Jony Ive and OpenAI's secret AI device: Questions and answers on ambition, reality, and prospects

After the Humane & Rabbit disaster: Is the hyped OpenAI device also facing a major flop?

The vision is monumental: Jony Ive, the legendary designer behind Apple's greatest successes, and OpenAI, the powerhouse of the AI ​​revolution, are collaborating on a device that will usher in nothing less than the post-smartphone era. At its core is a screenless, intelligent assistant that constantly perceives its surroundings via cameras and microphones and proactively assists in everyday life—a seamless connection between artificial intelligence and the physical world, known as ambient computing.

But behind the shiny facade, things are crumbling. Recent reports paint a picture of a project hampered by fundamental hurdles, whose success is anything but certain. The recent, spectacular failures of competing products like the Humane AI Pin and the Rabbit R1 cast a long shadow over the entire device category and demonstrate how rocky the road is beyond the established app ecosystem. From unresolved design issues to technical limitations in processing power and battery life to the formidable ethical and privacy concerns of an "always-on" device—the list of challenges is long.

Humane was a startup that developed the Humane AI Pin, a small, wearable device designed to display information via laser projection. It was advertised as an innovative, AI-powered assistant, but it suffered from technical issues, slow processing, low user relevance, and high costs ($700 plus $24 monthly). Humane discontinued the product at the end of February 2025, sold the remaining assets to HP, and the AI ​​Pin became nonfunctional. The company is considered a clear example of an ambitious but failed tech gadget.

Rabbit AI also released the Rabbit R1, an AI-powered wearable that initially received a lot of hype. Criticism was directed at the lack of features important to users, hardware issues, and a lack of a clear benefit or target audience. Despite criticism and declining interest, the Rabbit R1 is still being supported, with new features like a memory diary. However, there is a risk that Rabbit, too, will fail if it fails to improve its AI technology (Large Action Model) and define clear unique selling points and target audiences.

Both products—Humane AI Pin and Rabbit R1—are considered failures, primarily due to misplaced user needs, technical flaws, and a lack of market orientation. They serve as warning signs for OpenAI, which is launching its own hyped AI device for 2026 and faces the same challenges, such as balancing useful functionality, usability, and data protection.

Short description:

• Humane AI Pin: Wearable with laser projection for AI assistant, technically immature, expensive, discontinued, HP bought leftovers.
• Rabbit R1: AI wearable with voice assistant, weak features, hardware issues, still active, but at a crossroads.

The question of whether OpenAI could also fail with its new device is considered realistic due to the difficult market situation and the known problems with spoken AI gadgets.

This article examines the current status of this ambitious project. It answers the most important questions about its ambition and reality, explains the specific technical and conceptual hurdles, and offers an outlook on whether this grand vision can become a real product—or whether it is headed for the next high-profile AI flop.

Life without screens: Jony Ives and OpenAI's great AI vision is on the verge of collapse

The key finding upfront: The screenless AI device that OpenAI and Jony Ive are working on is going through a difficult phase. Technical hurdles, unresolved design decisions, computing capacity, and data protection issues are slowing down the schedule and raising fundamental market questions. Success is possible, but by no means guaranteed; the recent failures of other AI gadgets demonstrate how challenging the step beyond the smartphone is.

What is the OpenAI/Jony Ive project actually about?

It's about a new, screenless AI device roughly the size of a smartphone that perceives its surroundings via cameras, microphones, and speakers and interacts with users exclusively through voice, audio, and context. The goal is an "always-on" assistant that continuously collects sensor information without a traditional activation word, understands the situation, and proactively supports users in everyday life. The idea: ambient computing instead of app tiles, immediate interaction instead of touchscreens.

Why is the device relevant when smartphones can do everything?

Smartphones are generalist devices with app ecosystems, but interaction and context are fragmented. A specialized AI device could unify interaction, minimize response times, and use context—visual, acoustic, and spatial—as the primary signal. The promise is less cognitive load, more real-time assistance, more natural control, and a bridge to the next era of computing beyond the display.

Where is the development – ​​and what does Golem report?

Several, sometimes fundamental, problems are slowing progress. According to the report, the project is struggling with:

• unclear design decisions (form factor, interaction model, “always-on” behavior),
• technical limitations in computing power, energy efficiency and infrastructure,
• unresolved data protection and behavioral issues of the assistant,
• Dependencies in the supply chain and final assembly planning.

In addition, OpenAI would need to substantially expand its computing capacity to effectively operate a consumer product. Luxshare is named as the manufacturing partner, but final assembly could take place elsewhere.

Is “always-on” without an activation word technically and ethically realistic?

For technical reasons, such behavior requires extremely efficient, low-latency sensor processing, robust environmental analysis, and excellent on-device classification to prevent the cloud from being constantly flooded with raw data. From an ethical and data protection perspective, constant listening and viewing places high demands on transparency, consent, data minimization, edge processing, and access control. The balance between help and intrusiveness is considered a core problem—in other words: a helpful friend, not a "weird partner."

What specific technical hurdles are mentioned?

The biggest hurdles can be assigned to four areas:

• Computing infrastructure and LLM serving: Scalable inference for multimodal, context-rich answers with low latency is expensive and complex. OpenAI needs to expand its capabilities before a mass product can be reliably served.
• Energy and on-device AI: Always-on sensing, wake-word-free detection, and continuous context tracking require extremely efficient on-device models and hardware accelerators to meet battery life and thermal budgets.
• Data protection by design: Always-on without an activation word requires robust architectural decisions regarding edge processing, pseudonymization, local buffering, granular permissions, and user control interfaces.
• Form factor and UX: Without a screen, a clear, consistent interaction logic is needed, including voice, haptic feedback, and possibly projection or light signals. Market comparisons show that diffuse interaction leads to frustration and returns.

Why is computing power such a bottleneck?

The inference costs of modern multimodal models remain high despite cost reductions; latency requirements are increasing if devices are to continuously respond with context awareness. Scaling requires massive GPU/accelerator capacity, power supply, and a robust global latency architecture. At the same time, there is increasing pressure to offload functions to the device (edge/on-device AI) to improve privacy, latency, and cost—which in turn requires new hardware, model distillation, and compromises in quality.

What distinguishes the Ive/OpenAI vision from previous AI gadgets?

The goal: deeper integration of multimodality, context, and assistance, a "terminal for AI" instead of an app launcher. The implementation: a deliberately screenless approach with always-on sensing and ambient interaction, placing greater emphasis on elegant, unobtrusive presence. The focus is on merging hardware-software coherence—an approach that Ive has historically helped to shape. In contrast, the Humane AI Pin and Rabbit R1 suffered from heat, battery life, poor field performance, and returns.

What can we learn from the failures of Humane AI Pin and Rabbit R1?

Several things. First, simply transferring familiar smartphone functions to a new gadget isn't enough. Second, battery life, thermal stability, and real-world task performance matter more than video vision. Third, without superior service orchestration, voice control remains slower than app usage. Fourth, users won't tolerate constantly "listening" technology without tangible added value and transparent control. Fifth, return rates and loss of trust are destroying young categories.

What role does Luxshare play in manufacturing?

Luxshare is considered a core supplier in the Apple ecosystem and, according to a report, is a manufacturing partner for at least one OpenAI device. Other potential suppliers, such as Goertek, could contribute speaker modules, for example. Indications suggest a first series production will not begin before late 2026/early 2027. Luxshare has experience with wearables, AR optics integration, and highly automated assembly, which speaks for its quality and yield rates.

Is the market ready for screenless AI hardware?

The conditions are ambivalent. On the pro side: enormous AI diffusion, declining inference costs, advances in edge AI and sensor-rich hardware, and strong demand for simpler interaction. On the con side: habituation to app ecosystems, high privacy sensitivity, unresolved hallucinations, and robust alternatives through smartphone-based AI assistants (Apple, Google, Perplexity). Recent product flops point to an early, error-prone market phase without a clear "killer use case."

How do the major platforms position themselves?

• Apple is consolidating AI on the iPhone ecosystem with on-device pipelines, privacy-focused features, and deeply integrated services. The smartphone user base is strong, and user habits are stable; external gadgets, on the other hand, are struggling.
• Google is expanding Gemini to device classes like TVs, pushing AI capabilities into existing hardware categories; users interact without switching platforms.
• With the “AI Phone,” Deutsche Telekom and partners demonstrate that “app-free” can also function as a UX layer on the smartphone – with AI orchestration in the background.
• With Ive, OpenAI is focusing on new form factors beyond the display – a risky but potentially category-defining path.

Which product category is OpenAI targeting – smart speaker, wearable, desktop device?

Reports outline various options: an always-on desktop device, a smart speaker-like device without a display, a mobile assistant, or even a small product family. The sources describe prototype stages and an uncertain form factor; the common denominator is deep embedding in OpenAI's LLMs and context-aware interaction.

How could data protection be effectively addressed with always-on technology?

Practical ways would be:

• strict edge processing for detection and pre-filtering stages,
• local buffering and selective sending of minimized, pseudonymized features,
• granular operating modes (privacy zones, mute/cover/shutter hardware),
• visible indicators during active recording/transmission,
• Auditability and user control panels for data paths, deletions and releases,
• standardized AI risk monitoring (“always-on” defense for AI use on data).

A new dimension of digital transformation with 'Managed AI' (Artificial Intelligence) – Platform & B2B Solution | Xpert Consulting - Image: Xpert.Digital

Here you will learn how your company can implement customized AI solutions quickly, securely, and without high entry barriers.

A Managed AI Platform is your all-round, worry-free package for artificial intelligence. Instead of dealing with complex technology, expensive infrastructure, and lengthy development processes, you receive a turnkey solution tailored to your needs from a specialized partner – often within a few days.

The key benefits at a glance:

⚡ Fast implementation: From idea to operational application in days, not months. We deliver practical solutions that create immediate value.

🔒 Maximum data security: Your sensitive data remains with you. We guarantee secure and compliant processing without sharing data with third parties.

💸 No financial risk: You only pay for results. High upfront investments in hardware, software, or personnel are completely eliminated.

🎯 Focus on your core business: Concentrate on what you do best. We handle the entire technical implementation, operation, and maintenance of your AI solution.

📈 Future-proof & Scalable: Your AI grows with you. We ensure ongoing optimization and scalability, and flexibly adapt the models to new requirements.

More about it here:

• The Managed AI Solution - Industrial AI Services: The key to competitiveness in the services, industrial and mechanical engineering sectors

What schedule is realistic?

Public indications point to late 2026 to 2027 for the first hardware waves, assuming the design and infrastructure issues are resolved in 2025/2026. However, current reports indicate ongoing hurdles that could further slow the schedule.

Why is designing without a screen so difficult?

Because screens elegantly solve many problems: confirmation, contextual display, error handling, multitasking. Without a display, auditory and speech interfaces, haptic signals, and possibly projection/lighting must cover the entire interaction space. This requires outstanding speech recognition, dialogue management, error robustness, and time-based behavior. Furthermore, the pressure increases for the device to anticipate "correctly" before a question is asked—a significant hurdle.

What business risks lurk – and how can they be mitigated?

• Cost risk: High inference and hardware costs can drive up prices. Countermeasures: Edge AI, model distillation, layered service levels, partnerships, and service subsidies.
• Liability and trust: Always-on and hallucinations harbor misconduct. Antidotes: conservative defaults, secured functional domains, strict governance, and human-in-the-loop mechanisms.
• Market acceptance: Without clear added value, weak sales and returns are a risk. Antidotes: focused "jobs to be done," segmented target groups, pilot projects, and reliable usefulness over broader product range.
• Competition: Smartphone platforms are integrating AI more deeply. Antidotes: unique usage contexts, faster task resolution, and true ambient intelligence beyond app orchestration.

What does a credible “killer use case” look like?

Probably not a singular “killer”, but a threshold of functions that together noticeably facilitate interaction:

• Immediate, error-free task orchestration in everyday life (appointments, routes, reminders) without manual app steps, faster than with a smartphone.
• Visual daily living assistance: context-based recognition, explanations, assistance with household and work activities, with very low latency.
• Proactive safety/comfort in the room: intelligent notifications that only trigger when relevant, embedded in presence and activity context.
• Edge-first data sovereignty: clear, visible boundaries, local learning, and privacy design that generates trust.

What do independent reports say about the current problems?

Several German tech media outlets and industry portals consistently report on technical and conceptual hurdles: always-on behavior, data protection, computing power, unclear positioning between smart speaker and mobile assistant, difficult market launch, and potential delays. The core message: The demands are high, but the implementation is not yet mature.

How are other providers reacting to the “app-free” vision?

• Telekom: With the AI ​​phone and Perplexity integration, it demonstrates how app orchestration can work instead of app usage on the smartphone. This reduces the need to carry another device and leverages existing habits.
• Google: Distributes Gemini across categories such as TV and expands its presence in everyday life without a new device class.
• Apple: Drives on-device and system integration, bringing AI interactions seamlessly into existing hardware channels.

What role does ambient computing play as a framework?

Ambient computing precisely describes the goal: technology that fades into the background, helps proactively, and acts context-awarely – without dedicated inputs. The OpenAI/Ive device would be a vehicle for such interaction. But ambient intelligence requires mature sensor technology, reliable AI interpretation, edge processing, and an ethically sound data protection architecture. The path to this goal is evolutionary, not gradual.

What indicators point to progress by 2026/2027?

• Edge AI/NPUs are becoming faster, cheaper, and more energy-efficient; model compression and distilled multimodality are evolving rapidly.
• Supply chain partners like Luxshare have highly automated manufacturing, wearable/AR expertise, and yield experience.
• Cross-platform AI availability (Gemini, Copilot, Perplexity) increases user acceptance of conversational interaction; expectations for device intelligence are rising.

What warning signs remain?

• Persistent hallucinations and misbehavior in open contexts.
• Latency and reliability issues in real-world applications beyond demos.
• Data protection skepticism towards always-on sensors in private spaces.
• The “Why not just use a smartphone?” reflex when added value is not evident.

Could a product family make more sense than a single device?

Yes. A tiered portfolio—e.g., a stationary ambient device with room intelligence and a high-energy/performance profile, plus a mobile, power-saving companion device—could better cover different usage scenarios. The modular approach allows for clearer value propositions and technical optimizations for each environment.

Is cooperation with ecosystems inevitable?

Probably. For tasks like telephony, messaging, navigation, media, and smart home control, interfaces to operating systems, services, and devices are needed. Without deep integrations, the same hurdle looms as with Rabbit/Humane: slow, fragile orchestration of external services. Strategic partnerships, SDKs, and standardized agent APIs are crucial.

What does this mean for the European market and data protection requirements?

Europe sets high standards for data protection, consent, purpose limitation, and transparency. An always-on device would need to offer granular control mechanisms, local processing, and clear opt-in/opt-out paths. A differentiated EU product profile could become a competitive advantage – if it succeeds in building trust rather than limiting functionality.

How should a market entry be designed to avoid setbacks?

• Target group selection: Application domains with a clear “job to be done” and high tolerance for new interactions (e.g., home assistance, care, accessibility, knowledge work).
• Multi-stage pilots: controlled rollouts, quality telemetry, iterative tuning of edge/cloud components.
• Transparency offensive: functions, data paths, local processing, indicators and manual overrides.
• Service bundles: Added value beyond hardware – premium orchestration, knowledge graphs, offline modes, data sovereignty packages.

Will OpenAI create a new category with Ive – or build the smart speaker 2.0?

The spectrum ranges from a refined smart speaker without a display to a truly new terminal for AI. The outcome depends on practical performance: Can the device solve recurring everyday tasks faster, more discreetly, and more reliably than a smartphone plus assistant? Will a trustworthy always-on model succeed? If so, a new category is within reach. If not, the threat of "smart speaker, only more expensive" looms.

Which signals from the supply chain are trustworthy?

Reports about Luxshare as a manufacturer, possible Goertek involvement, and a late 2026/2027 timeline are repeated in several sources. Luxshare's expansion into AR/wearables, fully automated lines, and high yield rates are compatible with an ambitious but later rollout. Reliability increases when EVT/DVT/PVT milestones become publicly or indirectly visible.

How does the project differ from Telekom's "app-free" AI phone?

The Telekom model shifts the UX layer on the smartphone: The AI ​​assistant orchestrates apps in the background, while apps remain accessible. OpenAI/Ive want to remove app layers and the display in favor of ambient interaction. The Telekom model lowers barriers by preserving familiarity; the Ive/OpenAI model relies on a changed paradigm of interaction – higher risk, potentially greater leap.

Which industries could be early adopters?

• Smart home and comfort: discreet assistance, contextual automation, security.
• Care/Assistive Technologies: hands-free interaction, monitoring with strict privacy.
• Knowledge work/household: transcribing, visual and organizational support in real-life tasks.
• Teaching/training: situational feedback, explanations in the room.

What is the sober outlook for 2026/2027?

A realistic approach is a staggered market entry with initially limited quantities, focused usage scenarios, and conservative functional limits while edge AI and inference operations mature. A disruptive mass launch seems unlikely given the hurdles; however, a credible, useful launch with clear domains is achievable – provided data protection, latency, and reliability are convincing.

What hard no-gos should the project avoid?

• Overpromising from demo videos without robust practical performance.
• Diffuse interaction without clear states, signals and control options.
• Lack of transparency in always-on recording.
• Missing offline/degradation modes in case of loss of connectivity.
• Unclear update strategies for models and security fixes.

What should prospective customers pay attention to when prototypes are shown?

• Latency in real-world environments, not just in the lab.
• Battery life under true always-on usage.
• Error handling: What happens in case of misunderstandings?
• Privacy interface: displays, switches, logs, local processing.
• Reliable, reproducible daily jobs that are faster than smartphones.

What is the overall rating?

The project addresses a real gap: ambient intelligence that reduces complexity and makes context the primary interface. The combination of OpenAI's AI power and Ive's design expertise is extraordinary—but the hurdles are equally extraordinary. The industry has just learned that vision fails without everyday robustness. If it succeeds in building always-on responsibly, balancing edge/cloud sensibly, delivering clear added value, and ramping up the supply chain and infrastructure in a timely manner, the project has a realistic chance of becoming a new, credible device class. If this balance fails, the device will, at best, be a refined smart speaker—or join the list of prominent AI gadget failures.

The next 12-24 months will determine whether an ambitious concept will become a viable category – and whether ambient computing will become firmly established in the mass market for the first time beyond the smartphone.

☑️ Our business language is English or German

☑️ NEW: Correspondence in your national language!

Konrad Wolfenstein

I would be happy to serve you and my team as a personal advisor.

You can contact me by filling out the contact form or simply call me on +49 89 89 674 804 (Munich) . My email address is: wolfenstein ∂ xpert.digital

I'm looking forward to our joint project.

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitalization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs

Our global industry and business expertise in business development, sales and marketing - Image: Xpert.Digital

Industry focus: B2B, digitalization (from AI to XR), mechanical engineering, logistics, renewable energies and industry

More about it here:

• Xpert Business Hub

A topic hub with insights and expertise:

• Knowledge platform on the global and regional economy, innovation and industry-specific trends
• Collection of analyses, impulses and background information from our focus areas
• A place for expertise and information on current developments in business and technology
• Topic hub for companies that want to learn about markets, digitalization and industry innovations