OpenAI and Jony Ive are wrestling with unexpected challenges in their secretive AI device project, according to new Financial Times reporting. The palm-sized gadget won't launch until late 2026 at earliest as the team grapples with fundamental questions about AI personality and whether the company has enough computing power to make it work.
OpenAI just hit a wall that no one saw coming. The company's highly anticipated AI device, developed with former Apple design chief Jony Ive, is stuck in what can only be described as an existential crisis about personality. "The concept is that you should have a friend who's a computer who isn't your weird AI girlfriend," one source briefed on the plans told the Financial Times. The quote reveals just how deeply the team is thinking about the fundamental nature of human-AI interaction - and how much they're struggling to get it right. This isn't just about picking a voice actor or tweaking some dialogue. OpenAI is grappling with core questions about when an AI should speak, when it should stay quiet, and how to make it feel genuinely helpful rather than creepy or intrusive. The device itself sounds deceptively simple - a palm-sized gadget without a screen that communicates through microphones, speakers, and cameras. But simplicity in hardware often means complexity in software, and that's exactly what's happening here. The team wants to create something "accessible but not intrusive," according to FT sources, aiming to beat Apple's Siri at its own game. But beating Siri might be the least of their problems. The real challenge is infrastructure, and it's massive. "Amazon has the compute for an Alexa, so does Google [for its Home device], but OpenAI is struggling to get enough compute for ChatGPT, let alone an AI device," a source close to Ive revealed to the Financial Times. This admission exposes a critical weakness in OpenAI's strategy. While the company dominates headlines with ChatGPT advances, it's operating with far less computational muscle than the tech giants it's trying to challenge. The "always on" approach they're pursuing would require constant processing power to analyze conversations, understand context, and respond naturally throughout the day. That's a completely different computational load than Alexa, which only springs to life when someone says "Hey Alexa." The privacy implications are staggering. An always-listening device that processes conversations continuously raises questions that neither nor Ive have fully answered yet. The confirms privacy issues are among the "number of technical challenges" the team must overcome before launch. This puts in direct competition with established players who've already navigated these waters. has spent years refining Alexa's privacy controls and building trust with users. has the infrastructure and experience from Google Assistant and Nest devices. has positioned Siri as privacy-first, processing many requests on-device rather than in the cloud. The timeline reveals how serious these challenges really are. Initial reports suggested the device might appear sooner, but the pushes the launch window to late 2026 or early 2027. That's a significant delay for a partnership that was first reported in September 2023. The device represents the first in what's planned as "a family of devices" from the -Ive partnership. But if they can't solve the personality and infrastructure problems for device number one, the entire product roadmap could be at risk. Industry insiders are watching closely because this project could reshape how we interact with AI. But success depends on solving problems that go far beyond design and into the fundamental architecture of AI systems.