Silicon Valley giants Amazon, Meta, OpenAI and other companies are all racing towards developing AI. “operating systems” for AI-powered devices—and 2026 is likely the year these efforts will start to take off. Devices are designed around the idea that AI agents will be able take actions for users, and not require them to use an application or website.
This sounds, in theory, like a harmonious relationship between technology and humans. This could completely change the business models of many consumer tech firms.
AI devices are promising that users will no longer have to tap through ads or upsells in apps. Instead, they can ask for a result and it will happen. An AI assistant will—in theory—book your travel, order you lunch, or restock your paper towels. The AI assistant should be able determine which services can provide what you want quickly and at a reasonable price.
AI agents can pose a nontrivial challenge. fairly unreliableAI-powered systems could also separate businesses from their customers. It could turn out to be a nightmare scenario for app developers.
Uber and DoorDash rely on their users staying in their app to serve them ads, offer other services and develop loyalty.
Artificial intelligence agents could reduce these companies to only their core offerings, and users may not even need to use the apps. If a customer pays the same for DoorDash to deliver their order through AI glasses but the company does not sell additional services, it could lose out.
Anjney Midha, an investor and board member at Sesame—the AI device startup cofounded by former Oculus leaders including Brendan Iribe—says that if companies don’t have “deep control over the supply of their product,” In a world in which they must reach out to users via an AI agent, it will be extremely difficult to do business.
In the past, creating an active developer platform was a crucial part of developing a successful OS. The companies must give developers good reasons to create apps on their platform. The relationship is not without its tensions. Apple, while historically giving developers an easy way to get their products in front of many users has also taken a 30% cut from all purchases made within apps.
Some tech firms are reluctant to allow AI agents created by smaller startups to stand in the way of their customers.
Rabbit was the start-up behind the R1 gadget that made its debut at CES in 2024. negative reviewsThis wall has already been reached. Jesse Lyu of the CEO Jesse Lyu said to WIRED, major app developers including Uber didn’t want to collaborate with the company in the beginning. Instead they refused to allow access to the API that could have allowed the R1 device to directly hail rides. Rabbit developed workarounds that allowed its R1 device to tap into apps such as Uber without formal permission.
“You have to understand why they’re not super happy: They sell fucking advertisements. That’s where many of them make money,” He said this in reference to the broader tech industry. “They decided we were too small, and it wasn’t worth it for them to work with us.”
This issue has also been experienced by Perplexity with its early versions of the shopping agent. In November. Amazon sued Perplexity Over an agent that it offered to purchase items for users. Amazon is investing heavily in its shopping agents and demanded Perplexity to stop scraping their site. It was then nearly impossible for Perplexity’s agent’s ability to make purchases on Amazon.
However, other app and marketplace developers are willing to partner with the larger platforms. ChatGPT has signed agreements with companies like DoorDash and Instacart to develop early AI applications, while Ticketmaster and Uber have launched early AI features on Alexa+. Many tech companies have shown cautious optimism about AI’s potential to create novel apps.
