Tech giants join forces to break NVIDIA's software monopoly, with Qualcomm, Google, and Intel all in the mix

Source: Hard AI

Author: Fang Jiayao

NVIDIA, with its advanced AI chips and powerful software platform CUDA, has become the leader in the field of AI, attracting over 4 million developers worldwide and driving its market value to over $2.4 trillion. Its chips have become the core infrastructure for many developers, including startups, Microsoft, OpenAI, and Google.

However, tech giants Qualcomm, Google, and Intel are joining forces to challenge NVIDIA's dominant position. They plan to break NVIDIA's software monopoly to weaken its market influence. Vinesh Sukumar, head of AI and machine learning at Qualcomm, stated:

" We are guiding developers on how to migrate from the NVIDIA platform to other platforms."

Tech Giants' "Alliance"

NVIDIA not only relies on its hardware but also on the software code accumulated over twenty years, becoming a leader in the AI field and attracting over 4 million developers globally to use its CUDA software platform.

CUDA, as a parallel computing platform and application programming interface (API), allows software developers to use NVIDIA's GPUs for high-performance computing, enabling NVIDIA's GPUs to play a crucial role in artificial intelligence, scientific computing, and graphics processing.

However, tech giant alliances, including Qualcomm, Google, and Intel, are challenging NVIDIA's dominant position. They plan to attack key software targeting NVIDIA chips and strive to provide alternative solutions for developers, reduce dependence on NVIDIA, and drive application migration to other platforms.

In addition, it is reported that the UXL Foundation, composed of multiple tech companies, plans to leverage Intel's OneAPI technology to develop a set of software and tools that support various AI accelerator chips.

The project aims to achieve cross-platform code execution, allowing developers to write code once and then run it on different hardware platforms without the need to customize code for each hardware. This reduces the technical barriers developers face when dealing with different hardware platforms, simplifies the development process, improves efficiency, and accelerates innovation and application of AI technology.

Google is one of the founding members of the UXL Foundation. Recently, Bill Hugo, head of Google's High-Performance Computing Department and Chief Technology Officer, stated that the UXL Foundation is committed to building an open ecosystem under the machine learning framework, providing more hardware choices to help developers efficiently run AI applications. The foundation plans to finalize technical specifications in the first half of this year and strive to complete technical details by the end of the year to achieve a stable and reliable technical solution.

Furthermore, the UXL Foundation is expanding its partner network, including cloud computing companies like Amazon and Microsoft Azure, as well as chip manufacturers, to attract external technical contributions and enrich the technical framework. Currently, the cross-platform AI software tool suite based on Intel's OneAPI technology has been put into use, marking an important step in the development of UXL technology As time goes on, UXL also plans to support NVIDIA's hardware and code, demonstrating its emphasis on compatibility with existing mainstream technologies. In the face of efforts to challenge NVIDIA's dominant position in the industry, NVIDIA executive Ian Buck said, "The world is accelerating, and new ideas for accelerated computing will drive the development of AI and computing capabilities, benefiting the entire ecosystem."

"Encirclement" of Over a Hundred Startups

According to PitchBook data, in order to weaken NVIDIA's dominant position in the AI software field, venture capitalists and corporate funds have invested over $4 billion in 93 different projects, with the UXL Foundation's plan being just one of many efforts.

In 2023 alone, startups dedicated to challenging NVIDIA's leadership received over $2 billion in investments. This is a significant increase from $580 million just a year ago, indicating a growing interest in shaking NVIDIA's position through software vulnerabilities.

However, startups face a significant challenge, as NVIDIA's CUDA software is deeply entrenched in the AI field. CUDA, developed by NVIDIA, is a feature-rich software that has not only received continuous investment but also widespread support and contributions from the developer community.

Jay Goldberg, CEO of D2D Advisory, pointed out:

"The importance of CUDA software lies not only in its functionality, but also in the fact that people have been using CUDA for 15 years, building a large amount of code around it. That's what really matters."

CUDA has been deeply integrated into many artificial intelligence and high-performance computing projects, and changing this status quo requires overcoming significant inertia and dependence. For startups hoping to succeed under NVIDIA's shadow, they not only need to offer competitive technology but also overcome users' deep reliance on existing technologies.

As more funds and resources pour into this field, the industry remains cautious about whether these startups can pose a substantial challenge to NVIDIA in the AI software market