- Qualcomm builds servers racks and processors, looking at AI inference clusters in a hyperscal data center
- The promised Qualcomm processor could arrive in 2028, long after the competitors have evolved the infrastructure
- But investors were not excited by the plans of the Qualcomm data center, because the equity prices fell on the profits version
Qualcomm could finally be ready to go beyond its mobile bastion and take a real step on the data centers market.
During its recent call for results in the third quarter, the company confirmed that it was in “advanced discussions” with a hyperscal client concerning the new silicon targeting server infrastructure.
This could mark a significant change for the company, which has long teased the ambitions to bring compatible ARM processors to cloud computing spaces but never quite followed on a large scale.
The ambitions grow, but the execution will be critical
The CEO of Qualcomm, Cristiano Amon, has known the moment as an opportunity created by the evolutionary nature of the workloads of the AI, arguing that cloud service providers are increasingly focused on efficiency measures such as Watt tokens and tokens by dollar, not just performance.
This, associated with a wider distance from the industry of X86 processors to alternatives based on personalized arms, could finally give Qualcomm the entry point it was looking for.
However, the market is already crowded with established players who have been building AI infrastructure dedicated for years, and Qualcomm will need more than ambition to gain a significant traction.
Amon explained that the company is developing “a processor for general use” targeting hyperscalers.
“Although we are in the early stages of this expansion, we are engaged with several potential customers and we are currently in advanced discussion with a leading hyperscaller,” he said, “in the event of success, we expect revenues to start during the period for the year 2028.”
Amon has also mentioned other projects, including accelerator cards and even comprehensive servers racks, additions that are intended to support AI inference clusters, Signaling Qualcomm aims to become more than a simple processor supplier.
However, questions remain on the calendar and competitiveness of these efforts.
While Qualcomm expects the revenues of this decision to begin around 2028, this delay could leave it behind more rooted competitors like Broadcom, whose own personalized accelerator enterprise seems to adapt aggressively.
Despite the optimistic tone of the announcement, the reaction of investors was silent.
Qualcomm’s actions briefly dropped following the press release, suggesting that the market remains cautious about the viability of its expansion plans.
The company’s PC activity, while showing modest gains, remains weak – its share of Windows Premium laptops has only reached nine percent, and there is little indication, however, that it holds the fastest CPU in any major computer category.
Competition is also looming in the main segment of Qualcomm, and its rival, Samsung, indicated that it was preparing to bring its own advanced SoCs in flagship mobile devices by 2026.
This rivalry suggests that Qualcomm could try to diversify by necessity, rather than a position of strength.
For the moment, Qualcomm’s jump in the workstation and the data center space is still mainly theoretical.
It remains to be seen if he is a serious player in the credible IA infrastructure and to become a serious player in the IA infrastructure.
Via the register