The pace of innovation and development in generative artificial intelligence (A.I.) is unprecedented. And this is just the beginning. How, where, and why we will use this technology will be broadand beyond our imagination. Early estimates value the generative A.I. market at $1 trillion.
Not only does this technology have the potential to change how we search for and create content, but it can also improve our daily lives. With generative A.I., your smartphone can become a true digital assistant, enabling you to communicate naturally and receive thoughtful answers. PC users can use it to read or compose emails, draft documents, and create presentations automatically. In vehicles, a conversational in-car assistant can provide recommendations for recharging the vehicle, purchasing a parking permit, or ordering dinner on the way home. Store A.I. kiosks and smart carts will assist shoppers by making menus with recipes using weekly specials, budget constraints, and family preferences.
To realize its full potential and address growing demand, generative A.I. needs both the cloud and billions of connected devices capable of high-performance A.I. computing at low power, such as smartphones, PCs, and vehicles. Enter hybrid A.I. A hybrid A.I. computing architecture distributes processing among the cloud and devices to optimize efficiency and improve the overall user experience.
Generative A.I. processing can run directly on the device, be sent to the cloud as needed, or a combination of bothall seamlessly to the user.
Users will expect a similar experience to a traditional search, which can display results in under one second. Meeting these expectations would be too costly to scale using cloud-based processing, especially during peak demand, with guaranteed quality of service.
Data centers are energy intensive and expensive. The cost for each generative A.I. web search query is estimated to be 10 times higher than using traditional search. With more than 10 billion queries per day, the incremental cost could be in the billions of dollars annuallyand web search is just one of the many ways this technology will transform multiple industries.
Beyond cost, running all the inference processing in the cloud presents challenges to privacy, reliability, and performance. When a request goes to the cloud, data leaves the device, resulting in potential security concerns. In fact, regulatory and compliance issues have already resulted in models being disabled, at least temporarily, due to the collection and storage of personal data.
Hybrid A.I. is inevitable. As we continue to find new ways of using generative A.I., the demand on cloud infrastructure will explode. Hybrid A.I. processing is the next transition in computing, just as we saw the evolution from mainframes to desktops to todays mix of cloud and devices in our hands.
Harnessing the processing power of high-performance, low-power devices will allow generative A.I. to efficiently scale. The cloud and devices will work together to deliver next-generation experiences through powerful, efficient, and highly optimized A.I. capabilities.
Cristiano Amon is the CEO of Qualcomm.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
