Alternatives to NVFP4 KV Cache

OpenAI is an AI research and deployment company dedicated to ensuring that artificial general intelligence benefits all of humanity.

Global technology company renowned for its GPUs, SoCs, and APIs, driving advancements in AI, gaming, and high-performance computing.

Develops low-energy AI processors with spatial dataflow architecture for extended device operation.

Etched is building the hardware for superintelligence by burning the transformer architecture into their chips.

NVIDIA is a global leader in accelerated computing, pioneering GPU-accelerated computing for gaming, professional visualization, data centers, and autonomous machines.

GPU dev environments

AI chipmaker focused on energy-efficient hardware as an alternative to Nvidia.

An enterprise AI platform for building and deploying generative AI applications and machine learning models.

Sama provides high-quality training data and AI annotation services for computer vision and generative AI.

AI-powered platform for converting PDFs and images to editable text, specializing in STEM content.

AI data solutions company providing annotation, validation, and fine-tuning for machine learning.

AI knowledge foundation for regulated enterprises, turning complex data into AI you can trust.

Optimizes Snowflake queries using LLMs to reduce cloud compute costs.

Building a fault-tolerant quantum computer using photonics and silicon manufacturing.

WEKA is a data platform for AI and HPC workloads, delivering extreme performance for GPU clusters, genomics pipelines, and data-intensive applications.

Industry-leading AI infrastructure for ultra-fast AI inference and training powered by Wafer-Scale Engines.

Develops optical chips and software for AI computing, accelerating computations with light.

Developing the Omni Processing Unit (OPU) that unifies CPU, GPU, memory, and connectivity for AI workloads.

AI-native SaaS company empowering businesses with AdTech and MarTech solutions for data-driven decision-making.

Velaura AI develops ultra-low power silicon and software technologies for AI infrastructure, focusing on energy efficiency for data centers and physical AI applications.

A leading semiconductor manufacturer providing advanced chip manufacturing services.

Astera Labs provides purpose-built connectivity solutions for rack-scale AI infrastructure.

Develops AI and Deep Learning processors for data centers and cloud environments.

Designs and licenses RISC-V based custom processor cores and IP for diverse applications.

Custom silicon design and manufacturing for advanced networking and data center applications.

Quantum computing company focused on building fault-tolerant quantum computers and quantum software.

Deepgram is an enterprise voice AI platform providing APIs for speech-to-text, text-to-speech, and voice agents, used by over 200,000 developers to build voice-first products.

Groq develops a high-performance AI inference chip and compiler for ultra-low latency AI applications.

Designs, develops, manufactures, and sells high-performance, high-efficiency server and storage technology and green computing solutions.

Rethinking compute architectures for future processing needs with novel VLSI designs.

Neurophos develops photonic AI chips (OPUs) for extreme efficiency and performance in AI compute.

Commercial quantum computing systems for complex problem-solving.

Semiconductor hardware company accelerating data analytics query processing speed and scale.

Develops the C200 Analytics Processing Unit (APU), an ASIC for massive-scale analytics.

Developing an AI chip with integrated optical components for faster and more efficient AI inference.

Developer of high-performance RISC-V CPU cores and chiplets for data centers and AI.

Building next-generation, sovereign, Nvidia-powered data centers for enterprise AI.

Expedera provides scalable Neural Processing Unit (NPU) semiconductor IP for edge AI inference.

Pasqal develops quantum computing solutions for complex scientific and industrial problems.

South Korean semiconductor company designing AI accelerator chips and software for LLMs and multimodal AI.

Fabless semiconductor company designing and licensing graphics IP for embedded systems.

Developing full-stack embodied AI systems, including humanoid robots and foundational AI models for versatile automation.

Developing a new network switch architecture to break through AI networking bottlenecks.

A Japanese advanced semiconductor manufacturer focused on developing and mass-producing next-generation 2nm logic chips.

Designs high-performance AI processors and RISC-V based compute solutions.

Annapurna Labs invents the next generation of cloud computing infrastructure, focusing on compute, machine learning, storage, and networking.

d-Matrix develops high-performance AI compute platforms and chips designed for efficient inference of large language models and generative AI workloads.

AI computing company providing programmable edge AI hardware and software solutions.

Quantum computer hardware and software developer for various industries.

Google Research's AI compression algorithm that drastically reduces LLM memory requirements by compressing the KV cache.

Designs and manufactures ultra-low power microcontrollers and processors for IoT and wearable devices.

AI chip startup focused on energy-efficient inference acceleration, acquired by AMD.

Develops high-performance AI accelerators and computing solutions for intelligent data centers and edge devices.

Designs AI-specific processors (IPUs) for machine learning and AI workloads.

Develops model-specific AI chips that hardwire AI models directly into silicon for improved performance and efficiency.

Touch develops advanced semiconductor solutions for immersive and interactive experiences.

Designs and manufactures high-performance microprocessors and graphics processors for computing and data center markets.

AttoTude develops high-speed wired (copper) THz interconnects for datacenters.