PFN and SEMIFIVE Collaborate to Develop the MN-Core L2000 Series of Next-Generation AI Inference Chips
- Targeting over 2x the memory bandwidth of the MN-Core L1000 Series – joint exhibition at SC26
CHICAGO, Nov. 16, 2026 /PRNewswire/ -- Preferred Networks, Inc. (PFN) and SEMIFIVE Inc. (SEMIFIVE) today announced a collaboration to develop the MN-Core L2000 Series*, PFN's next-generation of chips designed specifically for AI inference.

The MN-Core L2000 Series is the next-generation successor to the MN-Core L1000 Series of AI inference chips, which is scheduled for release in 2027. The new series is planned to deliver more than twice the memory bandwidth of the MN-Core L1000 Series to enable faster and more efficient AI inference infrastructure. The collaboration is already underway, with tape-out (completing the chip design and handing over the final design data to the manufacturer) scheduled in the first quarter of 2027.
The MN-Core L Series (www.mn-core.com) is a family of AI chips that PFN is developing with a focus on memory bandwidth, a key determinant of generative AI inference performance. By vertically stacking logic chips based on MN-Core architecture, which combines high computational performance and low power consumption, with DRAM (dynamic random access memory) in a 3D structure, MN-Core L Series achieves substantially higher memory bandwidth than GPUs and other chips used for AI processing. This 3D-stacked structure enables high-speed AI inference and improves power efficiency for applications such as AI agents, providing a new option for AI infrastructure.
Since beginning development of MN-Core in 2016, PFN has spent more than a decade refining the technology through its use in the company's own services and products. Its proprietary MN-Core architecture combines high computational performance and low power consumption. Supercomputers powered by the first generation of MN-Core have topped the Green500 ranking of the world's most energy-efficient supercomputers three times. PFN is now focusing on developing the MN-Core L series, designed specifically for AI inference. For the MN-Core L2000 Series, PFN is responsible for processor architecture and logical design, leveraging the knowledge and expertise gained through its previous development work.

As a turnkey solution provider for custom semiconductors, SEMIFIVE delivers design implementation and comprehensive supply chain management for the MN-Core L2000 Series, powered by its advanced 3D-IC design solution. To navigate the structural complexities of vertically stacking logic chips and DRAM, SEMIFIVE leverages deep technical co-optimization with leading partners—seamlessly bridging DRAM supply, volume manufacturing, advanced packaging, and testing.
To bring this innovative architecture to life, the logic chip will be manufactured using Samsung Foundry's platform technology. Samsung Foundry is also providing a comprehensive 3D-IC design infrastructure—ranging from optimized I/O interfaces for 3D die connectivity to advanced design methodology for power structure optimization—ensuring seamless integration for the DRAM-stacked architecture.
"The MN-Core L2000 Series is the next generation of AI chips, designed to deliver even greater memory bandwidth than the MN-Core L1000 Series, taking another step toward overcoming bandwidth limitations, a major bottleneck for generative AI," said Daisuke Okanohara, Co-Founder and CEO of PFN. "Developing AI chips with a 3D-stacked structure presents more complex technical challenges than conventional chip development, particularly in areas such as power consumption, heat dissipation, and manufacturing yield. With SEMIFIVE's proven track record in turnkey solutions for custom semiconductors, we aim to tackle these technical hurdles and accelerate progress toward overcoming the challenges of today's AI."
"With the paradigm of AI chip competition shifting from pure compute speed to data transfer efficiency, 3D-IC is the ultimate key to breaking through the performance limits of AI silicon," said Brandon Cho, CEO and Co-Founder of SEMIFIVE. "The MN-Core L2000 project with PFN represents a monumental milestone as SEMIFIVE's first full-scale 3D-IC initiative. Leveraging our deep design expertise and robust turnkey execution, we will successfully translate PFN's visionary MN-Core architecture into volume production and set a new standard for AI infrastructure."
"We are pleased to collaborate with PFN and SEMIFIVE to apply Samsung Foundry's platform technology with valuable 3D-IC design infrastructure to an advanced AI inference chip," said Taejoong Song, vice president of Foundry Technology Planning at Samsung Electronics. "This collaboration accelerates the expansion of the 3D-IC ecosystem through SAFE™ and marks a critical milestone in delivering the high-bandwidth, energy-efficient silicon solutions required for next-generation AI infrastructure."
PFN and SEMIFIVE will jointly showcase the MN-Core L2000 Series at PFN's booth (Booth 4606) during SC26 in Chicago, Illinois, USA, from November 17 to 19, 2026.
* The MN-Core L2000 Series is being developed with support from Japan's Acquisition, Technology & Logistics Agency (ATLA) under its Innovative Science and Technology Initiative for Security program, through the project "Research on Ultra-High-Speed Deep Learning and Inference Using 3D-Integrated LSI Technology" (Type S; FY2024–FY2028). |
About Preferred Networks
Guided by its mission to "Make the real world computable and create the future together," Preferred Networks, Inc. (PFN) develops advanced software and hardware technologies through a vertically integrated approach spanning the entire AI value chain—from AI chips and computing infrastructure to generative AI foundation models and solutions across a wide range of industries. Founded in 2014 in Tokyo, PFN currently develops and delivers the MN-Core™ series of AI processors, the PFCP™ cloud platform for AI computing, and the Japan-developed generative AI foundation model PLaMo™. https://www.preferred.jp/en/
About MN-Core: www.mn-core.com
About SEMIFIVE
SEMIFIVE Inc. (KOSDAQ: 490470) is a pioneer of platform based SoC design, working with customers to implement innovative ideas into custom silicon in the most efficient way. Our SoC platforms offer a powerful springboard for new chip designs and leverage configurable domain-specific architectures and pre-validated key IP pools. We offer comprehensive spec-to-system capabilities with end-to-end solutions so that custom SoCs can be realized faster, with reduced cost and risks for key applications such as data center or AI-enabled IoT. With a strong partnership with Samsung Foundry as a leading SAFE™ DSP partner, as well as the larger ecosystem, SEMIFIVE provides a one-stop shop solution for any SoC design needs. For more information, please visit www.semifive.com.
View original content to download multimedia:https://www.prnewswire.com/news-releases/pfn-and-semifive-collaborate-to-develop-the-mn-core-l2000-series-of-next-generation-ai-inference-chips-302940559.html
This press release may not be republished, redistributed, resold or sublicensed without the prior written approval of PR Newswire.
AlphAI analysis
Written by AlphAI from the release above, not by the issuer or PR Newswire. Informational, not investment advice.
The 30-second read
Why it matters
The collaboration adds specificity on the next-generation inference chip’s performance target (more than 2x memory bandwidth vs L1000) and a near-term engineering milestone (tape-out in Q1 2027), with Samsung Foundry providing the manufacturing and 3D-IC design infrastructure.
Market read
Traders may view this as a strategic roadmap update for AI inference silicon, but without financial metrics or signed commercial volume, it is more of a medium-term sentiment catalyst than an immediate earnings driver.
What to watch
No disclosed contract economics, customer commitments, or benchmark results beyond bandwidth targets; execution risks in 3D-stacked power, heat dissipation, and yield are acknowledged but not mitigated with measurable milestones.
Background
PFN is developing the MN-Core L series of AI inference chips, with the L2000 positioned as the successor to the MN-Core L1000 and designed around 3D-stacked logic plus DRAM for higher memory bandwidth.
Ticker impact
PFN announced collaboration with SEMIFIVE to develop the MN-Core L2000 inference chips, targeting 2x+ memory bandwidth and tape-out in Q1 2027.
Likely modest upside bias as investors price improved long-term differentiation in inference efficiency, with volatility tied to future execution milestones.
The article discloses a concrete next-gen product direction (L2000), a quantified performance goal (2x+ bandwidth), and a specific design milestone (tape-out in Q1 2027), but no immediate financial guidance or order volume.
Market effects
Reinforces the industry shift from raw compute toward memory/data-transfer efficiency and 3D-IC packaging for AI inference accelerators.
Highlights Japan-based PFN and Korea-based Samsung Foundry involvement, potentially supporting regional semiconductor ecosystem sentiment around advanced packaging.
If executed, could strengthen the competitive narrative for non-GPU inference silicon and accelerate adoption of 3D-stacked memory architectures worldwide.
Counterpoint
A 2027 release and Q1 2027 tape-out do not guarantee manufacturability, yield, or performance in real deployments, so the market may discount the announcement until results are validated.
Key entities
- companyPreferred Networks (PFN)
Announced MN-Core L2000 development responsibilities for processor architecture and logical design.
- companySEMIFIVE
Will provide turnkey SoC design implementation and supply-chain management for the MN-Core L2000 3D-IC effort.
- companySamsung Foundry
Will manufacture the logic chip using its platform technology and provide 3D-IC design infrastructure.