ARM is selling its Artisan foundation IP business, including its engineering team, in a fundamental shift.
The processor core designer has always believed that it needs to control the underlying physical IP to deliver the best performance for its chip designs. Back in 2004, the company acquired Artisan for a whopping $913 million. Artisan's foundation IP business includes standard cell libraries, memory compilers, and general-purpose I/O (GPIO), all optimized for advanced process nodes at leading foundries.
The sale to Cadence Design Systems will help Cadence compete with Synopsys in the IP business, especially in the chip space. However, the deal also prepares ARM for a shift to independent chip manufacturing. Controlling the foundational IP that other partners rely on will exacerbate conflicts of interest in licensing agreements.
Cadence has been expanding its design IP product line, with core products including leading protocol and interface IP, memory interface IP, SerDes IP at the most advanced nodes, and embedded security IP with the upcoming acquisition of Secure-IC. The deal is said to have little impact on Cadence's revenue and earnings, so it looks like the deal is more about ARM's strategic positioning. ARM's parent company SoftBank is acquiring ARM-based data center chip design company Ampere for $6.5 billion.
The Artisan acquisition will expand Cadence's presence in the SoC design space and includes an asset purchase agreement and a concurrent technology license agreement. As part of the transaction, Cadence will gain a talented and experienced engineering team with a strong reputation in the industry to help accelerate the development of related and new IP products.
"Over its 25-year history, Arm's Artisan IP has established a strong presence and reputation in the global foundry and SoC partner ecosystem. With the addition of the Artisan IP business and team, Cadence will be able to capture new growth opportunities by entering the foundational IP market," said Boyd Phelps, senior vice president and general manager of the Silicon Solutions Group at Cadence.
With this move, we will gain key technologies and expertise to enhance our design services and silicon offerings, enabling us to implement a comprehensive IP strategy and deliver greater value to our customers. By leveraging Cadence's complete stack of IP, libraries, tools and services, we are committed to growing our foundational IP business while increasing PPA.
"We are committed to ensuring that the foundational physical IP required to deploy ARM technology in all markets remains available to the entire ecosystem," said Kevork Kechichian, executive vice president of solutions engineering at ARM. "The Artisan brand is well established and we believe this technology will continue to play an important role in the semiconductor industry in the future, and Cadence is the ideal partner to drive its growth."
The transaction is expected to close in the third quarter of 2025, subject to regulatory approvals and other customary closing conditions.
In August 2004, British semiconductor design company ARM Holdings Plc agreed to acquire US technology company Artisan for $913 million to expand its product range, but ARM's stock price plummeted on Monday due to concerns that the bid was too high.
ARM said that Artisan Components Inc. shareholders will receive $9.60 in cash and ARM shares equivalent to 4.41 ARM American Depositary Shares (ADS) per share. Based on the closing price of ARM ADS on August 20, 2004, the implied value is $33.89 per Artisan share, totaling about $913 million. Based on this calculation, ARM's acquisition price is about 42% premium to Artisan's closing price of $23.88 on Friday.
Artisan provides designs that help manufacture integrated software and electronic circuits, and its customers include many semiconductor manufacturers.
"Artisan has a mature sales channel that will be highly complementary to ARM's sales channel," said ARM chairman Robin Saxby in a statement.
He added: "The combined entity will be able to reach more customers with a broader product portfolio to better serve our joint user base."
ARM denied that it overpaid for Artisan.
Warren East, then-CEO of Arm, told reporters on a conference call: "We have no record of overpaying."
East added that the acquisition would not have an impact on earnings and that ARM may make further acquisitions.
Arm recently released a report titled "New Thinking in Chips: Laying a New Foundation for the AI Era", which points out that the design architecture of chips in the AI era will undergo many changes. The compatible operation of various AI tool software and hardware is a challenge for edge computing, because many AI developers need to run software on multiple hardware platforms. The lack of standardization between platforms may drag down the performance of AI on some hardware platforms.
Arm Application Engineering Director Xu Dayong pointed out that AI systems are becoming increasingly complex, from large language models (LLMs) to advanced reasoning agents, and the demand for computing power has also increased significantly. These are all challenges for chip designers such as Arm. The workload of AI basic models will be divided between CPU and GPU. A new basic model methodology will make AI inference more complex and require more CPU architectures.
The report points out that the AI revolution is affecting the computing architecture and facing two new trends: the demand for dedicated AI acceleration and the demand for powerful main processors to support new workloads.
In particular, advanced AI models have moved from inference to advanced inference, and even more complex "agents", with a significant increase in computing requirements, bringing huge computing. Advanced inference itself will bring huge computing power requirements. In the long run, enterprises will deploy "agents" that can handle complex tasks, greatly increasing the demand for more advanced training and powerful inference.
More and more AI users are developing their own AI accelerators to meet their own needs. Arm pointed out that not only the cooperation between GPU and CPU, but also Google's self-developed TPU chip, Microsoft's acceleration chip Maia, or AWS's Tranium and Inferentia customized chips still require a host CPU.
However, the cooperation between CPU architecture and other chips has become more complex than before. In the process of building a more powerful and efficient computing system, it is a challenge and a huge opportunity. Opportunities and challenges brought by AI.
1. Improved cooperation in the ecosystem: IP suppliers, foundries, packaging plants and system integrators must work more closely together. In plain words, IP designers cannot develop IP without understanding how the silicon intellectual property is actually implemented in the chip.
2. System-oriented: Think from a system level, not just individual component planning and design, from computing, memory, power supply and thermal management optimization,
3. Standardization requirements: The industry must establish new standards to support the interface, power supply and thermal management of small chips to achieve modular chip design.
4. Power-aware design: Every aspect of chip design takes energy efficiency as the core indicator.
5. Specialization: Different workloads will increasingly require exclusive architectures, making chip design more diverse.
Looking to the future, he said that the semiconductor industry is at a crossroads. Moore's Law of the past is giving way to a complex situation consisting of trade-offs and choices, but challenges will also be opportunities for innovation.
Reference link https://www.eenewseurope.com/en/arm-makes-fundamental-shift-with-artisan-sale/
Source: Content compiled from eenewseurope
View more at EASELINK
2025-01-17
2023-11-13
2024-08-14
2024-12-10
2023-09-26
2025-01-06
2024-04-26
2023-10-13
Please leave your message here and we will reply to you as soon as possible. Thank you for your support.
Sell us your Excess here. We buy ICs, Transistors, Diodes, Capacitors, Connectors, Military&Commercial Electronic components.
Leave Your Message