The News: At Mobile World Congress (MWC) Barcelona, Qualcomm’s AI Research arm is showcasing a range of important advancements for on-device AI such as large multimodal models (LMMs) and customization of large vision models (LVMs) running on Android smartphones and Windows PCs, and more, highlighting both the increasing importance of on-device AI capabilities and Qualcomm’s market opportunity in bringing these capabilities to its OEM partners. Read more about Qualcomm’s on-device AI releases at MWC 2024 on the Qualcomm media page.
Qualcomm Creates Opportunities for the On-Device Generative AI Market
Analyst Take: Qualcomm is hitting the gas on its on-device use case offensive for AI, as well it should. When generative AI hit the world by storm a little over a year ago, most of the discussions we were having were about data center capacity and scale to support large language model (LLM) training and inference workloads for apps such as ChatGPT. Fast-forward to today, and the discussion has shifted from “how do we scale generative AI in the data center” to “how much AI training and inference can we assign to a device?” Just like NVIDIA had the right product at the right time with its data center GPUs, Qualcomm had the right platform at the right time with its Snapdragon family of SOCs and its intellectual property (IP).
I remember pointing out that Qualcomm was an AI company long before anyone else noticed, and it feels good to see that observation validated. Qualcomm’s focus on building more AI capabilities into its platforms accelerated in the past year, for obvious reasons, redefining the parameters of Qualcomm’s market relevance for the next decade and opening entirely new opportunities for growth for its platforms. The company is leveraging the 2024 Mobile World Congress in Barcelona to not only showcase its latest advancements in on-device AI features but drop some important hints about how the future of AI-powered user experience (UX) is likely to be shaped by the on-device capabilities of handsets, PCs, IoT, XR, automotive, and more.
Mobile
For the first time running on an Android smartphone, Qualcomm AI Research is demonstrating Large Language and Vision Assistant (LLaVA), a 7+ billion parameter LMM that can accept multiple types of data inputs, including text and images, and generate multi-turn conversations with an AI assistant about an image. This LMM runs at a responsive token rate on device, which results in enhanced privacy, reliability, personalization, and cost. LMMs with language understanding and visual comprehension enable many use cases, such as identifying and discussing complex visual patterns, objects, and scenes.
Qualcomm AI Research is highlighting its first demonstration of Low Rank Adaptation (LoRA) on an Android smartphone: Running Stable Diffusion with LoRA, users can create high-quality custom images based on personal or artistic preferences. LoRA reduces the number of trainable parameters of AI models, enabling greater efficiency, scalability, and customization of on-device generative AI use cases. Beyond enabling fine-tuned language vision models (LVMs) for different artistic styles, LoRA is broadly applicable for customized AI models, such as large language models, to create tailored personal assistants, improved language translation, and more.
Qualcomm is showcasing a range of flagship commercial AI smartphones powered by its Snapdragon 8 Gen 3 Mobile Platform at MWC Barcelona, including the HONOR Magic6 Pro, the OPPO X7 Ultra, and the Xiaomi 14 Pro. These devices include impressive new generative AI feature like AI-generated image expansion (Xiaomi), AI-powered video creation and AI-powered calendar creation (HONOR), and image object eraser (OPPO), which together establish a pretty solid baseline for everyday generative AI applications on smartphones.
PC
Qualcomm’s accelerating focus on PCs is also on display at MWC Barcelona with the new Snapdragon X Elite PC platform and its impressive 45 TOPS NPU, built for on-device AI. Using GIMP (the popular and free image editor) with a Stable Diffusion plug-in, Qualcomm Technologies is showing that on-device generative AI can create an image on a PC in as little as 7 seconds – 3x faster than x86 competitors. I have been waiting for a gauntlet moment in the emerging AIPC arms race… and this feels like it.
Qualcomm AI Research is showcasing a world’s first on-device demonstration of a 7+ billion parameter LMM On a Windows PC that can accept text and audio inputs (like music, bird calls, etc.) and then generate multi-turn conversations about the audio.
Automotive
Qualcomm is also leveraging its industry-leading AI hardware and software solutions in the automotive space, applying traditional and generative AI capabilities to the Snapdragon Digital Chassis Platform. While the company looks to” deliver more powerful, efficient, private, safer, and personalized experiences for drivers and passengers,” creative and differentiated implementations outside of privacy, safety and personalization have thus far been a little thin. To be fair, this is still very early in the automotive design cycle for in-vehicle generative AI to find powerful use cases, so it may take a year or two to begin seeing real use cases for this. (There isn’t necessarily a lot of value in asking the car to generate images of dogs on skateboards while trying to find a parking spot or a restaurant, but there could be value in generating customized music to match a mood, instantly building out themed itineraries for a date night, or generating a customized city tour to make the most of a day trip).
IoT
In the consumer IoT space, Qualcomm is showcasing Humane’s AI Pin, which, if you didn’t know, runs on a Snapdragon platform. The device offers users the ability to take AI with them everywhere in an entirely new, conversational, and screenless form factor.
Modem-RF
Qualcomm is also showcasing its new Snapdragon X80 Modem-RF System, which integrates a second-generation 5G AI processor to enhance cellular performance, coverage, latency, and
power efficiency. Look for separate coverage for details about the solution.) Qualcomm is also introducing its new FastConnect 7900 Mobile Connectivity System, the first AI-optimized Wi-Fi 7 system, which leverages AI to radically improve adaptable, high-performance, low latency, and low-power local wireless connectivity. (Look for separate coverage of this connectivity solution as well.)
Infrastructure
Finally, on the infrastructure side of the business, since this is MWC Barcelona after all, Qualcomm is showcasing three groundbreaking AI-based enhancements for network management at the show, including a generative AI assistant for radio access network (RAN) engineers to simplify network and slice management tasks, an AI-based open RAN application (rApp) that reduces network energy consumption, and an AI-based 5G network slice lifecycle management suite.
In addition to being aware of these developments, it is important to consider the impact that these increasingly powerful on-device AI capabilities will have on device refresh cycles, particularly phones and PCs, both as a reset and an acceleration vector. Look for a more thoughtful analysis of this topic very soon.
Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.
Other Insights from The Futurum Group:
On-Device AI, Part 2 | The AI Moment, Episode 6
Qualcomm Raises Bar for On-Device Generative AI at Snapdragon Summit
Qualcomm Snapdragon 8 Gen 3 Brings Generative AI to Smartphones
Author Information
Olivier Blanchard is Research Director, Intelligent Devices. He covers edge semiconductors and intelligent AI-capable devices for Futurum. In addition to having co-authored several books about digital transformation and AI with Futurum Group CEO Daniel Newman, Blanchard brings considerable experience demystifying new and emerging technologies, advising clients on how best to future-proof their organizations, and helping maximize the positive impacts of technology disruption while mitigating their potentially negative effects. Follow his extended analysis on X and LinkedIn.
