GlobalFoundries to Acquire MIPS to Accelerate AI and Compute Capabilities

[Daniel Nenni]

Acquisition will expand GF portfolio with cutting-edge RISC-V processor IP and software tools for real-time computing in autonomous mobility, industrial automation, datacenter and intelligent edge applications

MALTA, N.Y., and San Jose, Calif., July 8, 2025 – GlobalFoundries (Nasdaq: GFS) (GF) today announced a definitive agreement to acquire MIPS, a leading supplier of AI and processor IP. This strategic acquisition will expand GF’s portfolio of customizable IP offerings, allowing it to further differentiate its process technologies with IP and software capabilities.

“MIPS brings a strong heritage of delivering efficient, scalable compute IP tailored for performance-critical applications, which strategically aligns with the evolving demands of AI platforms across diverse markets,” said Niels Anderskouv, president and chief operating officer at GlobalFoundries. “Through this acquisition, we will expand our capabilities to offer customers more flexible, RISC-V-based open platform solutions, paired with our differentiated process technologies and world-class manufacturing to help them build world-class products. This acquisition will be a powerful step forward to push the boundaries of efficiency and performance across a broad range of applications in automotive, industrial and datacenter infrastructure.”

MIPS recently expanded its processor IP offerings, based on open RISC-V specification, with the launch of its Atlas portfolio, a comprehensive suite of compute cores designed for real-time and application processing as well as specialized AI edge processing cores. Additionally, MIPS also introduced Atlas Explorer, a virtual platform that enables optimization of performance, power and area to shift-left in the design cycle.

“Becoming part of GlobalFoundries marks the start of a bold new chapter for MIPS,” said Sameer Wasson, CEO of MIPS. “GF’s proven track record in delivering differentiated technologies through a secure, global manufacturing footprint will enhance our ability to accelerate innovation and scale our solutions—unlocking new opportunities in the Physical AI space and driving even greater value for our customers.”

The acquisition of MIPS is subject to the satisfaction of customary closing conditions, including the receipt of required regulatory approvals, and is expected to be completed in the second half of 2025. Following the acquisition, MIPS will continue to operate as a standalone business within GF and serve its customers across a broad range of technologies.

About MIPS
MIPS is the leading provider of compute subsystems for autonomous platforms in automotive, industrial, and embedded markets. With a 40-year heritage in RISC computing innovation and safety capable processing, MIPS is uniquely positioned to simplify the adoption of Physical AI in industrial robotics and automotive applications. MIPS technology is based on the open specification RISC-V instruction set architecture, featuring MIPS pioneering, patented, multi-threaded capabilities to move beyond proprietary legacy architecture lock-ins. For more information, please visit MIPS.com.

About GF
GlobalFoundries (GF) is a leading manufacturer of essential semiconductors the world relies on to live, work and connect. We innovate and partner with customers to deliver more power-efficient, high-performance products for the automotive, smart mobile devices, internet of things, communications infrastructure and other high-growth markets. With our global manufacturing footprint spanning the U.S., Europe and Asia, GF is a trusted and reliable source for customers around the world. Every day, our talented and diverse team delivers results with an unyielding focus on security, longevity and sustainability. For more information, visit www.gf.com.
©GlobalFoundries Inc., GF, GlobalFoundries, the GF logos and other GF marks are trademarks of GlobalFoundries Inc. Or its subsidiaries. All other trademarks are the property of their respective owners.

Forward-looking Information
This news release may contain forward-looking statements, which involve risks and uncertainties. Readers are cautioned not to place undue reliance on any of these forward-looking statements. These forward-looking statements speak only as of the date hereof. GF undertakes no obligation to update any of these forward-looking statements to reflect events or circumstances after the date of this news release or to reflect actual outcomes, unless required by law.
Media Contact:
Erica McGill
erica.mcgill@gf.com
+1-518-795-5240

GlobalFoundries to Acquire MIPS to Accelerate AI and Compute Capabilities

Acquisition will expand GF portfolio with cutting-edge RISC-V processor IP and software tools for real-time computing in autonomous mobility, industrial...

www.globenewswire.com

 

[blueone]

So, now I know what it takes to make Intel look smart lately.

 

[Daniel Nenni]

I have known MIPs since they started in 1984. This has been one crazy journey! MIPs raised $200M+ and I'm guessing this is an asset sale. Rest in Peace MIPs.

• Founded: 1984
• Founder: John L. Hennessy, Stanford University professor (later President of Stanford and Chairman of Alphabet)
• Original Name: MIPS Computer Systems
• Purpose: To commercialize RISC architecture research at Stanford University
• Focus: High-performance RISC-based microprocessors

1992
Acquired by Silicon Graphics (SGI) to develop its own processors

1998
SGI spins off MIPS Technologies as a separate public company

2008–2013
Struggled in mobile chip market; sold patents to Imagination Technologies in 2013

2017
Imagination sells MIPS to Tallwood Venture Capital

2018
MIPS acquired by Wave Computing, a startup in AI hardware

2020
Wave Computing files for bankruptcy; restructures as MIPS LLC

2021
MIPS announces it is transitioning to RISC-V, abandoning its own MIPS ISA

2025
Acquired by GlobalFoundries
MIPS will continue to operate as a standalone business within GF and serve its customers across a broad range of technologies.

 

[blueone]

Oddly, Microsoft uses MIPS-64 cores with four threads per core in their DPUs. Microsoft acquired a company called Fungible in 2023, which was co-founded by Pradeep Sindhu, who also co-founded Juniper Networks. Fungible developed a SoC which falls into the category of "smart NICs" for Ethernet, though Sindhu is credited with popularizing the term DPU, or Data Processing Unit, to describe smartNICs for network-connected storage. (Personally, I think the DPU terminology is not especially descriptive or accurate, but whatever.) So far as I can tell Microsoft is still using the original Fungible architecture, so MIPS cores are integrated into Azure networking architecture. But I wonder... are these Fungible cores legacy MIPS cores (using the original ISA) or just a proprietary version of RISC-V?

 

[coldsolder215]

The IBM ASIC business unit cracked $1B a while back and was intentionally, internally kneecapped for selling parts to competitors in the mainframe business. Those teams had a lot of rarified early ASIC expertise, building processors for major game consoles and at one point even teaching Intel how to design CPUs during their pivot away from memory. So they limped along with IBM and eventually became GF, and then Avera, and then sold off to Marvell where they are doing quite well for themselves nowadays.

Now this noob McKinsey/Mubadala hack is just gonna start walking back all the hard work Tom Caufield did to put this miserable company out to pasture? This is definitely some kind of money laundering scheme, like a pump-and-dump to ride the AI wave and offload a bunch of debt before taking MIPS to bankruptcy. Remember a year or so ago when GF brought on a new CFO who lasted about 2 weeks before quitting? I'm going to speculate wildly and say it was because of whatever this grift will become.

 

[cliff]

Can you guys please explain why you think this is a bad move?

 

[blueone]

Can you guys please explain why you think this is a bad move?

Why would anyone use CPU IP locked-in to GF processes when GF isn't competitive as a foundry?

 

[cliff]

I assume that the RISC-V CPU will adhere to the RVA23 standard, so the software will work with other CPUs that stick to the same extensions. Am I wrong?

We are designing ASICs for drones. Are we dumb for using RISC-V and designing on GF 14, or are we dumb^2?

 

[cliff]

Why would anyone use CPU IP locked-in to GF processes when GF isn't competitive as a foundry?

Mr. Blue, please expand on GF not being competitive as a foundry.

A) Is the problem with their yield?
B) The accuracy of their models?
C) The IP they provide (standard cells, SRAM)?
D) Their prices?
E) Or that they didn't waste their money on fighting physics?

What letters apply?

 

[blueone]

I assume that the RISC-V CPU will adhere to the RVA23 standard, so the software will work with other CPUs that stick to the same extensions. Am I wrong?

You're correct to my knowledge. Due to RISC-V's permissive licensing WRT instruction set extensions, I hear more about companies extending RISC-V architecture to gain a competitive advantage than I do sticking to the ratified specifications.

We are designing ASICs for drones. Are we dumb for using RISC-V and designing on GF 14, or are we dumb^2?

Cliff, seriously? You are not dumb. GF is focusing on specific ASIC markets that are, to be charitable, emerging, and it is too soon to tell if RISC-V will gain traction in those markets which include cores, usually embedded cores, or not. Arm IP is expensive, so I can see the RISC-V temptation. With IP tied to one foundry, and one which focuses on niches, it looks like a stretch to assume it'll achieve broad deployment, which is why I'm a skeptic about the MIPS acquisition for GF. It looks like it limits MIPS more than it helps GF.

 

[Paul2]

MIPS had very, very, very bad sales, which only got worse after a spinoff. The laziest salespeople ever seen in the tech industry.

ARM will send you salesman with a USB stick with all the files, even if you are a 1 man fabless.

 

[Daniel Nenni]

MIPS had very, very, very bad sales, which only got worse after a spinoff. The laziest salespeople ever seen in the tech industry.
ARM will send you salesman with a USB stick with all the files, even if you are a 1 man fabless.

Agreed. MIPs was a poorly managed company from day one. I thought for sure Imagination Technologies would run with the technology but they fumbled it badly and selling it to a VC made no sense whatsoever. The same VC mashed it together with another floundering client (Wave) and it all came crashing down. So many missed opportunities.

I do not know how much GF paid for MIPS but I will find out. It should be enough to require public disclosure. MIPS has a big legacy royalty stream from set top boxes and gaming consoles. Mobileye is a customer as well as Marvell, Microchip and MediaTek. MIPS was in the NOT Arm market which is now dominated by RISC-V.

The only thing I can think of is that GF will go into the chiplet business. How else are they going to grow revenue without adding new process technology?

 

[blueone]

The only thing I can think of is that GF will go into the chiplet business. How else are they going to grow revenue without adding new process technology?

Good thought. They are indirectly in chiplet business already, by being the fab for the Ayar Labs silicon photonics UCIe chipet. If you're correct, the next step could be an Ayar Labs acquisition. Ayar Labs is valued at over $1B, so it would be expensive, but it seems to have a unique product with a bright future.

 

[Paul2]

The only thing I can think of is that GF will go into the chiplet business. How else are they going to grow revenue without adding new process technology?

The only thing I can think of is that GF needs an 8051-replacement-like offering, for which a borderline giveaway cores are needed. At the current price, MIPS ISA, and cores will satisfy the "borderline giveaway" criteria.

They look at Intel, trying to forming a design alliance with "free Xtensa cores if you fab at our factory", and they think they need to do the same beyond their current relationship with EDA makers.