The logo of American multinational corporation Qualcomm is pictured at the Mobile World Congress ... [+]
Mobile World Congress Barcelona always brings a flurry of announcements from companies such as Qualcomm, and MWC 2024 was no different. With this year’s releases, Qualcomm leaned into its leadership products in 5G and Wi-Fi and enhanced its spatial computing and AI capabilities.
One common thread through almost all of the major announcements from Qualcomm was the deeper integration of AI into products across its business units. Qualcomm’s investments in AI are longstanding, and now the company is communicating those investments and capabilities more clearly than ever. For some companies, AI has been a fad or a trend to hop on, but for Qualcomm, AI has long been an accelerator for its products that has enabled “special sauce” capabilities that its competitors can’t easily match. Qualcomm also continues investing in its AI solutions to empower OEMs and developers to better leverage its hardware platforms’ built-in capabilities.
The Snapdragon X80 5G modem to RF chipset in Lucite
Snapdragon X80 — 5G Modem
This is Qualcomm’s seventh-generation 5G modem and second-generation AI modem with dedicated tensor cores. It features a maximum theoretical throughput of 10 Gbps downlink and 3.5 Gbps uplink. Qualcomm says it is the first 5G modem with integrated NB-NTN for satellite communications. It also supports six-carrier aggregation in Sub6 bands for extremely fast non-millimeter-wave speeds, which should improve users’ experience of cellular networks that have fragmented spectrum holdings. The X80 is already sampling, and OEM devices are expected to ship with it in the second half of this year. I expect we will probably see the X80 in Qualcomm’s next-generation Snapdragon 5G SoC at the end of this year when it is announced for next year’s flagship phones.
FastConnect 7900 — Wi-Fi 7 Chip
The FastConnect 7900 is Qualcomm’s next-generation Wi-Fi 7 solution, leveraging deeper wireless integration and AI to enhance the user’s wireless experience. The FastConnect 7900 adds UWB integration to Bluetooth and Wi-Fi into a single package manufactured on the 6nm process node for enhanced power efficiency. While performance is still at 5.8 Gbps of maximum throughput across 320 MHz of spectrum, Qualcomm claims a 40% power reduction compared to its FastConnect 7800. This is a considerable improvement and could mean markedly better battery life for devices that are continuously sending or receiving data over Wi-Fi—something that could be especially useful for compact XR devices where every ounce of power matters. Qualcomm also redesigned the RF front-end module offerings that are paired with the FastConnect 7900; together, these enable 50% RF power savings and, when paired with new High Band Simultaneous modules, also offer a 50% reduction in size on the PCB.
Bringing Wi-Fi 7 To Automotive
Just ahead of MWC 2024, Qualcomm also announced its first Wi-Fi 7 solution for automotive applications, the QCA 6797AQ. Qualcomm says this is the industry’ first automotive-grade Wi-Fi 7 solution and will seamlessly integrate with the other connectivity solutions it already has in the car. Bringing Wi-Fi 7 into the car enables a multitude of improvements to range, throughput and latency while also improving power consumption. Considering how many vehicles require Wi-Fi for OTA updates, having a more performant Wi-Fi solution in a car could make the difference between downloading updates in seconds versus hours. This connectivity is a two-way street, meaning that an auto service center can also use this 5.8 Gbps Wi-Fi link to quickly download data from the vehicle when it is being serviced without ever having to connect any cables. Adding Wi-Fi 7 to vehicles will also bring auto manufacturers up to speed on the latest Wi-Fi technology that consumers’ phones are already using, given that most Android flagship phones are shipping with Wi-Fi 7 today.
XR Hub With Vodafone
At MWC 2024, Qualcomm talked more about its XR Hub, a concept that it has developed in partnership with Vodafone. This concept disaggregates compute from the display in XR devices; it uses a pair of glasses with a Snapdragon AR2 Gen 1 chipset and Qualcomm’s latest Wi-Fi products to stream graphics from a separate compute box run by a smartphone-class processor. Then, XR Hub can further enhance the experience by using a fiber or 5G network to connect to even more computing capability in the cloud. Vodafone calls this HyperRealityHub. It also offers a multi-user experience called HyperRealityHub+ that has multiple AR headsets sharing the same compute box.
It’s clear that this solution is still in its infancy, but it offers a way for enterprises to disaggregate compute and reduce the size, cost and complexity of the XR glasses themselves. This is all tied together with Snapdragon Spaces, Qualcomm’s cross-platform software solution that enables developers to engineer software for many different models of XR devices simultaneously. I believe that ISPs and carriers might be the most interested in deploying this type of solution to consumers, especially if it also runs these end users’ TVs and other entertainment services from the same provider. And having a device that can potentially deliver enhanced TV experiences as well as XR could be the way to subsidize the cost of the compute.
Qualcomm AI Hub For Developers
Qualcomm is doing a lot of heavy lifting to ensure that developers have as easy a time as possible deploying AI models on its many different chipsets and types of devices. In support of this, it has introduced the AI Hub, which is a library of fully optimized AI models for deployment across Qualcomm’s platforms. The company already has 75 pre-optimized AI models, which it claims improve inferencing by up to 4x. These models are available on Qualcomm’s website and on Hugging Face and GitHub, which are logical places that developers would go to pull these models. Qualcomm’s AI Hub supports devices and chipsets beyond smartphones, including XR devices, robotics and more.
Gigabit 5G Phones For Under $99
In addition to the many high-end experiences that Qualcomm offers, it is also looking for ways to enable 5G smartphones at price points under $99. This is important because the world’s cellular networks are all transitioning to 5G, and no one wants to lose customers because of that switch. Additionally, making 5G phones more accessible will enhance the user experience and potentially enable new 5G services. (For more on this, see my recent report on the state of 5G and what U.S. carriers are doing to improve it.)
Ericsson’s November 2023 Mobility Report said that out of 8.5 billion mobile subscriptions worldwide, only 1.5 billion are 5G. While North America is currently at 61% 5G saturation, other regions including Sub-Saharan Africa, the Middle East and Central and Eastern Europe are well under 50%—and aren’t projected to hit 50% by 2029. Even Latin America is expected to hit only 51% 5G saturation by 2029. This is why Qualcomm’s goal is to simplify 5G phones by pushing for a two-antenna 5G Standalone solution. Qualcomm claims this would bring 5G to 2.8 billion more users and increase speeds by up to 5x compared to 4G in the same price tier. Qualcomm is calling this initiative “5G for All.”
Qualcomm's satellite testing demo using a robotic arm to simulate a satellite's movement
5G Advanced And 6G R&D Demos
Ahead of MWC 2024, I had a chance to visit Qualcomm’s headquarters in San Diego to witness some of the R&D demos it would be showing off at the conference. After a detailed presentation about all the different areas of innovation, the other audience members and I got to check out some of the R&D devices in action. The first demo was a Giga-MIMO prototype showing the Giga-MIMO antenna array operating in the 13 GHz band with 4096 antenna elements. Qualcomm is actively working with governmental and non-governmental organizations to prepare the 7 GHz to 16 GHz band for 5G Advanced or even 6G use, considering how congested the existing frequency bands are. Qualcomm says that this Giga-MIMO technology would enable mid-band-like coverage at these higher frequencies using the same cellular site density, avoiding costly densification projects. At the preview event, Qualcomm showed how this antenna array would work and actively demonstrated reflection testing and signal behaviors in a controlled environment.
In the same warehouse Qualcomm likes to use for R&D demos, it also demonstrated how it tests the non-terrestrial network satellite capabilities of its new X80 5G modem. To do this, Qualcomm created a robotic arm designed to simulate the signal and speed of a satellite flying overhead. This allows the company’s engineers to test how the system manages the handoff between satellites, among other things. Qualcomm simulated a video call over this test satellite network to demonstrate that it could maintain a connection reliably, even for an application as challenging as video calling. As impressive as this R&D demo was, I do believe we’re many years away from anyone being able to make video calls using NTN.
For the last two demos, Qualcomm turned to spatial computing. First, it created a digital twin of a warehouse and simulated its wireless coverage using a model based on a physical scan of the space. Then, it used an automated guided vehicle to verify the coverage and fine-tune the simulation model. This application of digital twins and real-life AGVs to verify coverage and validate the physically based simulation can be used in factories and warehouses for companies looking to deploy wireless communications.
Anshel Sag using Qualcomm's Boundless AR prototype
Last but certainly not least, Qualcomm gave me a hands-on demo of one of its Boundless AR capabilities, leveraging Snapdragon AR2 Gen 1 reference design AR glasses connected to a smartphone paired with cloud rendering for dynamic distributed computing over a 5G connection. Qualcomm demonstrated this with a 5G sub-7 GHz network using Ericsson infrastructure and Hololight’s cloud rendering platform. It connected these through Snapdragon Spaces to seamlessly transition between edge and cloud rendering for both automotive visualization and digital twin demos in Barcelona. Qualcomm demonstrated how its system analyzes network performance and frame rates to ensure the best user experience by seamlessly switching between local and cloud rendering. Dynamic distributed compute like this will enable more people to access higher quality AR experiences when connectivity is good, and it helps establish an experiential floor using local compute until connectivity improves.
Conclusion
Qualcomm once again grabbed Mobile World Congress by the horns, as what I’ve covered here is just a small portion of the announcements the company made at the event this year. There are literally too many of these announcements to cover in a single article, but I recommend that you follow my colleague Will Townsend’s coverage of MWC for the 5G infrastructure component of the show.
With its releases at MWC, Qualcomm has made it clear that it stands squarely at the intersection of 5G, AI and spatial computing and that there aren’t many, if any, companies in that space innovating as aggressively as it is. After all, Qualcomm recently surpassed IBM as the number-two patent assignee in the U.S.. At the same time, Qualcomm’s IP licensing arm, QTL, is doing quite well for itself. I don’t think we’re remotely done hearing about AI, 5G or spatial computing innovations from Qualcomm, so expect this to continue into the next series of events and launches this year. Mobile World Congress 2024 Barcelona is ongoing this week, so expect more coverage from me and the rest of the Moor Insights & Strategy team.
Moor Insights & Strategy provides or has provided paid services to technology companies, like all tech industry research and analyst firms. These services include research, analysis, advising, consulting, benchmarking, acquisition matchmaking and video and speaking sponsorships. Of the companies mentioned in this article, Moor Insights & Strategy has had or currently has paid business relationships with Ericsson, Microsoft and Qualcomm.
