Qualcomm CEO: AI agents will end apps as we know them

What happened?

In an interview with CNBC, Qualcomm CEO Cristiano Amon stated that artificial intelligence agents will become the new model of digital interaction, replacing traditional applications as the main interface. According to Amon, apps will not disappear completely, but AI agents will take over their role. Qualcomm plans to launch 40 new devices with integrated AI capabilities in the coming months, powered by its Snapdragon chips. The statement comes at a time when the industry is seeking alternatives to app saturation and betting on more proactive intelligent assistants.

This is not the first time the end of apps has been predicted. In 2016, Microsoft CEO Satya Nadella spoke of "conversations as a platform," and in 2023, OpenAI's Sam Altman suggested that chatbots could replace graphical interfaces. However, Amon's statement carries special weight because Qualcomm is the leading manufacturer of chips for Android phones, with a market share close to 40% in the high-end segment according to IDC. His vision is not just theoretical: the company is already investing in specific hardware for AI, such as the NPU (Neural Processing Unit) in its Snapdragon 8 Gen 3 and future models.

Why is it important?

Amon's vision reflects a growing trend toward conversational and autonomous interfaces. If AI agents replace apps, the business model for developers, app stores, and platforms like iOS and Android could be radically transformed. Apple, for example, generated more than $85 billion in services revenue in 2025, a significant portion coming from the App Store. If AI agents reduce the need to download apps, that revenue stream could be affected.

Qualcomm, as a key chip supplier for mobile devices, is positioning itself to lead this transition by offering hardware optimized to run AI models locally. This could reduce reliance on the cloud, improve privacy, and enable smoother experiences. According to a Counterpoint Research report, the market for chips with integrated AI will grow 25% annually until 2028. However, mass adoption of AI agents faces technical challenges, such as the need for smaller, more efficient models, and user trust, as many are wary of assistants accessing their personal data.

Consequences for the industry

• Developers: They will need to adapt their services to be consumed through AI agents, rather than traditional graphical interfaces. This could involve creating specific APIs and optimizing content for conversational responses. According to a Gartner study, by 2027, 40% of interactions with applications will be mediated by AI agents.
• Device manufacturers: They will compete to integrate more powerful AI assistants, differentiating themselves by the ability to run advanced models locally. Samsung has already launched its Galaxy AI with on-device AI features, while Apple is developing its own generative model for iOS 19. Qualcomm, by supplying chips to multiple manufacturers, could standardize the AI experience on Android.
• Users: They will experience more natural and predictive interaction, but with potential privacy risks and dependence on a closed ecosystem. For example, if an AI agent manages your appointments, messages, and purchases, it could be difficult to switch platforms without losing functionality. Additionally, data collection to personalize agents could increase exposure to cyberattacks.
• Competitors: Apple, Google, and Samsung are already working on their own AI agents. Google has integrated Gemini into Android, and Apple has announced improvements to Siri with Apple Intelligence. Qualcomm seeks to get ahead with its chip infrastructure, but the battle is not only technical: it is also about user trust and integration with existing services.

What should readers know?

The transition to AI agents will not be immediate. Qualcomm expects the 40 AI devices to arrive throughout 2026, but widespread adoption could take years. Historically, predictions of interface replacement have been premature: in 2010, it was said that web apps would kill native apps, and in 2015, that chatbots would end apps. However, the convergence of advances in language models (such as GPT-4 and Gemini) and specialized hardware makes this time more plausible.

Users should be aware that AI agents will require access to personal data to function effectively, raising questions about privacy. For example, an agent that books restaurants will need to know your location, preferences, and payment history. Qualcomm has noted that its on-device AI approach minimizes data transmission to the cloud, but does not eliminate it entirely. Furthermore, reliance on a single chip supplier could create interoperability issues: if Qualcomm dominates the market, device manufacturers may have fewer options to differentiate themselves.

Finally, Amon's prediction should be taken with caution: while it is a plausible vision, the consumer market is unpredictable and apps have proven highly resilient. In 2025, more than 250 billion apps were downloaded worldwide, according to Statista. AI agents could complement, rather than replace, apps in the short term. However, the direction is clear: digital interaction is becoming more conversational and autonomous, and Qualcomm wants to be the engine of that change.

"AI agents will not only change how we use devices, but will redefine the app economy," says Amon.

In summary, Amon's statement is a milestone in the AI narrative, but its realization will depend on adoption by developers, manufacturers, and users. Time will tell whether apps become a digital relic or whether AI agents integrate as an additional layer in our technological ecosystem.