The 2025 deals reshaping the semiconductor industry

Feature articles | December 30, 2025

By Nick Flaherty

The semiconductor and Electronic Design Automation (EDA) sectors saw significant consolidation in 2025, driven primarily by the transition to the next generation of power-hungry chips for AI data centres.

Synopsys completed its $35bn deal to acquire Ansys for physical modelling, particularly of chiplets, while Marvell acquired Celestial AI for its chip-to-chip optical interconnect. And in the dying days of the year, Nvidia announced it was to acquire the technology of AI firm Groq, bringing native language processing to its graphics processing units (GPUs).

The biggest deal was undoubtedly the completion of the acquisition of the Ansys physical modelling tools by Synopsys in July, a move that was held up by regulators in the UK, Europe and China for over a year. Keysight Technologies was the big winner as a result, taking on the PowerArtist tool from Ansys and the Optical Solutions Group from Synopsys. PowerArtist is used for early-stage power analysis and reduction in semiconductor designs and fits well with the Keysight PathWave suite of tools.

Japanese conglomerate SoftBank has also ramped up its semiconductor activity. The acquisition of CPU designer Ampere Computing will sit alongside AI accelerator chips from Graphcore to provide a complete system for server makers, a step that Graphcore was not able to achieve with its custom POD systems. This will open up competition with AMD’s x86 and MI350 AI cards, as well as Nvidia with its Rubin GPUs and custom Intel x86 chips with NVlink interconnect.

SoftBank also acquired the robotics division from Swiss engineering firm ABB for $5.37bn in October, just over twice its revenue. ABB had been acquiring various companies and technologies for the division, including Spanish robotics company ASTI in 2023.

“SoftBank’s next frontier is Physical AI. Together with ABB Robotics, we will unite world-class technology and talent under our shared vision to fuse Artificial Super Intelligence and robotics — driving a groundbreaking evolution that will propel humanity forward,” said SoftBank Chairman and CEO Masayoshi Son.

Top Semiconductor and EDA M&A Deals in 2025

Month	Acquirer	Target	Deal Value	Significance
Jan	onsemi	Qorvo’s SiC JFET Portfolio	$115 Million	Acquisition of the United Silicon Carbide subsidiary to bolster power semiconductor leadership.
Jan	Cadence	Secure-IC	Not Disclosed	Expansion into embedded security IP for automotive and IoT chiplets.
Feb	NXP Semiconductor	Kinara	$307 Million	Expansion of NXP’s AI edge processing capabilities for industrial and automotive markets.
Mar	SoftBank	Ampere Computing	$6.5 Billion	SoftBank’s move to secure a high-performance ARM-based server chip designer for AI workloads.
Apr	Cadence	ARM Artisan IP	Not Disclosed	Cadence enters the foundation IP market by acquiring ARM’s standard cell and memory compiler business.
May	Siemens	Excellicon	Not Disclosed	Strategic EDA acquisition to integrate advanced timing constraint verification into the Siemens flow.
Jun	Qualcomm	Alphawave Semi	$2.4 Billion	Accelerates Qualcomm’s expansion into datacentre connectivity and high-speed custom CPUs.
Jul	Synopsys	Ansys	$35 Billion	Completion of the mega-merger creating a “silicon-to-system” leader in EDA and simulation.
Sep	Cadence Design Systems	Hexagon D&E	€2.7 Billion	The acquisition of the Hexagon MSC Software simulation suite strengthens its system design and multiphysics analysis portfolio.
Oct	Skyworks	Qorvo	$9.8 Billion	A massive consolidation of the two largest radio-frequency (RF) chip suppliers for the mobile industry.
Dec	Marvell	Celestial AI	$3.25 Billion	Acquisition of photonic fabric technology to lead the shift to optical interconnects in AI data centres.
Dec	NVIDIA	Groq (Assets)	~$20 Billion	A proposed “aqui-hire” and asset purchase to dominate the dedicated AI inference (LPU) market.

The 2025 deal for Celestial AI brings key interconnect technology to Marvell for designing the next generation of chiplet based systems for AI datacentres, particularly for partner Amazon Web Services (AWS). See our CEO interview here.

“The acquisition of Celestial AI is a transformative step in Marvell’s evolution and expands our leadership in AI connectivity, as scale-up becomes the next frontier in AI infrastructure,” said Matt Murphy, Chairman and CEO of Marvell. “This builds on our technology leadership, broadens our addressable market in scale-up connectivity, and accelerates our roadmap to deliver the industry’s most complete connectivity platform for AI and cloud customers.”

“AI infrastructure is transforming faster than ever, and the future demands scale-up fabrics that deliver unprecedented bandwidth, power efficiency, and reach,” said Sandeep Bharathi, President, Data Center Group at Marvell. “By combining our UALink scale-up switch roadmap and Celestial AI’s breakthrough optical scale-up interconnect, we will enable customers to build AI systems that scale beyond the limits of copper and redefine what’s possible in AI data center architecture.”

“At AWS, we aim to be at the forefront of major technology inflections, and we believe optical interconnects will play an important role in the future of AI infrastructure. Building a scalable, high-performance, and power-efficient cloud starts with an approach built upon differentiated technologies,” said Dave Brown, Vice President of Compute and Machine Learning Services at AWS. “Celestial AI has made impressive progress, and we expect their combination with a large-scale semiconductor company like Marvell will help further accelerate optical scale-up innovation for next-generation AI deployments.”

Qualcomm acquiring Alphawave for its IP, and Cadence acquiring ARM’s Artisan physical IP library are both signs of increasing consolidation around chiplets, particularly for 2nm process technology.

However there are signs that the deals are weaker than might be expected.

Celestial had raised $515m as the leading optical interconnect startup, which makes the 5x multiple a little weak for the backers. There will also be challenges for Synopsys to integrate the Ansys tools rather than just offering bundles. Competitor Siemens EDA has already re-architected its database to accommodate multiple physical analysis tools, and Synopsys may have to follow a similar approach to get the full benefit of the acquisition.

Cadence Design Systems has also been focussing on the physical modelling of chips and systems with security at the heart of the design process, and also acquired the MSC simulation tool from engineering firm Hexagon to boost its multiphysics portfolio for system design.

The Nvidia deal for Groq on Christmas Eve also highlights the shift in the industry and Nvidia’s more innovative financial engineering. The deal sidesteps regulatory restrictions by hiring the staff and signing a non-exclusive deal for the inference AI technology. This is also part of the consolidation of AI technology detailed in our roundup:

AI in 2025: chips, robots and the race for scale

As part of the agreement, Jonathan Ross, Groq’s Founder, Sunny Madra, Groq’s President, and other members of the Groq team will join Nvidia to help develop the licensed technology. Groq will continue to operate as an independent company with Simon Edwards stepping into the role of Chief Executive Officer, and GroqCloud will continue to operate without interruption.