Semiconductors, Data Centers & AI

The world of computing is growing in line with increasing demands for problem solving and robotic integration across many major sectors. Quantum computing, semiconductor technologies, and the rise of humanoid robots are some examples of how diverse computing technologies can be.

Silicon, silicon carbide, gallium nitride, and ultra-wide bandgap (UWBG) semiconductors all compete for space in the power electronics market. While data centers drive further commercialization of wide bandgap semiconductors to support increasingly power-hungry servers, other industries such as wind energy are reluctant to switch to newer, less proven technologies. At the same time, UWBG R&D continues to push these materials closer to commercialization, further complicating the power electronics industry's future.

As semiconductor packaging technologies evolve, 2.5D packaging and 3D Cu-Cu hybrid bonding are becoming essential routes to higher performance and improved power efficiency. However, manufacturing these packages at high yield requires careful material selection, process control, and continued innovation in packaging manufacturing techniques.

Advanced semiconductor packaging has become a central technology platform for high performance computing. For AI and HPC processors, performance depends not only on transistor density, but also on memory bandwidth, I/O density, power delivery, thermal management, and the ability to integrate more functional dies within a single package. This article explores the technology development trends in advanced semiconductor packaging technologies.

In this article, IDTechEx analyzes the potential of nuclear small modular reactors (SMRs) to satisfy the growing power demand of AI data centers. Tech hyperscalers like Amazon, Google, and Meta are already making investments in SMR startups, with IDTechEx forecasting the global SMR market to reach US$53.8 billion in 2036 and almost US$300 billion by 2046.

Robots take on a multitude of forms, each with technologies that make them best suited to their own specific environments, either working independently or in the company of humans. This article covers cobots, service robots, humanoids, drones, and in-vehicle AI assistants.

The transition to 800V in electric vehicles has accelerated the commercialization of wide bandgap semiconductors, and has led to increased vehicle range, faster charging, and greater e-powertrain efficiency. AI data centers are being pushed to ever-increasing power ratings to support future generations of AI training and inference. Drawing inspiration directly from EV power electronics, the data center industry is working to develop the HVDC (800VDC) data center, enabled by wide bandgap power electronics innovations.

In December 2025, American semiconductor and technology giant Marvell acquired Celestial AI in a deal worth US$3.25bn. This is part of a flurry of activity in the wider photonics space, with NVIDIA investing $2bn each in Coherent and Lumentum, and IDTechEx's recent industry and technology report - "Silicon Photonics and Photonic Integrated Circuits 2026-2036: Technologies, Markets, and Forecasts" - indicates that 2026 is seeing explosive growth in the photonics industry driven primarily by AI interconnect demands.

The rise in demand for data center usage in line with increasing high-performance computing and AI applications can create negative environmental effects. This article covers solutions to increasing sustainability surrounding data centers, and battery storage technologies that can remove some reliance on energy from fossil fuels.

Join IDTechEx Technology Analyst Matthew Fall on Thursday 21 May for this expert-led webinar, "The 1MW Rack Challenge: Redesigning Data Center Power Delivery for the AI Era". Choose your preferred time zone - the webinar runs three times in one day.