Jesse Lyu's AI hardware startup is launching OS3, an agentic operating system available via desktop, Telegram, and iMessage, while defending the commercial legacy of the controversial R1 gadget. Despite scathing initial reviews and internal labor disputes, Rabbit is shifting its focus to multi-screen software as the broader tech industry rushes toward AI agents.
When Jesse Lyu and his team at Rabbit introduced the R1 at CES 2024, they positioned it at the vanguard of a hardware gold rush aimed at creating dedicated artificial intelligence gadgets. The premise was straightforward: users could speak directly into the orange-red, Teenage Engineering-designed device to orchestrate digital tasks, such as booking an Uber or ordering food through DoorDash. However, the commercial reality fell short of the visionary marketing. Early reviews were scathing - WIRED notably awarded the device a dismal 3/10 rating - largely because the agentic features struggled to execute reliably. The technology simply could not keep pace with the lofty ambitions of its creators.
Yet, the technological landscape has undergone a dramatic transformation in the months since that rocky debut. The capabilities of AI assistants have vastly improved, and agentic software that can perform complex digital workflows has become the dominant conversation across Silicon Valley. Recognizing this shift, Rabbit is attempting a major strategic reentry with the introduction of OS3. Described as a standalone "agentic operating system," OS3 is designed to operate seamlessly across multiple screens. Crucially, while it serves as a software update for existing R1 owners, users no longer need the physical hardware to access its capabilities; the system can now be utilized directly through a desktop browser, Telegram on a mobile device, or even iMessage.
This move marks a striking tactical pivot for an enterprise that initially staked its entire brand identity on the superiority of standalone handheld hardware over traditional smartphone applications. By decoupling its software from proprietary casing, Rabbit is adapting to an ecosystem that has rapidly ripened for versatile, cross-platform agents. Industry observers note that this pivot mirrors broader movements within the tech sector, where companies are increasingly realizing that users prefer integrated software experiences over carrying single-purpose hardware gadgets in an already crowded digital toolkit.
Despite the widespread critical backlash and the infamous 3/10 review score, founder Jesse Lyu firmly rejects the notion that the Rabbit R1 was a commercial failure. From a strictly financial perspective, Lyu maintains that the hardware rollout was tremendously successful, particularly given that the enterprise does not rely on a recurring subscription model for its AI features - a monetization strategy common among many of its industry competitors. According to internal figures cited by Lyu, the company successfully delivered more than 100,000 units of the R1 worldwide, maintained return rates of less than 5 percent, and secured a robust gross margin of roughly 45 to 50 percent on the hardware itself.
Reflecting on the engineering decisions that defined the R1's development, Lyu insists that the team did not make fundamentally incorrect strategic bets. Instead, he attributes the product's early operational friction to extreme resource limitations, pointing out that Rabbit operated with a fraction of the funding and manpower available to competing AI studios that raised hundreds of millions or billions of dollars. In total, Rabbit has raised approximately $60 million in venture capital and currently maintains a lean corporate structure supported by a dedicated team of just 15 employees.
This lean approach, however, brought severe operational vulnerabilities. Last year, the startup faced profound financial strain when a planned commercial launch in India was derailed by regulatory hurdles. The resulting cash-flow crunch became so severe that a handful of employees briefly went on strike, alleging delayed payroll disbursements. Lyu acknowledges these turbulent episodes with candor, stating that he respected the staff's position while the company worked aggressively to secure additional funding. Today, Lyu asserts that the organization is on significantly stronger financial footing, emphasizing that internal health metrics are stable and that the worst of the crisis has passed.
The broader race to dominate the AI agent market has intensified fiercely, thrusting Rabbit's core architectural approach back into the industry spotlight. Recently, Meta encountered significant friction when its new Muse agent was blocked from crawling Amazon's e-commerce platform to execute shopping transactions on behalf of users via a virtual machine. Bringing up this high-profile industry clash, Lyu offered a knowing smirk, highlighting that Rabbit utilized a similar virtual machine architecture to execute app actions a year and a half ago. This historical parallel prompts a lingering question in tech circles: was Rabbit simply too far ahead of the curve, deploying controversial infrastructure before the market and major platform holders were prepared to accommodate it?
As major technology conglomerates collide over web-scraping rights, API access, and automated browser control, smaller innovators like Rabbit must navigate an increasingly complex regulatory and platform-enforcement landscape. The debate over whether AI agents should be permitted to act autonomously inside third-party digital environments touches on critical questions regarding terms of service, cybersecurity, and consumer consent. While tech giants spar with regulators and web platforms over automated crawling, Rabbit is attempting to position OS3 as an agile alternative that can interface across multiple communication channels without relying exclusively on closed hardware ecosystems.
Furthermore, the competitive pressure from well-funded AI studios forces lean startups to constantly reevaluate their product-market fit. By shifting focus toward a flexible, multi-platform operating system, Rabbit is attempting to sidestep the capital-intensive hardware manufacturing cycle while still capitalizing on the surging consumer and enterprise demand for autonomous digital assistants that can streamline everyday workflows across desktop and mobile environments.
Rather than rushing headlong into the development of a successor hardware device following the R1's polarizing debut, Rabbit's leadership made a deliberate choice to focus engineering resources on continuous software refinement. Over the past year and a half, the company has deployed roughly 50 distinct software updates designed to polish the user experience and address core functional complaints. A critical engine behind this iterative process has been a highly active Discord community comprising more than 12,000 dedicated users who provide direct feedback, helping developers prioritize new features and integrations for the platform.
Significantly, this feedback loop has pushed Rabbit toward a much more open ecosystem philosophy. Users can now integrate third-party solutions such as OpenClaw or Hermes AI agents directly into the R1 hardware, moving away from a strictly closed, proprietary model. Emphasizing this cultural shift, Lyu notes that the device today bears little resemblance to its launch-day iteration, declaring that the hardware ultimately belongs to the consumer to configure as they see fit. During a recent demonstration, Lyu even showcased incoming web orders arriving for the R1 through Rabbit's proprietary website, proving that niche consumer demand persists even as retail channels like Amazon, Target, and Best Buy clear out remaining inventory.
Looking ahead, Rabbit is gradually winding down hardware manufacturing for the R1 as existing stock dwindles, cementing the transition toward OS3 as the cornerstone of the company's commercial future. By preserving the passionate enthusiast base cultivated through its hardware experiment while simultaneously expanding software accessibility to mainstream messaging platforms like Telegram and iMessage, Rabbit is attempting a difficult corporate reinvention. Whether OS3 can successfully capture market share in a crowded field dominated by tech behemoths remains to be seen, but Lyu's enterprise has definitively proven its resilience in surviving the volatile opening chapter of the consumer AI hardware boom.
Apple is reportedly in the early stages of developing a screenless, Whoop-like fitness tracker featuring a thin fabric band and a sensor-equipped computing module. The conceptual device, which faces a potential release window no earlier than 2028, signals Apple's growing strategic ambitions in continuous health monitoring and athletic analytics.
Apple is aggressively repositioning its Mac hardware lineup to challenge the hegemony of Microsoft and Nvidia in the generative AI space. By prioritizing localized processing power, Cupertino aims to slash enterprise AI costs while redefining the professional workstation market.
Microsoft is restructuring its gaming division by cutting hundreds of jobs across Washington state and consolidating major studios. In a sweeping corporate shift, iconic franchise 'Halo' is being moved under the operational umbrella of the Call of Duty publisher.
