OpenAI making phones: a lifeline, or a step toward self-destruction?

Original Author: Ada, TechFlow

TF International Securities analyst Ming-Chi Kuo released a message stating that OpenAI is collaborating with MediaTek and Qualcomm to develop a mobile phone processor, with Luxshare Precision as the exclusive manufacturing partner, and mass production is expected in 2028. This news has also been confirmed and reported by multiple media outlets.

As the news spread, supply chain stocks rallied first. Analysts began calculating MediaTek's incremental orders, Luxshare Precision's customer structure optimization, and Qualcomm's licensing revenue from its baseband solutions.

But the question is, why would a company not expected to turn a profit until 2030, with cumulative cash consumption potentially reaching approximately $115 billion, venture into making phones?

The Subscription Model Trap

OpenAI achieved $20 billion in ARR by 2025, with revenue growing 3,628 times since 2020. ChatGPT has 500 million weekly active users, making it a top-tier consumer internet product globally.

However, according to a Deutsche Bank report from October 2025, only about 5% of users have paid.

The remaining 95% are free users, and every single one of their conversations burns through computing power, electricity, and GPUs. Even Sam Altman has admitted that the $200/month Pro subscription is losing money. The full-year cash consumption in 2025 was approximately $9 billion, with 70% of revenue evaporating on servers.

Furthermore, according to Deutsche Bank analyst Adrian Cox's report, ChatGPT's consumer-side paying user growth in Europe nearly stalled in the second half of 2025. The ceiling for the paying user rate might be much lower than imagined, and the growth model driven by subscriptions is hitting a wall.

The problem with the subscription model is that while costs grow proportionally with the user base, revenue growth plateaus at a certain point. This isn't an issue in traditional SaaS, but it's fatal for AI companies.

So, what's the solution?

Advertising is one path. OpenAI has already started testing ads within ChatGPT, hiring Meta's monetization expert Fidji Simo as its consumer app CEO. However, advertising means directly competing with Google, which has hundreds of billions in annual cash flow from search ads alone and a very deep moat. It will be difficult for OpenAI to snatch market share from its rival.

Enterprise services represent another path. Enterprise revenue currently accounts for over 40% of OpenAI's total revenue and is indeed growing fast. However, Anthropic's annualized revenue from enterprise coding tools surged to $30 billion by March 2026, with its secondary market trading price once surpassing OpenAI's. This path is also crowded with competitors.

That leaves the third path: hardware.

Hardware Isn't a Dream, It's Financial Anxiety

OpenAI CFO Sarah Friar stated in a CNBC interview, "Hardware will be the next layer of value added to ChatGPT and will help drive user upgrades and subscription growth."

In other words, OpenAI needs a vehicle to convert free users into paying ones. Sell a phone, bundle it with a ChatGPT subscription, and users get charged automatically every month. No more waiting for users to open a browser and upgrade to Pro themselves. Hardware locks in the entry point, making the subscription the default option. It's the same principle as Apple bundling iCloud storage with the iPhone.

So, the vision Ming-Chi Kuo painted—redefining the phone with an AI Agent, where users execute tasks directly through the phone instead of opening a bunch of apps—makes for good tech narrative. But the underlying driver is more primal: OpenAI needs new monetization channels to fill its nearly ten-billion-dollar annual deficit.

OpenAI's core motivation for making a phone has nothing to do with innovation. It needs a way to transfer computing costs off its own balance sheet, and hardware is that transfer vehicle. When a user buys the phone, they implicitly pay for the cloud inference costs.

OpenAI plans an IPO as early as Q4 2026, targeting a $1 trillion valuation. Before going public, it must tell Wall Street a growth story beyond just "models are getting better." Enterprise revenue is being chased by Anthropic, advertising is just starting, and AI Agents are still conceptual. A phone is a good story. Over a billion phones are sold globally each year. Capturing even a small slice paints a promising revenue curve.

Lessons from the Past

The gap between a good story and a good business in AI hardware has been proven time and again.

Humane AI Pin raised $230 million, priced at $699 plus a $24 monthly subscription, and shipped fewer than 10,000 units. In February 2025, it sold itself to HP for $116 million, and the product was effectively bricked, with all user devices ceasing to function.

Rabbit R1 was the flashy little orange box at CES, selling 100,000 units before facing massive returns as users found many demoed features were unusable. The device's 10-second voice response latency made it impractical for real-time interaction. By early 2026, media reported the company couldn't even pay its employees. Users also discovered it was essentially a skinned Android app.

These two cases share a common fatal flaw: mistaking technological novelty for product-market fit. The demos were explosive, waiting lists were long, and the teams thought this was market validation. But when users got the product, they found it was better to just install the ChatGPT app on their phone.

Jony Ive himself publicly evaluated Humane AI Pin and Rabbit R1 as "very bad products," stating the entire industry lacks "products expressed with new thinking." He then sold his own company, io, to OpenAI for $6.5 billion.

The Rivals in 2028 Won't Be Today's iPhones

The OpenAI phone is expected to enter mass production in 2028. That's two years from now.

What will the phone market look like in two years?

Apple has already integrated both Google Gemini and ChatGPT into the iPhone, with a major AI overhaul for Siri expected by 2026. Samsung Galaxy AI covers its flagship and mid-range lines. Google Pixel natively runs Gemini. Android XR glasses are on the way.

In other words, by 2028, every mainstream phone on the market will be an "AI phone." AI capabilities will become a standard feature, just like cameras, GPS, and fingerprint sensors.

So where is OpenAI's differentiation?

Ming-Chi Kuo's answer is that an AI Agent needs to continuously understand the user's context, and only a phone has access to all of a user's real-time state information. Since OpenAI has the best model, the phone's experience will be different.

This answer has a glaring loophole: model capabilities can be provided via API. OpenAI is already selling its models to Apple and Samsung through APIs. If the model is the core advantage, selling the model to all phone manufacturers would be more profitable and less risky than making the phone itself.

Unless OpenAI believes that just selling model APIs isn't generating enough revenue.

This brings us back to the core question: Is the phone venture for technological ideals or financial survival?

Tech history is filled with failed hardware ventures, and successful cases of software companies making hardware are few and far between. Google has been making Pixels for a decade, holding less than 2% of the global market. Microsoft's Surface line lost money for years before barely breaking even. These companies had tens of billions in cash flow to support their trial and error, but OpenAI doesn't.

The $852 Billion Bet

OpenAI's phone story is, at its core, a narrative necessity corresponding to its $852 billion valuation.

Model capabilities are converging. The lead time for new models may only be a few months, with Gemini, Claude, and Llama all in pursuit. When the model itself becomes a commodity, profit margins for selling models will only thin out further.

Meanwhile, subscription revenue is hitting its ceiling. The 5% paid conversion rate already reveals the market's true willingness to pay. The enterprise market is also being poached by Anthropic. In the secondary market, Anthropic's trading price has already surpassed OpenAI's—investors are voting with their feet.

Against this backdrop, "making a phone" gives investors a new space for imagination. If OpenAI can sell 100 million AI phones, each tied to a $20 monthly subscription, that's $24 billion in new annual revenue. Combined with hardware revenue itself, total revenue could double instantly.

The math is easy. But the math for Humane and Rabbit looked good too at the time—beautiful on paper, disastrous in sales. Consumers aren't willing to pay for a phone without an application ecosystem. Without WeChat, without TikTok, without the complete Google Play ecosystem, even the most powerful AI Agent can't meet daily life needs.

Ming-Chi Kuo stated that OpenAI might adopt a business model bundling subscriptions with hardware sales. In other words, hardware might be sold at a loss, with costs recouped through subscriptions. It's another "lose now, profit later" story. OpenAI has been telling this story for the past three years, and investors have listened for three years.

The question is: when the phone reaches mass production in 2028, how much longer can the story last? By then, OpenAI will have burned through over $100 billion cumulatively. If the phones don't sell, the flywheel won't just stop spinning—it will spin backward.

CFO Sarah Friar has already expressed doubt about OpenAI's IPO timeline, stating the company isn't ready to go public and voicing reservations about the massive $600 billion spending plan over the next five years. According to Bloomberg, a research firm that contacted hundreds of institutions found that "not a single one was willing to buy OpenAI in the secondary market."

The most likely outcome for OpenAI's phone venture isn't redefining the mobile phone industry, but adding one more slide to its IPO roadshow deck. As for how much that slide will ultimately be worth, it probably depends on conditions over which OpenAI has no control whatsoever.