Elon Musk has once again shut down a speculation that had been growing for days around SpaceX: the possibility that the company is developing a portable artificial intelligence device, more similar to a phone or connected terminal than a traditional aerospace product. His response was definitive. According to Reuters, Musk called the information from The Wall Street Journal that claimed SpaceX had shown investors a prototype of a handset-like device before its IPO “utterly false.”
The denial doesn’t eliminate the interesting part of the story. Instead, it shifts it. Even without confirmation of a product, the rumor aligns with three themes that SpaceX, Starlink, and Musk’s business universe have long been approaching: direct satellite connectivity, AI at the edge, and a potential layer of mobile services to reduce dependence on carriers, manufacturers, and app stores.
The information from The Wall Street Journal, reported by Reuters, described a prototype with its own operating system, xAI technology, and a Qualcomm Snapdragon chip. The project would be in an early stage, with a still-open design and no certainty it will be manufactured. The Verge added, citing the same report, that the device would be described as thinner than an iPhone with an AI-enabled operating system.
A firm denial about an uncomfortable possibility
Musk had already denied in February that SpaceX was working on a phone. Reuters then reported that the company was considering a mobile device connected to the Starlink constellation, and Musk responded on X that “we are not developing a phone.” In that context, he also mentioned that a potential Starlink phone “is not ruled out at some point,” but it would be “very different” from current mobiles and optimized to run neural networks with maximum performance per watt.
That distinction is important. Musk can deny that SpaceX is manufacturing “a phone” in the traditional sense, without completely closing the door to a different device: an AI terminal, an interface for Grok, a personal computing node, a device connected to Starlink, or a part of a larger ecosystem. Language matters because the smartphone market is saturated, but the market for native AI devices is still finding its shape.
Recent history suggests caution. Many attempts to create AI devices have faced a simple problem: smartphones are already in users’ pockets. To justify new hardware, it’s not enough to say it includes AI. It must do better than an iPhone or Android with an AI app, mobile connection, camera, payments, messaging, and app store.
Here, SpaceX could have a potential advantage others lack: Starlink. The company doesn’t need to imagine an alternative connectivity; it already operates a satellite constellation and is pushing direct-to-phone services. According to Reuters, SpaceX has communicated plans to investors to launch a Starlink mobile service for consumers in the U.S., with a possibility of competing more directly with Verizon, AT&T, and T-Mobile. Reuters also noted that SpaceX already offers direct-to-cell connectivity with T-Mobile as a satellite-based supplementary coverage.
Starlink, AI, and the old dream of controlling the interface
The idea of an in-house device makes strategic sense even if no commercial product has been confirmed. Musk controls or influences components that could form a complete tech chain: SpaceX for space infrastructure, Starlink for connectivity, xAI for models, X as a social layer, and Tesla for hardware experience, data, and connected users.
The challenge is that smartphones are among the most difficult markets in the world. Apple and Google control dominant operating systems, app stores, software distribution, much of the developer relations, and much of the user experience. Samsung, Xiaomi, Oppo, and other manufacturers add industrial scale and global channels. Entering this space isn’t just launching a gadget; it’s competing against entire platforms.
Therefore, a hypothetical SpaceX device would need to avoid seeming “just another phone.” Its narrative should focus on something more: coverage outside networks, local or semi-local AI, deep integration with Grok, satellite services, security, energy autonomy, sensors, or use cases where current mobiles are insufficient.
Direct satellite connectivity to the device is already a partial reality. T-Mobile offers T-Satellite with Starlink for texts and certain compatible apps in outdoor areas with a clear sky, explaining that Starlink’s direct-to-cell satellites supplement their network by acting as space-based cellular towers.
Standardization is also progressing. 3GPP has worked on non-terrestrial networks (NTN) within recent releases to integrate satellites into the 5G framework. For example, Release 18 work relates to NTN bands for satellite communications.
But moving from basic connectivity or limited services to a global device with an experience comparable to terrestrial telephony is another scale. Challenges include bandwidth, power consumption, antennas, modems, transmission power, latency, satellite availability, spectrum agreements, regulatory compliance, and costs. The commercial magic of a “Starlink phone” runs into physics and regulatory hurdles.
Qualcomm, xAI, and edge hardware’s role
Mentioning Qualcomm in the supposed prototype isn’t accidental. If an AI device is to operate outside data centers, it needs efficient processors, mobile connectivity, and local inference capabilities. Snapdragon chips already compete in laptops, smartphones, extended reality, and connected devices, fitting into a terminal where battery life and efficiency matter more than raw power.
Nevertheless, most advanced generative AI still depends on data centers. Slim devices can run small models, voice and vision tasks, or local privacy functions, but they don’t replace the infrastructure supporting large models. If such a device ever exists, it would likely have a hybrid architecture: local inference for quick or sensitive tasks, and remote connection for heavier workloads.
In that context, Starlink could offer a unique advantage: access to AI even outside terrestrial networks. However, this promise would only be valuable if the experience is reliable, affordable, and fast enough. For emergencies, remote areas, industry, defense, mobility, logistics, or outdoor activities, it could make sense. But for the average city user, competing with current smartphones would be much harder.
The real question isn’t whether SpaceX can build a device. With enough capital and partners, probably yes. The challenge is whether there’s a large, profitable market for proprietary hardware when Starlink can expand through existing smartphones via operators and standards.
Interest in mobile exists even without a phone
Although Musk denies the device, SpaceX is moving into the mobile space. Reuters reported in June that the company is exploring a Starlink Mobile service for U.S. consumers and may build its own terrestrial mobile network. The same report mentioned SpaceX’s acquisition of spectrum licenses from EchoStar and ambitions to strengthen direct-to-cell services.
This suggests a more realistic possibility than a “SpaceX iPhone”: that the company aims to become a hybrid mobile connectivity layer, combining satellite and terrestrial networks, compatible with existing devices. This route would allow entry into telecom markets without the complexity of manufacturing and maintaining a global smartphone upfront.
It also aligns with Starlink as SpaceX’s core business. Reuters noted in February that Starlink was a key revenue generator for the company and that SpaceX operates over 9 million satellite internet users. If Starlink aims to grow beyond fixed antennas, boats, planes, and vehicles, mobile connectivity becomes the next logical step.
Musk’s denial may be literal and true: SpaceX isn’t making a phone or hasn’t shown the described prototype. But it doesn’t mean the company isn’t exploring ways to bring Starlink, xAI, and mobile services to more devices. In tech, the final product often differs from initial rumors, but the rumors point toward a strategic direction.
A market full of traps for AI devices
The rise of generative models has revived debates about new hardware formats. Glasses, pins, pendants, voice assistants, badges, pocket terminals, and screenless devices have attempted to occupy the space left by smartphones. Most face the same obstacles: lack of daily usefulness, dependency on phones, battery issues, latency, privacy concerns, high prices, or inferior experiences compared to opening an app.
SpaceX would need to overcome all these challenges and add its own. An AI-related device connected to Starlink would have to manage variable coverage, indoor use, power consumption, national regulations, and operator agreements. An AI device tied to xAI would need to demonstrate a sufficiently compelling experience. A device with a proprietary OS would need to handle apps, identity, payments, communications, security, and updates.
Thus, the most cautious approach would be to focus first on developing services, not hardware. Improving direct-to-cell, launching Starlink Mobile, integrating AI into existing apps, offering connectivity to third parties, and letting the market show where specific devices are needed. Rushing into manufacturing a terminal could be costly and distract from the main advantage: the network.
This approach echoes Musk’s other major tech ambitions, which often begin as denials, leaks, or contradictory signals before becoming actual products. Sometimes they remain ideas that never materialize. With the available information, the sensible view is to treat SpaceX’s AI device as a dismissed rumor, not an unannounced product launch.
What’s clear is that the boundary between telecom, satellites, AI, and personal devices is shifting. If Starlink manages to become a complementary or alternative mobile network, pressure on operators and manufacturers will grow. If xAI joins this layer, the discussion will move beyond coverage to interfaces, data, and control over user relationships.
Musk can deny the device. However, the convergence that would make it possible is already underway.
Frequently Asked Questions
Is SpaceX manufacturing an AI phone?
There is no confirmation. Elon Musk has denied the Wall Street Journal report and called it “utterly false,” according to Reuters.
What did the rumor say about the device?
It described a prototype handset, thinner than an iPhone, with its own OS, Qualcomm Snapdragon chip, and xAI technology.
Does a connected Starlink device make technical sense?
It’s possible in certain scenarios but complex. It requires addressing antennas, modems, power, spectrum, coverage, bandwidth, regulation, and user experience.
What is direct-to-cell?
A technology allowing satellites to provide coverage directly to compatible phones, usually as a supplement to terrestrial networks in areas without coverage.
Why might SpaceX want to enter the mobile market?
Because Starlink could evolve beyond fixed satellite internet into a layer of mobile connectivity. Building a device could be an option, or they could proceed via partnerships and existing phones.
via: tomshardware and X Twitter
