AI-RAN Alliance: Qualcomm Silently Joins, BUT as Player or Observer?

It appears that Qualcomm is now a member of AI-RAN Alliance, according to the alliance's official website, but there is no indication when the company joined. The company was not a founding member when the AI-RAN Alliance was formed in February 2024 and was not listed on the company website in March 2024. According to ABI Research’s understanding, Qualcomm joined the AI-RAN Alliance in January 2025. Unlike the typical fanfare when a Tier One industry player joins a major alliance, neither Qualcomm nor the AI-RAN Alliance have made any official announcement regarding this membership, which raises questions. This uncharacteristic silence suggests the company may intend to maintain only an observatory position, rather than taking an active role—perhaps merely ensuring it doesn't miss potential opportunities, while limiting its actual contributions.

Joining a prestigious Alliance that has grown from just 5 members at its launch during last year's Mobile World Congress (MWC) to over 72 members spanning the entire technology ecosystem, including 5 network operators, highlights its remarkable momentum. With additional members expected to join at this year's MWC, the alliance's rapid expansion demonstrates its growing industry influence and appeal.

Early specifications/recommendations will be available during 2025, while members will start revealing pilot projects and Proofs of Concept (PoCs) beginning this year and continuing through 2026. Mainstream implementations are not expected to reach the marketplace before 2027 at the earliest.

From the surface, it might seem attractive for a technology innovator like Qualcomm to join the AI-RAN Alliance, as many benefits could be appealing, including reinforcing Qualcomm's position as a technology innovator and leader in cellular standards, notably The 3rd Generation Partnership Project (3GPP) RAN specifications. The membership could also be appealing as a chipset supplier like Qualcomm could grasp industry pain points, technology requirements, and use cases to effectively guide its Intellectual Property (IP), product and solution roadmap, and how AI could be deeply integrated into the network. The membership could enable the company to forge strong alliances and collaborations with mobile operators and infrastructure vendors to integrate and commercially deploy AI-RAN infrastructure.

This is surface-level thinking, as this is what incentivizes any organization to join a successful alliance. However, the commercial reality is something entirely different. Below are some reasons why.

Because the AI-RAN Alliance seems to be heavily centred around NVIDIA's chipsets and systems, Qualcomm's innovation might be constrained and could potentially limit the company's ability to leverage its unique IP portfolio, mainly when it comes to RAN innovation. Indeed, NVIDIA is a dominant member in the alliance with a formidable ability to accelerate the alliance agenda, shadowing the influence of chipset competitors inside the alliance, and reducing its influence to one of a follower. There are also concerns that NVIDIA might be steering standards and requirements toward architectures that favor its CUDA and Graphics Processing Unit GPU-as-a-Service (GPUaaS) agenda, with the GPU potentially becoming the central engine for simultaneously computing Artificial Intelligence (AI) and RAN workloads.

Qualcomm must demonstrate its leadership in both RAN and edge AI, promote its open approach to innovation, and show the industry how to integrate AI natively within cellular infrastructure, particularly RAN. It needs to pursue alternative industry collaborations to the AI-RAN Alliance, and time is of the essence before it's too late (before the AI-RAN Alliance becomes the de facto standard for integrating AI within RAN infrastructure). It should invite like-minded organizations from the industry, including other chipset suppliers, to join an open intelligent RAN alliance that could rival the AI-RAN Alliance. This approach would allow it to maintain strategic autonomy, while still addressing the market opportunity that AI-RAN represents. ABI Research believes that the birth of a new alliance in this area would be highly welcomed by the industry and Qualcomm has a role to play here.

Alternatively, Qualcomm could play a central role in the O-RAN Alliance to shape its influence, mainly related to Radio Intelligence Control (RIC) and Service Management and Orchestration (SMO). Qualcomm's contributions to Open RAN specifications are particularly significant in Working Group 2 (responsible for non-Real-Time (RT) RIC, particularly resource management, policy optimization, and the provision of AI/Machine Learning (ML) models for RAN functions) and Working Group 3, concentrating on near-RT RIC control and optimization. Qualcomm has already been instrumental in advancing RIC technology in collaboration with industry partners such as Verizon and Samsung Electronics. The solution it announced yesterday, integrating Samsung's AI-powered Energy Saving Manager (AI-ESM) with Qualcomm's new Dragonwing RAN Automation Suite's RIC is designed to use AI to enhance energy efficiency within the network. The Qualcomm Dragonwing RAN Automation Suite provides a vendor-neutral platform that enables the development and deployment of rApps, which are applications designed to optimize various RAN functions. However, the O-RAN Alliance is currently losing momentum and the faith of mobile operators outside of the United States. This becomes evident as the financial health of some core members deteriorated in recent months, casting doubt on the organization's sustainability and long-term survival.

Qualcomm is already a pioneer in bringing AI to smartphones and Personal Computers (PCs), and must now extend its leadership by seamlessly integrating intelligence into radio infrastructure and network edges. This strategic evolution should form the cornerstone of its technology roadmap and market diversification strategy, ultimately enabling a truly distributed intelligence ecosystem that thrives at the network's edge.

Based on this ABI Insight, here are recommendations for both Qualcomm and the industry:

• For Qualcomm: Qualcomm should leverage its unique expertise in both RAN and edge AI to either form a new alliance with like-minded organizations focused on open approaches to AI-RAN integration, or significantly strengthen its role within the O-RAN Alliance by expanding its Dragonwing RAN Automation Suite. Whichever path it chooses, it needs to act quickly before the AI-RAN Alliance's specifications become industry standards, ensuring its IP portfolio and architectural approaches aren't sidelined by NVIDIA's GPU-centric vision.
• For the Industry: The broader telecoms and technology ecosystem should encourage healthy competition between different technical approaches, technology influencers, and alliances, ensuring that open standards do not favor specific hardware architectures. This could involve supporting multiple alliances that are exploring different AI-RAN integration methods, balancing participation across the AI-RAN Alliance and O-RAN Alliance, and demanding hardware-agnostic specifications that enable innovation across diverse silicon solutions, ultimately ensuring the most efficient and effective application of AI to enhance RAN performance.