Company’s Intelligent Compute Architecture delivers immediate acceleration without code changes for customers including Sandia National Laboratories

NextSilicon, a leader in next-generation computing solutions for AI and high-performance computing (HPC), today disclosed architecture and performance details for Maverick-2, the next-generation processor enabling order-of-magnitude better compute for HPC and AI:

This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20251022712360/en/

Maverick-2 Dual-Die OCP Accelerator Module

• Built on the Intelligent Compute Architecture, the dataflow-enabled hardware adapts to provide peak performance regardless of software branching or parallelism.
• In benchmarks released today, Maverick-2 achieves up to 10x performance over leading GPUs at 60% less power in algorithmically complex workloads.
• Maverick-2's architecture allows everyone to bring their own unmodified code - CUDA, Fortran, etc., to run out-of-the-box with no specific optimizations necessary.
• Also, we're unveiling Arbel, an enterprise-grade high-performance RISC-V core built on TSMC's 5nm process that surpasses current RISC-V competition as well as Intel Lion Cove and AMD Zen 5—the latest generation cores from industry leaders.

Already running at dozens of customer sites worldwide, including Sandia National Laboratories in its Vanguard-II supercomputer, Maverick-2 is revolutionizing workloads across scientific research, HPC, and AI.

"For eight decades, rigid hardware designs have forced software to adapt, with modern processors dedicating roughly 98% of silicon to overhead and just 2% to actual computation,” said Elad Raz, NextSilicon’s Founder and CEO. “Maverick-2 flips this paradigm on its head by devoting the majority of hardware real estate to compute, moving runtime overhead management to intelligent algorithms and software in real-time through our dataflow architecture.”

Designed for increasingly data-intensive and algorithmically complex workloads, whether today’s transformer models or tomorrow’s breakthrough algorithms, Maverick-2 introduces capabilities that conventional GPUs cannot achieve. Enabling out-of-the-box acceleration for both legacy and emerging applications simplifies deployment and future-proofs investments by ensuring flexibility as workloads evolve.

Market Validation

From HPC pioneers to AI innovators at the leading edge of scientific research, web search algorithms, social media analysis, and financial network modeling, Maverick-2 is seeing strong adoption while addressing today’s most demanding workloads with unmatched efficiency.

“Over the past three years, our collaboration with NextSilicon has focused on evaluating emerging architectures for next-generation HPC workloads critical to national security,” said James H. Laros III, Senior Scientist and Vanguard program lead at Sandia National Laboratories. “The out-of-the-box HPCG performance results are impressive, showing real promise for advancing our computational capabilities without the overhead of extensive code modifications. Maverick-2’s dataflow approach demonstrates strong potential in real-world scenarios, and we’re continuing to explore its full performance envelope."

By removing the need for time-consuming code rewrites or proprietary software stacks, Maverick-2 accelerates time-to-results, reduces operational costs, and unlocks new possibilities for applications ranging from agentic AI real-time decision making or advanced reasoning models processing complex queries to modeling irregular data patterns of complex proteins for new medical treatments.

Leadership Performance Out-of-the-Box

NextSilicon also released detailed performance data for Maverick-2 across a wide range of critical HPC benchmarks. The results match or exceed the industry’s most widely deployed accelerators and were achieved using unmodified, out-of-the-box application code, with no specialized programming, vendor-specific optimizations, or lengthy porting cycles.

Full benchmarking results are available at www.nextsilicon.com/maverick, with highlights including:

• PageRank: 10x higher graph analytics performance than leading GPUs, enabling workloads on 25GB+ graph sizes that GPUs cannot process.
• GUPS (Giga-Updates Per Second): 32.6 GUPS at 460 watts, 22x faster than CPUs and nearly 6x faster than GPUs, for high-throughput databases and AI inference.
• HPCG (High-Performance Conjugate Gradients): 600 GFLOPS at 750 watts, matching leading GPUs at half the power and delivering performance-per-watt comparable to the world's Top 500 systems..
• STREAM: Validated full performance HBM3e for your most demanding applications.

Silicon, Not Slideware

As an exploration of future architecture, NextSilicon today also unveiled Arbel, an enterprise-grade RISC-V core using its own patented intellectual property (IP) on TSMC’s 5nm process. Designed from the ground up to stand toe-to-toe with offerings from AMD and Intel. The test chip demonstrates NextSilicon’s ability to build world-class IP and explore multiple strategic paths from vertically integrated solutions to enabling a broader RISC-V ecosystem in datacenter environments.

“We don’t just talk about innovation, we build it in silicon while others share roadmaps,” stated Raz. “As AI and HPC workloads evolve, Maverick-2’s intelligent, adaptive design ensures our customers' investments remain future-proof while delivering unparalleled performance today. With its pioneering dataflow architecture already transforming computing, and Arbel showcasing our ability to engineer world-class, general-purpose silicon, we’re setting a bold new standard for what’s possible."

Maverick-2 is currently in the hands of leading partners and is available in product volumes now in Q4 2025. To learn more, visit us at booth #3824 during the SC25 Supercomputing Conference, held from November 16-21, 2025, in St. Louis.

About NextSilicon

NextSilicon is revolutionizing the future of high-performance computing with its Intelligent Compute Architecture (ICA), which blends adaptive computing with self-optimizing software-hardware integration. The company’s ICA solutions are designed to address the most demanding computational challenges by delivering unmatched performance, efficiency, and scalability. Industries such as finance, energy, healthcare, manufacturing, and scientific research that rely on intensive and complex computations benefit from its dynamic workload optimization.

Established in 2017, NextSilicon has grown to over 350 employees globally and has raised more than $300 million in funding from leading investors like Amiti, Aleph, Liberty Technology VC, Playground Global, Standard Investments, StepStone, and Third Point Ventures. Its rapidly expanding partner ecosystem includes native support for commercial HPC libraries with partners such as NAG and Bio Team, as well as hosting and integration services with partners like Dell Technologies, Penguin Solutions, Databank, Vibrint, E4, and Partec.

For more information, please visit: www.nextsilicon.com

To join our team and become a “Maverick,” please visit: www.nextsilicon.com/careers

Frequently asked questions about NextSilicon’s Maverick-2 Intelligent Compute Architecture:

1. What is the Intelligent Compute Architecture (ICA)?
   
   NextSilicon’s Intelligent Compute Architecture (ICA) is a new class of acceleration technology that combines software-defined hardware with intelligent runtime algorithms. This architecture enables real-time adaptability and 4x greater performance-per-watt compared to traditional GPUs, making it ideal for AI, HPC, and vector database workloads. Learn more.
   
   
   
   
2. How does Maverick-2 outperform GPUs?
   
   Maverick-2 dynamically adapts its hardware to each workload, delivering over 4x the performance-per-watt of GPUs through intelligent code path optimization. Unlike fixed-function GPU designs, it reconfigures in nanoseconds to accelerate the most critical parts of any application. Learn more.
   
   
   
   
3. How does the data flow architecture solve CPU and GPU bottlenecks?
   
   NextSilicon’s data flow architecture eliminates the inefficiencies of the 80-year-old Von Neumann model by allowing data availability – not a program counter – to drive computation. This approach dedicates most silicon to computation instead of complex control systems, unlocking massive parallelism and consistent performance. Learn more.
   
   
   
   
4. How easy is it to adopt Maverick-2 for existing software?
   
   Maverick-2 supports popular programming languages like C/C++, Fortran, and Python, so most existing applications can run with little to no modification. Developers save months of manual tuning, as the ICA automatically optimizes performance at runtime. Learn more.
   
   
   
   
5. What are the key differences between Maverick-2 ICA and GPUs?
   
   Unlike static GPUs, NextSilicon’s Maverick-2 ICA continuously learns and optimizes in real time, providing immediate performance benefits without costly code rewrites. It delivers superior energy efficiency, adaptability, and future-proof flexibility for AI and HPC workloads. Learn more.

View source version on businesswire.com: https://www.businesswire.com/news/home/20251022712360/en/