100 relevant articles · classified by Haiku 4.5 · ingested daily
Intel deploys $400M in ASML lithography equipment for Panther Lake chip production.

Q1 2026 EDA market shows sustained growth led by APAC demand as more companies insource chip design.

Samsung launches the 990 2TB QLC-based SSD, a budget Gen 4 drive with full throughput at common capacities.

China's chip exports hit $177B in H1 2026, up 96% YoY, driven by surging memory prices and global AI-demand.

Intel invests $5.7B to expand Ireland fab for higher output of Xeon 6 and next-gen Intel 3 chips.

Nvidia slashes its whitelist of authorized customers in Asia by more than half after US pressure to curb GPU smuggling into China.

Osaka Metropolitan University researchers create a programmable thermal device using magneto-optical and phase-change materials that steers heat and remembers its state without power.

Spain's semiconductor players, universities, and research centers meet to strengthen domestic industry collaboration.

Worldwide EDA and SIP revenue grew 12.7% to $5.748B in Q1 2026, driven by AI chip design tool demand and a Chinese sales surge.

Tesla's AI5 chip tapes out at Samsung Foundry on 2nm-class process, months after a separate tape-out at TSMC.

GlobalFoundries VP Sudipto Bose discusses FD-SOI and physical AI at a Taipei event.

Micron commits $500M to GlobalWafers' Texas wafer plant and raises total planned US spending to $250B through 2035.

Intel unveils Starfire, a space-grade Panther Lake SoC with 18A CPU cores and an Intel 3 GPU tile, targeting the U.S. government satellite market.

AMD's experimental Adrenalin driver hints at up to 8x FSR Multi Frame Generation support for current Radeon GPUs.

SiFive's Krste Asanović argues RISC-V is inevitable and its ecosystem is strong in a keynote at the RISC-V summit.

TSMC posts record June sales driven by AI chip demand, bringing first-half revenue to 113 trillion won.
Imec's ITF World 2026 conference frames the semiconductor industry's shift from transistor scaling to system-level design and heterogeneous integration.

An OCP-defined effort called ESUN aims to evolve Ethernet into a lossless, low-latency interconnect for AI accelerator clusters.

TSMC extends strong sales momentum into June, maintaining its lead in the global foundry chip market.

ASML insiders trim positions, but the company's €38.8B order backlog signals sustained demand.

SemiWiki interviews Kneron CEO Albert Liu about the company's edge AI NPU architectures and localized AI infrastructure.

ZEISS moves EUV tooling development to Korea, embedding in the AI memory supply chain.

Architect Labs builds AI-powered systems to accelerate semiconductor design and engineering.

South Korea invests $166M in AI chip startup Rebellions to build a domestic Nvidia competitor.

Georgia Tech researchers publish an open DRAM model for circuit-level analysis of processing-in-memory in 3D DRAM architectures.

ASML stock price holds steady amid ongoing chip-equipment demand trends.

University of Stuttgart researchers publish a study evaluating hardware abstraction layer concepts for software-defined vehicles.

SJSU and Sandia researchers publish a study evaluating self-heating and radiation hardness of 3nm GAA-FET SRAM with novel channel-BDI isolation.

Netrasemi, a Kerala-based startup, is developing an edge AI chip for the Indian market.

Researchers demonstrate a monolithic CMOS platform integrating piezo-optomechanical photonics on an electronic backplane.

Researchers demonstrate a quantum photonic processor built on standard CMOS-compatible processes, operating with single photons at room temperature.

UF-led research uses an AI-driven framework to model and manage thermal behavior in 3D photonic integrated circuits.

SambaNova raises $1B at an $11B valuation to build AI infrastructure.

ASML's upcoming Q2 report is preceded by a $26.5B order backlog and lingering US export policy uncertainty.

Cerebras plans to list on public markets next Thursday, per report.

Nvidia's $24.9B in circular financing via CoreWeave puts pension funds at risk from overconcentration in AI infrastructure debt.

TSMC now controls roughly 70% of the global chip foundry market.

Ireland's data centers consumed 23% of national electricity in 2025, nearly matching residential use, as AI-driven demand surged 10% year-over-year.

ASML faces conflicting signals as SK Hynix places a $26.5B order while Washington launches a probe.

Cerebras IPO analysis covers expected 2026 timeline, pricing, and valuation outlook.

Citi raises TSMC's price target by more than 30% ahead of the company's upcoming earnings call.

ASML's lithography systems face ongoing US-led export controls amid earnings uncertainty.

ASML's July 15 earnings preview argues DUV lithography remains relevant and EUV's dominance is overstated.

Cerebras shares continue falling despite a reported $20B OpenAI contract, as analysts cite broader sector rotation.

ASML heads into earnings amid overlapping export espionage allegations against the company.

DeepSeek plans to develop its own AI chip, per a report.

DeepSeek looks to raise $7B as China's AI arms race intensifies.

ASML faces a record backlog amid political pressures as the July 15 earnings date approaches.

DeepInfra deploys Nvidia Blackwell inference in Canada, addressing data residency but not full sovereignty.

Cerebras targets a $5.5B IPO as the AI chip race moves to public markets.

FuriosaAI's RNGD inference chip is deployed at Equinix Lisbon, bringing the Korean chipmaker to Europe.

TSMC scales its 3DFabric advanced packaging platform for chip stacking.

AMD's RX 9070 GRE GPU gets a $50 price drop to $499 via Newegg promo code.
Cerebras signals expansion plans at the Paris AI Summit, hinting at growth moves ahead.

SK Hynix CEO forecasts 2027 as the worst year of the memory shortage, with supply constraints lasting through 2030.

Microsoft's FY25 emissions rose 25% to 20.3M metric tons, driven by AI data center buildout, threatening its 2030 carbon-negative target.

A hobbyist proof-of-concept runs the 744B-parameter GLM-5.2 model on just 25GB of RAM via a novel sparse MoE approach, albeit at sub-0.1 tokens per second.

ASML launches the YieldStar 385D metrology system for EUV fabs.

Nvidia stock holds steady as sustained AI demand underpins its long-term growth outlook.

An unnamed Chinese chip startup launches the world's first 8-inch 2D semiconductor production line.

ASML's stock price reflects its central position in semiconductor manufacturing supply chains.

Samsung Foundry raises 4nm and 5nm wafer prices by approximately 15%.

ASML's upcoming earnings report will test whether EUV bookings signal sustained AI chip demand.

An analyst questions whether ASML's €1,574 stock price has outstripped the company's fundamentals and growth outlook.

Micron invests $250B to build the largest U.S. chip plant in Clay, New York.

SK hynix lists on the NYSE at a $26.5B valuation, setting a foreign IPO record for a U.S. listing.

SK Hynix debuts on U.S. stock exchange amid rising American demand for semiconductor fabs.

Apple sues OpenAI for allegedly stealing trade secrets through its chief hardware officer and recruited Apple employees.

SK hynix, TetraMem, and USC researchers tape out a memristor-based in-memory compute SoC for edge AI inference, though peak performance is 16x below Copilot+ requirements.

SK hynix raises $26.5B in largest foreign IPO in U.S. history to fund HBM manufacturing expansions.

Rapidus sets 2nm-class wafer pricing at ~$20K, undercutting TSMC's rumored $30K, with HVM planned for H2 2027.

Apple and Broadcom sign a $30B deal, strengthening U.S. supply chain and opening an opportunity for Intel.

Korean V-Die and Japanese MOSAIC proposals flip HBM dies vertically to improve bandwidth, density, and cooling for AI accelerators.

Micron commits $500M to GlobalWafers' Texas 300mm wafer plant via strategic financing and a 10-year offtake deal, raising total US spending plans to $250B through 2035.

Samsung is sampling its Gaia NPU, built on 4nm, with HP and Lenovo for validation ahead of a 2027/2028 system launch.

Micron raises DRAM investment to $250B+ through 2035; Apple commits $30B+ to Broadcom custom silicon.

U.S. chip export controls are pushing Russia's Sberbank toward Chinese processors like Huawei's Ascend 950, potentially creating a parallel supply chain.

Intel preps 28-core Nova Lake-S Xeon CPUs for Dunlow workstation platform with LGA1954 socket and 95W TDP.

AMD's Zen 6 Medusa Point 10-core APU surfaces on Geekbench, beating the Ryzen AI 9 HX 370 by 22% in single-core.

Elon Musk gets FTC clearance to acquire Mesh Optical, a startup building light-based networking hardware for AI data center interconnects.

Gartner forecasts AI-optimized servers will draw 258 TWh by 2027, surpassing all conventional data center hardware, as power availability becomes the binding constraint on AI expansion.

As AI shifts from training to inference, optical interconnect makers are moving silicon photonics closer to the chip to reduce power and latency.

AI infrastructure shifts focus from accelerator scale to heterogeneous rack-level system composition for agentic workflows, elevating the CPU's role in orchestration.

A White House executive order accelerates the federal timeline for migrating to post-quantum cryptography.

Rapidus races toward 2027 2nm mass production at its single Hokkaido fab with no volume customers yet committed.

China's NVDB flags Claude Code for alleged security backdoors that send user data to remote servers; Anthropic says it was an abuse-prevention experiment.

Memory shortages push budget smartphone costs up — DRAM and NAND now account for up to 64% of BOM, driving a projected 22% shipment decline.

JEDEC publishes the SPHBM4 standard, enabling cheaper high-bandwidth memory by using a narrower 512-bit interface on organic substrates.

SiPearl receives the first Rhea1 HPC CPU from the fab, targeting end-of-2026 availability for Europe's first sovereign HPC chip.

Oregon approves 29.7% rate hike for data centers using over 20 MW under the POWER Act, cutting residential bills by 1.3%.

Nvidia frames Vera CPU as a 'max single-threaded CPU at scale' for AI inference, reveals next-gen Rigel Arm core with improved per-core performance.

Two research groups present side-stacked DRAM designs (V-Die and MOSAIC) at IEEE VLSI Symposium to overcome HBM thermal limits.

SambaNova closes a $1B oversubscribed funding round at an $11B valuation led by General Atlantic, and signs JPMorganChase as a customer.

Ambarella uses AI agents to abstract its SDK, letting ISVs port applications to its hardware in days.

South Korea's $880B chip-and-AI push faces a 15-16 GW power deficit as fab construction outpaces grid and water infrastructure.

Longsys, Lexar's parent, forecasts a 60,000%+ profit surge to ~$1.5B for 1H26 on AI-driven memory demand.