In this article we will see some companies and technologies used also in the AI ​​field (blockchain and not). Surely the most famous is NVIDIA but we can also mention Intel, AMD (and Xilinx), Azure AI (Microsoft), TPU (Google), IPU (Graphcore), Cerebras Systems, Ascend (Huawei), Baidu Kunlun and Tenstorrent. Many blockchain-based cloud infrastructures (Akash, Bittensor, Render, Ocean Protocol, Fetch AI, Numeraire, SingularityNET, etc) use some of these services for AI.

NVIDIA (GPU)
NVIDIA can be used on cloud infrastructure, especially for compute-intensive workloads such as artificial intelligence (AI), machine learning, and deep learning. Akash allows anyone to provide compute resources, including NVIDIA GPU, which are widely used to run AI applications. NVIDIA is famous for its GPU (graphics processing units), which are used for gaming but are also suitable for AI. GPU have a higher parallel processing capacity than CPUs, which makes them ideal for training machine learning models, especially those based on neural networks.
With NVIDIA, it is possible to speed up the training of AI models, reducing the time from days (or weeks) to hours or minutes.
The CUDA (Compute Unified Device Architecture) platform developed by NVIDIA is also used, which allows developers to use its GPU for general purposes, such as artificial intelligence and machine learning.
NVIDIA has created a wide range of AI software and frameworks (such as TensorRT), Deep Learning SDK (tools and libraries to accelerate deep learning on GPU), GPU Cloud (AI resource hub that includes pre-configured AI containers), DGX Systems (hardware systems dedicated to AI, such as DGX servers.
Some AI projects, such as OpenAI, Tesla (for autonomous driving) use NVIDIA GPU to improve their models (they process real-time data from vehicle sensors, making decisions). NVIDIA is also an integral part of the development of AI supercomputers (Selene, one of the world's most powerful supercomputers, is based on NVIDIA technologies and is used to run AI applications at scale) and generative AI (models that generate text, images and videos). In the area of ​​edge AI, NVIDIA provides technologies for AI applications in smaller, distributed devices, such as drones, industrial robots (automation) and IoT devices, where AI computation must be done locally.

AMD (GPU AND FPGA)
AMD (Advanced Micro Devices) is another GPU manufacturer that, like NVIDIA, is used to accelerate AI and deep learning workloads. In particular, the Radeon Instinct series are designed for intensive computing and are used in some AI and blockchain projects.
AMD also offers the ROCm (Radeon Open Compute) platform, an open-source development environment for parallel computing and artificial intelligence, similar to NVIDIA's CUDA. The company Xilinx has been absorbed by AMD, in particular it produces FPGA (Field-Programmable Gate Arrays), useful for artificial intelligence applications since they can be configured to run AI models optimized for specific use cases. Unlike GPUs, FPGAs offer high compatibility and are often used in areas such as autonomous driving and decentralized computing, including Edge Computing (IoT and Augmented Reality).

GOOGLE TPU
Google has developed its own hardware accelerators called TPU (Tensor Processing Units), which are specifically designed to run machine learning models, especially TensorFlow, Google's AI platform. TPUs are highly optimized for processing deep learning models and are used by Google Cloud, as well as in various distributed AI projects. They can be used to develop AI models that could later be integrated into decentralized systems.

INTEL (CPU)
Intel is known in the hardware industry but also for artificial intelligence. Intel Xeon CPU are very popular for training and inferencing AI and machine learning models, especially in combination with FPGAs (Field-Programmable Gate Arrays) or other accelerators.
Intel Nervana has been used to create an architecture specifically for AI. Intel has also built its OpenVINO ecosystem, which makes it easy to optimize AI models for a variety of hardware.

GRAPHCORE
This is a company that specializes in producing hardware for AI. Their architecture called IPU (Intelligence Processing Unit) is specifically designed to accelerate machine learning models, offering an alternative solution to GPUs for AI. Graphcore IPU are used to improve the efficiency of deep neural network computation.

CEREBRAS SYSTEMS
Cerebras Systems has developed Wafer-Scale Engine (WSE), which is the world’s largest AI chip designed to massively accelerate the training of complex AI models. WSE has the capacity to handle massive AI workloads that could be integrated into AI-focused decentralized blockchain platforms. Cerebras is being used in AI blockchains to speed up decentralized computation.

HUAWEI ASCEND COMPUTING
Huawei has developed its own line of AI processors called Ascend, specialized for deep learning and machine learning. Ascend accelerators are used in various industries, from cloud applications to data centers, and are compatible with various AI frameworks.

MICOSOFT AZURE AI 
Microsoft offers Azure AI, a cloud computing infrastructure that provides access to accelerated AI computing power via CPU, GPU (including NVIDIA), and FPGA. Project Brainwave, in particular, is an AI platform from Microsoft that uses FPGA to accelerate AI inferences in real time. Microsoft has a growing ecosystem that combines blockchain and AI, and its AI computing technologies can be integrated with blockchain AI projects.

BAIDU KUNLUN
Baidu has developed its Kunlun line of AI chips, designed to optimize AI model processing in data centers and edge applications. These chips could find use in blockchain AI projects to improve the speed and efficiency of distributed computing.

TENSTORRENT
Tenstorrent is focused on developing high-performance AI chips. Their architecture is optimized to run intensive machine learning tasks, making them a potential alternative to traditional GPU.

Are you interested in ways to earn crypto bonus? Check it out here: Some Sites To Earn Crypto Bonus (Old & New)