In the artificial intelligence chip market, NVIDIA, with its powerful GPU performance and CUDA software ecosystem, has almost a monopoly position. However, supply shortages and high prices are driving its customers to seek alternative solutions.
<h2>Major customers and competitors join forces to participate in open-source projects as alternatives to CUDA</h2>
Major customers of NVIDIA such as OpenAI, Meta, Microsoft, Google, and Amazon, as well as competitors like Intel, AMD, and Qualcomm, seem to have reached a subtle consensus on overturning the dominance of CUDA. These companies are all participating in the open-source language project Triton initiated by OpenAI.
Tech giants like Meta, Microsoft, and Google, who have spent billions of dollars purchasing NVIDIA chips, hope that Triton can help them break NVIDIA's monopoly on AI hardware; Intel, AMD, and Qualcomm also hope to use Triton to poach NVIDIA's customers.
Triton was initially released by OpenAI in 2021 with the aim of enabling code to run smoothly on various types of GPUs.
According to OpenAI's official explanation, Triton aims to provide an open-source environment for writing fast code with higher productivity than CUDA, while offering greater flexibility than other existing DSLs.
Analysts believe that NVIDIA has been able to dominate the AI industry over the past year mainly due to its over 20 years of development of the CUDA system, which is a barrier that competitors find difficult to overcome.
NVIDIA CEO Jensen Huang has stated that his company &#34;not only produces chips, but also builds the entire supercomputer, from chips to systems to interconnects... but most importantly, software.&#34;
He refers to CUDA as the &#34;operating system&#34; of AI.
<h2>Challenging CUDA may be very difficult</h2>
Since the birth of CUDA in 2006, NVIDIA has invested billions of dollars in developing hundreds of software tools and services to accelerate and simplify running AI applications on its GPUs. In terms of the number of employees, NVIDIA has twice as many software engineers as hardware engineers.
David Katz, a partner at AI investment firm Radical Ventures, said:
<blockquote>
I think people underestimate what NVIDIA has actually built. They have built an efficient, user-friendly, and practical software ecosystem around their products, making complex things simple.
</blockquote>
Nevertheless, the high prices of NVIDIA products and the long waiting queues to purchase state-of-the-art equipment continue to drive its major customers to seek alternative solutions or develop GPUs independently.
However, since most AI systems and applications run in NVIDIA's CUDA ecosystem, rewriting code for other GPUs (such as AMD's MI300, Intel's Gaudi 3, or Amazon's Trainium) would require developers to spend a lot of time and take on risks
CEO of AI startup CentML and Associate Professor of Computer Science at the University of Toronto, Ganadi Peshmenko, told the media:
&#34;To compete with NVIDIA in this field, you not only need to build competitive hardware, but also need to make it easy to use. While NVIDIA's chip performance is indeed outstanding, in my opinion, its biggest advantage lies in software.
Compared to NVIDIA, building a competitive chip is already a big challenge, not to mention building the entire software ecosystem and getting people to start using it.&#34;
Although Triton may weaken NVIDIA's market share, Citigroup analysts estimate that by 2030, NVIDIA's share in the generative AI chip market will still be around 63%, meaning that it will continue to maintain a dominant position in the coming years

GOOG

GOOGL

Compared to NVIDIA, creating a competitive chip is already a major challenge, let alone building an entire software ecosystem and getting people to start using it

NVIDIA has been dominating the world for too long! Google, Meta, and Microsoft join forces, hoping to break the CUDA monopoly!