powering a rally driven by accelerating AI infrastructure spending. Broadcom secured long‑term AI chip contracts with Google through 2031 and expanded its partnership with Anthropic, while its fiscal 2026 revenue outlook tops $102 billion with...
Dual Copper and Optical Roadmaps Drive Interconnect Success
Guess what is a good business and technology strategy in the interconnect world? Having both a great copper and a great opical roadmap. $CRDO
DRAM Surge Highlights EWY as Underrated AI ETF
Nice explanation on why $DRAM is Hansel-level hot right now (already $50m in volume today, insane numbers for newbie) from @sumitroy2. This is why we had $EWY as our ETF of the Month for Feb, arguably most underrated AI play....
Nvidia’s AI Chip Empire Fuels Bullish Technical Outlook for NVDA
Nvidia reported a 73% revenue surge to $68 billion and analysts project a 77% increase for the next quarter, reinforcing a bullish technical stance on its shares. The company’s CUDA platform, installed on more than 100 million devices and powering 75% of...
TSMC Sole Supplier for Google's 2nm; Intel Rivalry Lies Elsewhere
TSMC remains the only one able to fabricate Google's 2nm chips. The Intel/TSMC competition is about the "backend," not the chip itself.
VFabTech Launches to Help Solve the Semiconductor Capacity Bottleneck Behind the Next Wave of AI and Advanced Manufacturing
VFabTech, a new semiconductor engineering and consulting firm, launched to address the growing capacity bottleneck that’s limiting AI, robotics and advanced manufacturing growth. The company offers end‑to‑end services—from cleanroom planning and equipment qualification to process integration and workforce training—covering the...
Alphabet Expands TPU Partnership with Anthropic, Adding 3.5 GW of Custom AI Chips
Alphabet and Broadcom have extended their Tensor Processing Unit (TPU) agreement, giving Anthropic access to an additional 3.5 gigawatts of custom AI chips starting in 2027. The move follows a $21 billion TPU order for 2024 and underscores Alphabet's push to...
AI Compute Hits Record High, Explore Supply Chain
AI compute at a record high. CASCI takes you inside the AI supply chain. https://t.co/uZIigc0UJL https://t.co/8rUTqg8Xxd
Intel, SambaNova Unveil Joint AI Inference Hardware Blueprint
Intel and SambaNova Systems revealed a co‑developed AI inference architecture that blends Intel Xeon 6 processors, GPUs and SambaNova RDUs. The blueprint, slated for release in the second half of 2026, aims to challenge Nvidia’s dominance by offering a multi‑layered hardware...
From Wooden Boards to White Gloves: How FPGA Prototyping and Emulation Became Two Worlds of Verification… and How the Convergence...
FPGA prototyping and hardware emulation originated from parallel needs—speed and system realism versus deep debugging of ever‑larger designs. Prototyping leverages re‑programmable silicon to run software workloads early, while emulation provides controlled, observable execution for complex verification. Historically served distinct vendor...
Week 15, 2026
Week 15 highlights a surge in semiconductor capital spending and market size, with equipment billings climbing to $135.1 billion as AI compute and advanced packaging drive demand. Gartner projects the industry to reach $1.32 trillion in 2026, underscoring the monetary impact of...
Untitled
PNY Technologies announced three new GeForce RTX 50 Series Slim graphics cards—RTX 5080, RTX 5070 Ti, and RTX 5070—featuring NVIDIA’s Blackwell architecture in a dual‑slot, ultra‑compact form factor. The cards deliver up to 16 GB of GDDR7 memory, PCIe 5.0 support, and high clock speeds, while employing...
China, Iran Weaponize Economy, Outmaneuver U.S. Sanctions
China, Iran weaponized the global economy to beat the U.S. at its own game US...expanded export controls to cover an estimated 20,000 Chinese companies, and tightened controls over advanced chipmaking equipment and jet engines to China. Yes, that is economic coercion. https://t.co/Kd7ckEDhaE
Global Chip Sales Jump over 60% in February on Strong Demand
Global semiconductor sales surged to $88.8 billion in February, a 7.6% rise from January and a 61.8% jump from a year earlier, according to the Semiconductor Industry Association. Growth was broad‑based, with Asia‑Pacific posting the steepest year‑over‑year increase at 93.5%, while...
Murata Starts Mass Production of High-Capacitance Automotive MLCCs
Murata Manufacturing has begun mass production of seven automotive‑grade multilayer ceramic capacitors (MLCCs) that deliver higher capacitance in smaller footprints. The new parts meet AEC‑Q200 standards, with five low‑voltage devices for ADAS and two 25 V power‑line models. Notably, Murata achieved...
Squishy Photonic Switches Promise Fast Low Power Logic
Researchers at the University of Ljubljana have created a liquid‑crystal photonic switch that controls light with light using two sub‑nanosecond laser pulses. The device exploits whispering‑gallery resonances and stimulated emission depletion to suppress the first pulse while amplifying the second,...
‘Memflation’ Pushes Semiconductor Market Across $1.3 Trillion Threshold
Gartner projects global semiconductor revenues to surpass $1.3 trillion in 2026, a 64% jump from 2025, and to reach $1.6 trillion in 2027. The surge in AI‑driven demand is inflating memory prices, with DRAM expected to rise 125% and NAND flash 243%...
Qualcomm Expands Strategic Advanced Driver Assistance Systems, Immersive Eyewear Collaborations
Qualcomm Technologies announced expanded collaborations with Bosch and Snap’s Specs Inc to advance advanced driver assistance systems (ADAS) and immersive eyewear using Snapdragon platforms. Bosch has already shipped more than 10 million Snapdragon Cockpit computers and will now integrate Snapdragon Ride Flex SoCs...
The Man Who Built NVIDIA (with Stephen Witt)
In this EconTalk episode, host Russ Roberts talks with author Stephen Witt about his book on Jensen Wang and NVIDIA, tracing the company’s rise from a modest 1993 startup focused on 3‑D graphics chips for video games to the world’s most...
China Mandates Export Licenses for Gallium and Germanium, Sparking U.S. Supply‑Chain Alarm
China has required export licences for gallium and germanium, two rare‑earth metals vital to U.S. semiconductors and optics. The policy, introduced in 2025 and expanded in 2026, triggered price spikes and a swift strategic response from Washington and Brussels, highlighting...
South Korea's Semiconductor Exports Surge 152% YoY, Driving 36.7% Export Growth
South Korea's total exports climbed 36.7% year‑over‑year in the first ten days of April, propelled by a 152% surge in semiconductor shipments to $8.6 billion. The jump set a new record for the country’s early‑month export performance and highlighted the accelerating...
Amazon Shares AI Revenue Surge, $15B Run Rate Fuels Stock Rally
Amazon CEO Andy Jassy disclosed that AWS AI services now generate a $15 billion quarterly run rate and its in‑house chip business exceeds $20 billion annually. The data, paired with a $200 billion 2026 capex plan and a $100 billion OpenAI contract, lifted the...
Qualcomm Expands Snap and Bosch Partnerships
Qualcomm announced expanded collaborations with Snap and Bosch, aiming to broaden its footprint beyond traditional mobile chips. The company will provide Snapdragon XR solutions to Snap's newly formed Specs Inc, powering its next‑generation smart glasses. Simultaneously, Qualcomm deepened its partnership...
TSMC: The AI Silicon Shortage Is About To Get Worse
Taiwan Semiconductor Manufacturing Co. (TSMC) has become the primary bottleneck in the AI semiconductor supply chain, with its 2 nm and 3 nm production lines fully booked through 2027. The company's advanced‑node and CoWoS packaging capacity are already operating at or above...
Researchers Created a Computer Chip That Can Survive the Heat of a Volcano
Researchers at the University of Southern California have unveiled a memristor‑based computer chip that can operate at temperatures above 700 °C, far surpassing the 200 °C limit of conventional processors. The device retains data for more than 50 hours without refresh and...
Aehr’s Quarterly Revenue Rebounds to $10.3m
Aehr Test Systems posted third‑quarter 2026 revenue of $10.3 million, up from $9.9 million last quarter but down from $18.3 million a year earlier. Non‑GAAP operating expenses rose to $6.31 million and the company recorded a net loss of $1.52 million. Cash and equivalents increased...
NearLink’s Next Phase: From Smartphones to Peripherals, but Still China-Centric
NearLink, branded as SparkLink, is gaining traction in China’s IoT market, anchored by an estimated 12 million NearLink‑enabled smartphones slated for 2026, primarily within Huawei’s ecosystem. Peripheral categories such as earbuds, smartwatches and PC accessories are projected to add another 11 million...
Startup Funding: Q1 2026
Q1 2026 saw private semiconductor startups raise over $8 billion across 80 companies, with 18 rounds exceeding $100 million and two mega‑rounds—Cerebras and Rapidus—reaching $1 billion each. AI‑centric chip designs for inference and high‑bandwidth interconnects dominated the capital, while photonics and agentic EDA...
Why Hardware Monitoring Needs Infrastructure, Not Just Sensors
Chipmakers are turning to comprehensive hardware monitoring infrastructures to handle the growing complexity of modern SoCs, which now contain billions of transistors and multiple power and clock domains. Traditional test and guard‑banding methods no longer provide sufficient visibility, prompting a...
ASRock Introduces Intel Arc Pro B65 Creator and Passive GPUs
ASRock has unveiled two professional graphics cards built on Intel’s Arc Pro B65 GPU: the Arc Pro B65 Creator with active turbine‑style cooling and a fanless Passive model for chassis‑cooled deployments. Both cards share a dual‑slot design, a large heatsink, reinforced backplate, and use...
Intel Nova Lake S Leak Points to 52-Core Desktop CPUs
Intel’s upcoming Core Ultra 400 “Nova Lake S” desktop line is rumored to span from 6‑core entry models to a flagship 52‑core configuration, extending the company’s hybrid architecture to unprecedented scale. The leak shows DDR5‑8000 memory support and a massive 144 MB–288 MB big...
ONEXStation Mini PC Packs Ryzen AI Max+ 395 Into Compact Chassis
OneXPlayer unveiled the ONEXStation Mini PC, a compact high‑performance system built around AMD’s Ryzen AI Max+ 395 APU. It pairs the processor with a Radeon 8060S iGPU, 128 GB of LPDDR5X memory, and a 1 TB PCIe Gen4 SSD, expandable to 8 TB. The...
Intel Brings Fab 34 Back: The Repurchase of the Apollo Stake Marks a Change of Course in Ireland
Intel announced on April 1, 2026 that it will repurchase Apollo’s 49 percent stake in its Fab 34 joint venture in Leixlip, Ireland, for an estimated $14.2 billion. The transaction will be funded with existing cash and roughly $6.5 billion of new debt. Fab 34, a high‑volume...
Helium Shortage Exposes Hidden Bottleneck Threatening Chip Production
A looming global helium shortage, sharpened by recent Strait of Hormuz tensions, is constraining ultra‑high‑purity helium supplies essential for semiconductor fabs. The bottleneck threatens to slow chip output at TSMC, Samsung and other makers, potentially driving up technology prices.