Monday, April 6, 2026
Market Intelligence for Hardware Professionals

Intel revives Rio Rancho fabs with $500M CHIPS Act boost to launch AI chip‑packaging business
Intel is bringing back its Rio Rancho Fab 9 and Fab 11X to build an advanced chip‑packaging operation, funded with $500M from the U.S. CHIPS Act and billions of internal investment. The company forecasts packaging revenue north of $1B and is courting AI‑focused customers such as Google and Amazon with its new EMIB‑T technology.
Also developing:
By the numbers: SpinQ Technology raises $87.2M Series C+

The lineup includes 1-, 2- and 4-channel options with rail-to-rail input/output for sensor and power monitoring uses. The post ROHM adds TLRx728 and BD728x CMOS op amp series appeared first on Engineering.com.
Engineering.com
As of the end of the fourth quarter of 2025, South Korea's SK Hynix and Samsung controlled 57% and 22% of the global revenue share, respectively, while US-based Micron held the remaining 21% share, according to Counterpoint Research.
ET Telecom (Economic Times)

The UK Ministry of Defence has confirmed that the DragonFire high-energy laser weapon will be installed on Royal Navy Type 45 destroyers by 2027, five years ahead of the original schedule.
Tom's Hardware
ASUS has introduced the ROG Strix OLED XG34WCDMS, a 34-inch ultrawide gaming monitor built around a QD-OLED panel and designed to address both performance and durability concerns associated with OLED displays.
Guru3D
Intel is advancing texture compression techniques with its newly introduced Texture Set Neural Compression (TSNC) technology, a neural network-based approach designed to significantly reduce the size of texture assets used in modern graphics workloads.
Guru3D
People are using AI to break CUDA’s moat. Given any pytorch model, it profiles it, ranks bottlenecks by amdahl's law, writes triton or CUDA C++ replacements, and runs 300+ experiments overnight with no human in the loop. - 5.29x over pytorch eager on rmsnorm - 2.82x on softmax - beats torch.compile by 3.44x on softmax and 2.94x on cross entropy - #1 on the vectorsum_v2 B200 leaderboard - single prompt triton FP4 matmul that beats CUTLASS by up to 2.15x
@Adyen — 195x speedup in model training with NVIDIA accelerated computing. Real-time routing & conversion optimisation at $1T+ in payment volume. https://t.co/HxmnbS6Y8V