Apple's new CEO John Ternus wants the company to seriously pursue the robotics business, but judging by the development progress of this planned first-generation consumer product, he may be facing an uphill battle.
The device features a 9-inch screen and is equipped with a movable robotic arm, resembling the desk lamp character from Pixar animation. Apple launched this robot project as early as 2019, internally codenamed J595. In its early stages, the project had a conservative positioning, mainly envisioned for video calls; the robotic arm's role was to follow the user around the room. But as management sought to layer AI capabilities onto it, the project's goals expanded dramatically. Under the current design plan, this robot is equipped with cameras and various sensors, capable of fully perceiving its surroundings and acting autonomously: when someone enters a room, it can turn on the lights, or even try to shoo a cat off the table; it also has built-in obstacle avoidance to prevent bumping into surrounding objects. People involved in the project revealed that most of the robot's hardware design was finalized about two years ago, but the software team has struggled to keep pace. The hardware team responsible for J595 wants to accelerate the timeline and release the product as soon as possible, while the software team advocates delaying the launch. Over the past year, the software team has been trimming the originally planned features, intending to narrow the product's core capabilities to FaceTime calls and simple home automation tasks like turning on lights. Apple plans to price this robot at around $1,000, though the final price could be higher due to component market conditions. Multiple former employees said it may be named Home View Ultra, following the naming convention of the stationary desktop device Home View that Apple released last month. Apple originally hoped to launch it next year, but the release could also be pushed back to 2028 (the hardware team initially planned to launch it this year).
The development difficulties of this robot reflect the challenges Apple faces in Ternus's early tenure. Apple excels at iterative annual updates to the iPhone and Mac, but is not adept at developing entirely new product categories from scratch, as the company has traditionally maintained a slow and steady pace. For example, the Vision Pro headset was refined over roughly 8 years before its official launch two years ago, and market reception was lukewarm. If J595 arrives next year, its development cycle would be close to that length. Amid the AI wave, Apple's pace is especially sluggish. The major overhaul of Siri took years, and even with new AI features launching this year, it still lags behind competitors like Meta's Muse and OpenAI's Dots in the AI agent race. Former CEO Tim Cook delayed the release of multiple new products to align with Ternus's September transition, syncing product debut timing with his arrival. Since then, Apple has successively released Home View, the foldable iPhone Duo, and the touchscreen MacBook, but market reviews have been mixed: the $2,899 MacBook has been criticized for being overpriced; the Duo generated strong buzz but has seen frequent complaints about cracked screens; and Home View was criticized as merely a larger version of the Apple Watch, heavily dependent on the still-underwhelming Siri.
When Ternus took over, a host of formidable competitors were already positioning themselves in AI hardware and robotics. A year ago, OpenAI acquired a hardware startup co-founded by former Apple chief design officer Jony Ive. Since then, Ive has devoted himself to building OpenAI's AI hardware product line, and OpenAI has poached more than 400 Apple engineers for related projects. OpenAI plans to launch a donut-shaped smart speaker next year, featuring a movable structure and camera to perceive users and their surroundings. Meta has also been rolling out hardware intensively, including AI glasses and VR headsets. Compared to Apple, some of Meta's projects move from concept to reality very quickly: its pocket AI device Charm reportedly took less than a year to develop and is expected to ship in December. Tesla Motors (TSLA) is developing the humanoid robot Optimus, which Elon Musk says will help with household chores. A large number of startups have also joined the competition, such as robotics company 1X, which originally planned to bring humanoid robots into homes this year. Its Neo robot went viral immediately after being unveiled last year, with videos showing it washing dishes and folding clothes. In addition to J595, Apple's robotics team, led by veteran executive Kevin Lynch, has explored other form factors: wheeled follow-along robots and indoor flying robots. Apple also has a project developing robots designed for specific daily chores, such as a clothes-folding robot. In theory, a robot focused on only a few fixed chores is far less difficult to develop than a general-purpose humanoid robot that must handle diverse tasks. Meeting records reviewed by insiders show that Ternus has a strong interest in the chore-specific robot direction. A former engineer who worked on J595 put it this way: Apple has been "torn apart trying to catch the AI wave."
An iPad on a stick
J595 has been in development for nearly a decade, passing through the hands of multiple Apple executives, many of whom simultaneously managed multiple projects and could hardly devote themselves fully to this device. Bloomberg had previously disclosed some information about this robot. The earliest prototype was completed in 2018. The following year, Cook observed a prototype demonstration and approved the project to proceed. Early prototypes were larger than the current design, and some Apple engineers joked that it looked like "an iPad stuck on a stick." The robot hardware team was led by Matthew Costello, who previously oversaw Beats (the headphone brand Apple acquired in 2014); software development was assigned to Apple's Special Projects Group, which at the time was primarily advancing Project Titan — Apple's autonomous vehicle program. For a long period, senior engineer Brian Lynch (a veteran of the iPod era) was simultaneously responsible for the car cockpit and J595: he envisioned the robot not only for home use but also mounted on the center console of an Apple car. In 2021, the leadership of the Special Projects Group changed: Doug Field departed for Ford, and Kevin Lynch took over. Lynch was known for developing health tracking features for the Apple Watch and saving that product line, but at the time most of his energy was on the autonomous vehicle, with the robotics business mainly handled by Chris Sari, who had led Apple Watch software development. From the early stages, project members were clear that to justify a high price, it could not just be a fancy FaceTime calling device. After ChatGPT's emergence in 2022 sparked the AI boom, the team began adding AI capabilities to create more complex experiences. Internally, the team hoped to incorporate a large number of proactive agent features, allowing the device to autonomously complete diverse tasks. The robotics division even established a dedicated agentic AI research branch led by Vladan Kortuun; he had previously managed AI business at Intel and joined Apple in 2021. But Apple's self-developed AI capabilities lag noticeably behind competitors. As a result, the team was permitted to use externally available models, including open-source models and Google-licensed models. The team originally planned to later integrate J595 with Siri, but in the early days of robot development, the Siri team and the robot team operated independently, and the Siri team was already under enormous iteration pressure. After Apple halted its autonomous vehicle project in 2024, Lynch shifted his full focus to home robots, and a large number of former car project employees were reassigned. Three people involved in the project said that many employees who had been working on robotics earlier were disappointed with his management style, believing he rarely communicated and seldom listened to frontline opinions. Lynch chose to lower the feature expectations for the first-generation J595: the plan was to launch a basic version first and gradually add capabilities through updates, similar to the approach with the Apple Watch. The first-generation Apple Watch launched in 2015, and health and fitness sensors were added later, which is what truly made the wearable device popular.
Ternus takes the helm
Before being promoted to CEO, Ternus, then head of hardware, was skeptical of many new projects, such as the defunct autonomous vehicle project and the Vision Pro headset. But three people familiar with the robotics project said he holds higher expectations for Apple's robotics ambitions. Last year, Kevin Lynch and the Special Projects Group's reporting line was shifted from senior vice president John Giannandrea to Ternus (Giannandrea had overseen Apple's AI business and departed earlier this year). In April, Apple officially announced that Ternus would succeed as CEO within the year. Even during the transition preparation stage, he attended almost every J595 and other robotics demonstration and product planning meeting. People familiar with the matter said Ternus is promoting a company-wide strategy: streamlining teams and accelerating execution. The robotics team remains relatively small, and Apple is scarce in professionals deeply versed in robotics, with some estimating that there are only about 10 true robotics experts. Lynch's software team numbers around 300, of which about 30 are responsible for research. J595 project members are already discussing a second-generation model, planning to carry over some features cut from this generation to the next. Meanwhile, a group of early engineers on the project has already departed: hardware lead Brian Lynch went to fitness hardware company Oura, and robotics research lead Zhang Jian jumped to Meta. Some J595 team members want the product to launch as soon as possible; others worry it will suffer the same cold market reception as Home View. But they are concerned that the later Apple delays the release, the more the industry's AI capabilities will continue to advance, and the higher external expectations for this robot will become.
Comments