Industry News

AI is moving beyond the screen—and into dedicated hardware.

OpenAI and Work Louder have introduced Codex Micro, a compact physical controller designed for managing AI coding agents. Priced at $230, the device includes:

Wi-Fi 7 is accelerating the next generation of smart devices — and PCB technology is becoming a key enabler.

Quectel has introduced new tri-band Wi-Fi 7 modules, including the FCE870X LCC module and FME170X M.2 module, featuring: • 2.4 / 5 / 6 GHz Wi-Fi 7

Altera Expands DDR5-6400 and LPDDR5-6400 Support for FPGA Systems

But selecting the device is only one part of the sourcing decision. For high-speed boards, laminate availability, copper profile, dielectric thickness, impedance capability and manufacturing tolerances need to be checked before the stack-up is fixed. A material that looks suitable on a datasheet…

High-mix electronics manufacturing is still expanding.

Velocity Group has opened a new 80,000-square-foot manufacturing facility in Ohio, tripling its production capacity and expanding support for high-mix, low- to mid-volume electronic and electromechanical products. That is an interesting signal for the wider supply chain.

AI hardware growth is moving investment further upstream into materials.

Fujifilm has completed a new semiconductor-materials production facility in Japan, expanding capacity for chemicals used in manufacturing advanced semiconductors, including AI chips. It is another useful signal about where the electronics supply chain is heading.

AI hardware complexity is moving in both directions—into the package and into the board around it.

SK hynix is breaking ground on a new advanced-packaging facility in Indiana focused on next-generation AI memory. The investment reflects how important HBM and advanced packaging have become to AI-system performance.

The AI hardware boom is moving upstream into semiconductor manufacturing equipment.

Lam Research has started construction on a new advanced R&D lab in Oregon, part of more than $3 billion of planned investment in its global laboratory network. The facility will support process development, materials research and hardware validation for next-generation semiconductors.

AI infrastructure is becoming an increasingly integrated electronics manufacturing business.

Luxshare Precision’s latest outlook points to continued growth across AI data-center infrastructure, alongside its broader communications and electronics businesses. What is interesting is how much of the hardware stack is starting to come together.

Quantum computing is starting to look less like laboratory equipment and more like deployable infrastructure.

Diraq and Equinix have announced plans to install a silicon-spin quantum computer inside a commercial data center in Sydney. The first system will integrate its cryogenic cooling and control electronics into a self-contained platform operating alongside conventional servers.

Edge AI becomes commercially interesting when it moves from the development kit into the product.

SiMa.ai and ARK Electronics have announced a production-ready hardware platform for autonomous drones, combining edge-AI computing with purpose-built drone electronics. That transition is important.

AI data centers are starting to redesign the power architecture around the rack.

Flex has agreed to acquire EPC Power for $4.4 billion, adding software-defined power conversion and 800V data-center technology to its existing compute and liquid-cooling portfolio. The scale of the transaction says something important.

AI processors are getting harder to manufacture before they even reach the PCB.

ASML is working with leading chipmakers including Nvidia, Intel, TSMC and Samsung on future High-NA EUV technology designed to support larger advanced processors. The semiconductor challenge is clear:

AI infrastructure is starting to lock in optical capacity years ahead.

Verizon and Corning have signed a multi-billion-dollar agreement covering more than 80 million miles of optical fiber through 2032. The agreement will support broadband expansion, but it is also intended to build the high-capacity backbone connecting future AI infrastructure.

Edge AI is moving far beyond cameras and factory robots.

A new $1 billion program is backing a 50-satellite constellation from Loft Orbital, with AI processing designed to happen directly in orbit. That means the spacecraft can analyze sensor data before sending the result back to Earth.

The AI hardware investment cycle starts much further upstream than the server rack.

Major chipmakers are committing to ASML’s next-generation High-NA EUV technology as they prepare future semiconductor processes for increasingly demanding AI hardware. Samsung and SK hynix are targeting High-NA production from 2028, while Intel says it has already processed more than one million…

Optical Interconnect Roadmaps Move Beyond 1.6T

At ECOC 2026, Coherent is demonstrating a 3.2T transceiver and a 6.4T Near-Packaged Optics engine for next-generation AI networks. It is also showing technology aimed at 400G-per-lane optical connectivity.

Industry News

Send your board specification for engineering review