Qualcomm's Snapdragon Tech Summit took place on the Big Island of Hawaii from the 30th of November to the 2nd of December. The central theme of the event was the reimagining of Qualcomm under its new CEO, Cristiano Amon. The cornerstone of this reinvention is Snapdragon which represents Cristiano’s vision for Qualcomm’s future and its strategy for expanding its business by pursuing and catalyzing new market opportunities where the emerging edge cloud and mobile computing meet.
The One Roadmap that is the Future of Qualcomm - Snapdragon
Snapdragon is no longer just a brand or a line of chips that powers our smartphones. It represents a single technology roadmap, a point that was brought home in Qualcomm executive keynotes in Hawaii and Investor Day in New York. Moreover, it is a strategy driven by Qualcomm’s premium technologies that will be infused into the current and future generations of Snapdragon chips. Quite simply, Snapdragon is about technological leadership enabling the company’s bold aspiration to diversify beyond the smartphone.
Technological Leadership – The summit focused on premium mobile experiences. Fundamentally that means leading edge technologies and features that advance the Android smartphone ecosystem. The new Snapdragon 8 Gen 1 SoC continues to deliver a bevy of catalytic feature and capability enhancements typical of every new iteration of SoC. Most notable this year were the computational photography and video features. The 8 Gen 1 allows handset OEMs to deliver 8K HDR video and provides the ability to synthesize up to 30 images with each press of the shutter for stunning HDR photos thanks to three 18-bit ISPs.
In addition to the chips, AI is an increasingly important pillar of technological leadership for Qualcomm. According to the Snapdragon team, today’s premium smartphones run over 100 machine learning (ML) models. The number is growing as Snapdragon continues to deepen its application of AI across many smartphone features ranging from computational photography to antennae tuning. Increasingly these models run concurrently and in larger numbers.
To meet the rapidly increasing demand for and complexity of AI processing, the Snapdragon team announced dramatic improvements to silicon and architecture to keep their OEM customers competitive in world of AI computing. Snapdragon announced its seventh generation AI Engine that leverages the 8 Gen 1’s heterogeneous compute architecture and expanded shared memory to more efficiently (70% better performance per watt) handle AI workloads with 4 times the performance from the previous generation 888.
Snapdragon also announced important enhancements to their developer tools for AI and neural processing SDK, namely their partnership and integration with Google Cloud’s Neural Architecture Search (NAS). The Vertex AI NAS uses AI to select the optimal ML models for an OEM’s specific device configuration. This not only speeds up the DevTest process for tuning ML models for the hardware they will run on, but also improves the performance of models across the diverse Android OEM ecosystem that uses Snapdragon chips. This same tooling can be used to optimize ML models for Snapdragon platforms for automotive, XR, PC, IoT, and more.
End Market Expansion – While the smartphone remains important, Qualcomm intends to leverage its single technology roadmap to enter new end markets, most notably automotive and the expansive and diverse market opportunities of consumer and industrial IoT. The idea is to bring advanced semiconductor technologies, image and audio processing, and AI compute and Edge AI to frontier markets that will expand the company’s revenue growth opportunities.
On the premium side of its market diversification ambitions, Snapdragon announced the G3x Gen 1 gaming platform which applies the company’s prowess in mobile gaming to the handheld gaming form factor. Snapdragon also announced the 8cx Gen 3 SoC for portable computing devices that promises to supercharge Qualcomm’s efforts to bring Windows-on-Arm to laptops and, dare we say, desktops to give x86 a run for their money in the personal computing market.
Undoubtedly, the premium Snapdragon features and technologies will provide technological advantages for Qualcomm to pursue new IoT opportunities with feature differentiations and compute efficiencies that drive down the cost of enablement. Snapdragon’s camera, audio, and AI chops will prove instrumental in Qualcomm’s foray into smart cities and intelligent transportation infrastructure where economical digitization capabilities are needed in order to scale.
Uplink the Great Hope for mmWave?
I was personally and professionally excited to witness Qualcomm CEO Cristiano Amon stand up on stage declaring that "Uplink is just as important." Why? For one, uplink will be important if 5G is to make an impact on the industrial IoT and deliver on its promise to revolutionize industries. One thing is certain, industrial 5G will not be about streaming 8K video to a PLC (Programmable Logic Controller). Quite the opposite.
Yet, the mobile wireless industry doesn’t talk about uplink speeds much. It has been fixated on downlink speeds touting how quickly you can download a 4K movie rather than how you can upload high-quality live videos to your favorite social platform or send massive data streams coming from hundreds if not thousands of PLCs on a factory floor to an edge cloud node for aggregation and analytics. Unfortunately, uplink is hard especially for smartphones.
In his Day 1 keynote, Cristano announced that Qualcomm achieved 3.5 Gbps uplink speeds using a X65-equipped test device and carrier aggregation over a 100 MHz sub-6 and 300 MHz of mmWave frequency. Sure, the results were from a lab test under lab conditions. Regardless, Qualcomm continues to rapidly advance uplink performance on top of what the company achieved with Verizon and Samsung earlier in October of this year. The trio reached uplink speeds of 711 Mbps in a lab trial that aggregated carriers across 400 MHz of mmWave frequency and 20 MHz of 4G frequency.
The big deal here is the shift in mindset, especially by Qualcomm. The application of mmWave to boost uplink performance could be a boon for US operators who invested billions in mmWave early on only to find a limited number of applications for new 5G services and revenue opportunities. According to Cristiano Amon, uplink will be vital for operators in bringing about new experiences and possibilities for innovation. I couldn’t agree more.
Influencing the Future of Computing
Amon’s vision for Qualcomm’s role in the future of computing is what he calls the Connected Intelligent Edge. As Cristano sees it, Qualcomm’s mission is to make billions of endpoint devices that populate the edge more intelligent and “always connected” to the cloud. In support of this mission, Snapdragon’s growing portfolio of AI-infused compute platforms and Qualcomm’s leadership in mobile wireless technologies such as 5G put the company in pole position to drive the evolution of distributed computing from the edge.
In New York and in Hawaii, Cristiano made it a point to emphasize that, “If you believe in the growth of cloud, then you believe in the Qualcomm growth plan.” The Connected Intelligent Edge is a weighty counterpoint to the traditional notion of cloud – central public cloud computing. Consider the massive impact that mobile computing has had on how we think of computing even today. While cloud computing has had a tremendous impact on the business of IT, it has yet to achieve the transformational impact that mobile computing has had in transforming industries and the way we work, play, and live.
The fact is cloud computing is changing, especially in the last three years. As we saw at AWS’s re:Invent 2021 conference which took place at the same time as the Snapdragon Tech Summit, hyperscale cloud players are increasingly expanding their cloud footprint out toward the “edge.” According to Werner Vogels, CTO of AWS, this shift of cloud computing toward edge computing is coming about by “the need to solve the problem of the speed of light.”
Why? Because the edge, more specifically edge devices, is where the data is being generated AND consumed, whether it is video, telemetry, photos, or documents. And if Qualcomm can help it, most processing will be done on device whether the application workload is AI inference for anomaly detection or a video game.
We can expect the convergence of edge cloud and mobile computing to be the spawning ground for applications and innovations that were not possible before. As the hyperscalers push to extend the cloud out to the “edge,” Qualcomm is uniquely positioned to meet them by wirelessly connecting the vast and rapidly growing universe of intelligent endpoint devices to the emerging edge cloud thus changing the way we design distributed applications and computing systems.
The Connected Intelligent Edge will work out great for Qualcomm if it is able to execute on its Snapdragon strategy because the edge is a diverse frontier beyond the smartphone. That is exactly where Qualcomm boldly intends to go.
Leonard Lee is the founder and managing director of neXt Curve, a research advisory firm focused on Information and Communication industry and technology research. He has worked as an executive consultant and industry analyst at Gartner, IBM, PwC and EY and has advised leading companies globally on competitive strategy, product and service innovation and business transformation. Follow Leonard on LinkedIn: linkedin.com/in/leonard-lee-nextcurve