Achronix Announces New VectorPath 815 FPGA Accelerator Card
Enables accelerated generative AI inferencing and high-performance computing (HPC) with industry leading efficiency
Santa Clara, Calif. – May 15, 2025 – Achronix Semiconductor Corporation, the leader in data acceleration FPGAs, today announced the launch of the VectorPath® 815 (VP815), a new PCIe accelerator card powered by the industry- leading Speedster® 7t1500 FPGA, and engineered to meet the growing demands of AI workloads and high-performance computing (HPC).
FPGAs are uniquely suited for AI inferencing due to their massively parallel architecture, customizable data paths, and efficient handling of sparse and irregular computations — characteristics that traditional GPUs and CPUs struggle with. Unlike fixed-function ASICs, FPGAs provide hardware adaptability, allowing AI models to evolve over time without requiring new silicon. This adaptability is becoming increasingly valuable as AI models trend toward more efficient, lower precision (such as 1-bit, 2-bit and 4-bit) inferencing architectures, which significantly reduce computational overhead through low-precision operations. Achronix FPGAs, featuring integrated machine learning processors (MLPs), are specifically optimized for these compact, quantized workloads, offering unparalleled performance and seamless programmability for rapidly evolving AI applications. The VectorPath 815 accelerator card exemplifies this capability by delivering unprecedented performance, flexibility, and accelerated time-to-market for high-efficiency AI inference deployments.
“As AI and HPC workloads continue to demand extreme performance and energy efficiency, Achronix has developed the industry’s most advanced power/performance-optimized FPGA acceleration platform,” said Jansher Ashraf, Director of AI Solutions Business Development at Achronix. “The VectorPath 815 card delivers greater than 2,000 tokens per second with 10 ms inter-token latency (LLAMA 3.1-8B Instruct) for unmatched generative AI inferencing performance — enabling customers to accelerate bandwidth-intensive, low-latency applications with a greater than 3× total cost of ownership (TCO) advantage vs. competitive GPU solutions.”
VP815 Enables Organizations to Achieve Breakthrough AI Inferencing Performance
- Generative AI and Large Language Models (LLMs): Achieve industry-leading LLM performance, powering agentic AI applications, accelerating text generation, chatbot responses, and AI-powered content creation while delivering significantly lower latency for AI workloads leveraging various size foundational or fine-tuned and task specific LLMs vs. traditional GPU and CPU solutions.
- Automated Speech Recognition (ASR): Transcribe 2,400 simultaneous audio streams with multiple language support, latency under 25 ms, and an industry-leading word error rate (WER) of less than 3%, ensuring real-time, low-latency speech-to-text transcription for real-time agent assist in contact centers, voice assistants, transcription services, and real-time conversational AI applications.
Why Choose the VP815?
- Extreme Performance and Scalability: Speedster 7t1500 FPGAs have thousands of advanced machine learning processors (MLPs) and a 2D network-on-chip (NoC) enabling high-speed data movement throughout the FPGA, combined with high-bandwidth GDDR6 memory interfaces delivering 4 Tbps bandwidth, PCIe Gen5 ×16 and 400G Ethernet interfaces to deliver superior inference performance vs. competitive solutions.
- Flexible AI Workload Acceleration: Unlike fixed-function AI processors, the VP815 allows customization and optimization for specific AI models, ensuring sustained efficiency as workloads evolve.
- Supply Continuity: The VP815 with GDDR6 chip-down memory, offers a lower-cost, more stable, and readily available alternative to other FPGA-based cards that use HBM2e which is under obsolescence.
VP815 Configuration
The VP815 accelerator card offers the following functionality:
- Speedster 7t1500 FPGA with:
- 2,560 MLPs with multi-number format support (capable of up to 82,000 INT4 MACs)
- 2D NoC architecture for high-speed data movement
- 32 112 Gb SerDes with support for PCIe Gen5 ×16 and dual 400 Gb Ethernet channels
- 692k LUTs of programmable logic
- 32 GB of GDDR6 memory –16 channels for 4 Tbps total bandwidth
- 16 GB of DDR4 memory operating at 3200 MHz delivering >200 Gbps bandwidth
- Dual QSFP-DD 112G interfaces
- PCIe Gen5 ×16 interface
The VP815 is fully supported by the Achronix ACE design tools, a leading FPGA development environment that supports Speedster7t FPGAs and Achronix Speedcore™ eFPGA IP.
Achronix Virtual Lab
Achronix introduces the Achronix Virtual Lab (AVL), a dedicated server cluster equipped with VectorPath 815 accelerator cards, enabling customers to efficiently test and evaluate FPGA designs optimized for AI inferencing and high-performance computing applications. The AVL provides remote, secure access to advanced hardware, accelerating development cycles and reducing time-to-market for innovative FPGA-based solutions.
Available Now
The VectorPath 815 accelerator card is now available in volume production. Customers can access the Achronix Virtual Lab (AVL) for remote, streamlined evaluation of FPGA designs optimized for AI inferencing and high-performance computing. For availability and AVL access, please contact sales@achronix.com.
About Achronix Semiconductor Corporation
Achronix Semiconductor Corporation is a fabless semiconductor corporation based in Santa Clara, California, offering high-end FPGA-based data acceleration solutions, designed to address high-performance, compute-intensive and real-time processing applications. Achronix is the only supplier to have both high-performance, high-density standalone FPGAs and licensed eFPGA IP solutions. Achronix Speedster®7t FPGA and Speedcore™ eFPGA IP offerings are further enhanced by ready-to-use VectorPath® accelerator cards targeting AI, machine learning, networking and data center applications. All Achronix products are fully supported by the Achronix Tool Suite which enables customers to quickly develop their own custom applications.
Achronix has a global footprint, with sales and design teams across the U.S., Europe and Asia.
Related Semiconductor IP
- High Bandwidth Memory 3 (HBM3/3E) IP optimized for Samsung SF4X
- ULL PCIe DMA Controller
- Bluetooth Dual Mode v6.0 Protocol Software Stack and Profiles IP
- SENT/SAE J2716 Transmitter
- SENT/SAE J2716 Receiver
Related News
- Achronix and BittWare Launch VectorPath Accelerator Card with Speedster7t FPGAs
- PCI-SIG Certifies VectorPath Accelerator Card for PCIe Gen5 x16 @ 32 GT/s
- Xilinx Expands Alveo Portfolio with Industry's First Adaptable Compute, Network and Storage Accelerator Card Built for Any Server, Any Cloud
- Socionext Adds 8K Encoding Feature to M820L Media Accelerator Card
Latest News
- Codasip: Toward Custom, Safe, Secure RISC-V Compute Cores
- The Arm Evolution: From IP to Platform for the AI Era
- Achronix Announces New VectorPath 815 FPGA Accelerator Card
- Perforce Partners with Siemens for Software-Defined, AI-Powered, Silicon-Enabled Design
- Geopolitical Tensions Fuel a Wave of AI Chip Independence as US and Chinese CSPs Race to Develop In-House ASICs, Redefining the Market Landscape, Says TrendForce