The Fight For The AI Reign: AMD vs Nvidia

AMD announced on 10th October that it will acquire open-source AI startup Nod.ai. This move is seen as a strategic step aimed at challenging NVIDIA's supremacy in the artificial intelligence and graphics processing unit sector. Let’s take a look at the current market landscape and how AMD’s stepping up the game.

What’s Going On In The Market?

Intel and AMD are the two main companies that make Central Processing Units (CPUs) for laptops and desktops. CPU is a processor that processes the basic instructions of a computer.

While Nvidia have always been the leading manufacturer of high-end graphics processing units (GPUs). GPU is a computer processor that uses accelerated calculations to render intensive high-resolution images and graphics.

This is how GPUs deliver substantial acceleration for specific functions like machine learning, data analysis, and various AI tasks.

How Big Is Nvidia In The Game?

Let’s look at the Data Center segment revenue for the past two years which includes chips like Central Processing Units (CPUs), Data Processing Units (DPUs), and Graphic Processing Units (GPUs). Companies share revenue data for their Data Center segment instead for AI chips specifically.

Nvidia's Data Center revenue has experienced a 4x increase in the past two years, and it is projected to hold a dominant market share of over 70% in the AI chip sector.

The company established its dominance by being an early adopter of the AI trend. Following its attainment of a $1 trillion market capitalization earlier in 2023, the company's stock continues to experience a remarkable surge in value.

How AMD Is Fighting This Battle

When we compare Nvidia and AMD, AMD has experienced a comparatively slower growth rate and generated lower revenue. AMD's MI250 chip was determined to be approximately 80% as fast as Nvidia's A100 chip.

AMD is making bold strides to elevate its game in 2023.

The MI300X chip

They introduced the MI300X chip by the end of Q2 2023, and it was claimed to be designed for large language models and other cutting-edge AI models.

The MI300X boasts a memory capacity of up to 192GB, enabling it to accommodate larger AI models compared to competing chips. For instance, Nvidia's counterpart, the H100, supports only 120GB of memory. This could potentially place AMD in a stronger competitive position within the field of AI.

The acquisition of AI start up

Nvidia is currently outfitting nearly every major AI company such as OpenAI. However, there still may be potential for expansion:

• Nvidia’s hardware is expensive

• AI accelerators are in shortage

Big tech companies like OpenAI and Microsoft are actively looking for alternatives.

Nod.ai has developed the SHARK automation software. SHARK enables efficient implementation of AI models on a variety of AMD products, from data centers to edge devices to endpoints.

The acquisition is anticipated to promote the advancement of open-source compiler technologies and the delivery of high-performance AI software tailored for AMD hardware.

To Conclude:

The rivalry between AMD and Nvidia ensures that we'll witness remarkable breakthroughs and cutting-edge solutions in the field of AI hardware.

Stay tuned for further updates on this dynamic competition, and feel free to share your thoughts with us.

That’s it for this week’s dive.