Lost Password


Qualcomm Wants to Introduce Chips Powered with Generative AI for Next Gen of Mobile Phones


“Next year, c will introduce top-tier chips with the power of generative AI inside premium phones.”

You have seen generative AI in the form of ChatGPT and Midjourney (a generative AI program for generating images from natural language descriptions called “prompts”), but these software products have disrupted industries, and they were limited to browser windows on desktops. Next year, the use of generative AI is going much better than what you are using on desktops.

What is the news?

Next year, Qualcomm will introduce top-tier chips with the power of generative AI inside premium phones. The company wants its new-age Snapdragon chips to use artificial intelligence for more than just enhancing the quality of camera shots.

Smartphones have been using AI for years to enhance photos and improve autocorrect, but generative AI software could do much better to improve the mobile experience. At the upcoming Qualcomm Summit in Hawaii this October, the company will unveil its latest premium chips that will feature generative AI technology.

Other Advantages of Qualcomm’s Generative AI-

There are various benefits to utilizing on-device generative AI. The most significant advantage is the assurance that queries and personal data are kept confidential and not transmitted through a remote server. Additionally, local AI processing is faster than relying on cloud computation, and it can function even when traveling on airplanes or in areas without cell service.

Also read: Kryndryl Extends Partnership with Cisco to Help Consumers Show Resilience to Cyber Attacks

Using an on-device solution is a smart choice for both business and efficiency reasons. The complexity of machine learning models has made it more expensive to operate servers that respond to queries. OpenAI was estimated to spend about $700,000 per day in April to get ChatGPT to respond to prompts. This prediction was based on the older GPT-3 model, and the newer GPT-4 model is even more complex and likely to be even more expensive to maintain at scale. Rather than requiring an entire server farm, Qualcomm’s solution is to use the device’s existing silicon brain to do all the necessary thinking at no extra cost.

Share This Post

Like This Post


Related Posts


    Leave a Reply

    Your email address will not be published. Required fields are marked *

    Thanks for submitting your comment!