SpaceX Eyes Boost Mobile Acquisition in Bid to Become Fourth Largest US Carrier

According to The Verge's podcast Vergecast, SpaceX is seriously considering entering the mobile communications market, and acquiring Boost Mobile could be a key step. The news was first reported by The Washington Post; if it materializes, it means the company known for rockets and satellites will directly challenge the three major US telecom carriers.

From Satellites to SIM Cards: SpaceX's Carrier Ambitions

SpaceX's intent to become the "fourth largest US carrier" is already quite clear. In addition to continuously launching satellites to build its constellation network, the company has also been purchasing spectrum resources in bulk and is reportedly in talks to access about 80,000 cellular base stations across the country. Pieced together, these moves point to a blueprint for a full mobile virtual network operator (MVNO) or even a facilities-based carrier.

Interestingly, Boost Mobile currently relies on T-Mobile and AT&T's networks to operate. If SpaceX acquires Boost, it would instead need to apply to these two future competitors for network access permission in order to launch wireless service. This awkward "board the train first, buy the ticket later" situation adds a layer of drama to the whole affair.

Strategically, SpaceX's path is entirely different from traditional carriers: it has a low-Earth-orbit satellite network (Starlink) that can cover areas ground base stations struggle to reach, and combined with spectrum and base station leasing, it has an opportunity to build an integrated space-and-ground communications service. But the challenges are equally enormous—regulatory barriers in the telecom industry, spectrum coordination, and roaming agreements with existing carriers are not things that can be solved as quickly as launching a rocket.

Mobile virtual network operators (MVNOs) and facilities-based carriers are two fundamental forms in the telecom industry. An MVNO does not itself own wireless spectrum or radio access network assets such as base stations; instead, it buys voice, data, and roaming services in bulk from licensed facilities-based carriers and resells them to users under its own brand. Typical examples include Mint Mobile, Google Fi in the US, and Boost Mobile mentioned in this article. A facilities-based carrier, by contrast, holds spectrum licenses itself, builds and operates base stations and core networks, and can fully control network quality and coverage. There is also a hybrid form: a company can first enter the market quickly as an MVNO, accumulate users, and simultaneously gradually build its own spectrum and base station assets, eventually transforming into a facilities-based carrier. SpaceX, currently purchasing spectrum in bulk, negotiating access to about 80,000 cellular base stations, and owning the Starlink low-Earth-orbit constellation, is advancing along exactly this hybrid path—first using ground networks to fill in indoor and urban coverage, then using satellites to cover remote areas, creating an integrated space-and-ground service capability.

Anthropic's Financials: Growth and Losses Soaring in Tandem

The same episode also mentioned AI company Anthropic's updated IPO filing, whose financial numbers are staggering. The filing shows that as of September 30, Anthropic's revenue for the year reached $33.24 billion, up 1,206% year over year; it brought in $17 billion in the last quarter alone. That said, the company's loss over the same period was as high as $23.29 billion.

Even more striking is its infrastructure spending commitment: Anthropic plans to invest nearly $900 billion in compute and infrastructure, with nearly $70 billion to be spent in the next 12 months alone. Capital expenditure at this scale shows that the AI arms race is far from cooling down, and leading companies are using massive investment to buy model capabilities and market share.

Revenue and losses both expanding at the hundred-billion-dollar scale reflect a typical feature of the current AI business model: demand is real and robust, but unit economics have not yet been proven. Whether Anthropic can prove this path is sustainable to the public market after its IPO will be one of the most noteworthy things to watch next.

Unit economics refers to the relationship between the revenue a company generates from a single customer, a single transaction, or a single product unit and the cost it incurs for it; it is a core metric for judging whether a business model can hold up over the long term. For AI companies, unit economics mainly depend on three cost blocks: the compute consumed by model inference (GPU hours and electricity), the R&D investment needed to train the next generation of models, and customer acquisition and service costs. If the revenue brought in by each user remains below the variable cost of serving that user over the long term, then the larger the scale, the larger the loss—growth itself cannot automatically lead to profitability. Anthropic's revenue growing 1,206% year over year while its loss over the same period was $23.29 billion shows that demand-side growth is real, but inference costs, compute depreciation, and R&D amortization still outweigh revenue. To make unit economics work, companies usually need to rely on improved inference efficiency, lower per-unit model costs, higher subscription prices, or a shift toward higher-margin enterprise API business—and this is also one of the variables the market watches most closely when evaluating AI company IPO pricing.

When Freezers Become Attack Targets: A New Weak Spot in IoT Security

The episode concluded by mentioning a wave of cyberattacks targeting industrial refrigerators and freezers. Several grocery stores in Europe reported that hackers broke in and directly tampered with in-store temperature settings; in one store, freezers were raised to 122 degrees Fahrenheit (about 50 degrees Celsius), putting food at risk of spoilage.

Such attacks expose a common problem with IoT devices: many industrial refrigeration devices are connected to networks, yet their firmware updates and security protections lag far behind consumer-grade products. Attackers do not need to steal data; merely changing a temperature setting can cause food waste and business losses. As more traditional devices become connected, similar security blind spots will only multiply.

The security dilemma of industrial IoT (IIoT) devices stems from their lifecycle and design orientation. Consumer devices usually receive regular firmware updates pushed by manufacturers, use relatively unified protocols, and have shorter replacement cycles; by contrast, commercial refrigeration equipment often has a design life of more than ten years, embedded controllers prioritize stable operation above all else, and upgrades after leaving the factory are rare. Default passwords, plaintext communications, and a lack of network segmentation are widespread. Many devices connect to the internet through building management systems or remote monitoring platforms, and once these entry points are compromised, attackers can issue temperature-setting commands in bulk. Such attacks are "operational technology (OT) attacks": the goal is not to steal data but to directly disrupt physical processes, potentially causing food spoilage, drug failure, or even equipment damage. Mitigation measures include isolating OT networks from IT networks, enforcing changes to default credentials, restricting remote write permissions, and setting automatic alerts and interlock protection for abnormal temperatures.

The Common Thread Behind These Three Stories

SpaceX's carrier dream, Anthropic's massive investment, and hacked freezers may seem unrelated, but all point to the same trend: the boundaries of infrastructure are being redefined. Satellite networks are becoming communications infrastructure, compute is becoming the core asset of AI companies, and even a freezer has become part of the cyberattack surface. Whoever controls these new types of infrastructure will control the initiative in the next decade.

For SpaceX, the path to becoming the fourth largest carrier is already laid out, but regulatory and competitive barriers should not be underestimated; for Anthropic, revenue growth proves the market is buying in, but the scale of losses also reminds people that the profitability inflection point for AI has not yet arrived.

One-line Summary

SpaceX plans to acquire Boost Mobile to enter telecom, while Anthropic's massive losses and hacked freezers both point to the redefinition of infrastructure boundaries.

Paragraph Summary

According to Vergecast, SpaceX is seriously considering acquiring Boost Mobile to become the fourth largest US carrier. It has already purchased spectrum and is negotiating access to about 80,000 cellular base stations across the country, but Boost currently relies on T-Mobile and AT&T networks, so after the acquisition it would instead need to apply to competitors for access permission. The same episode also disclosed Anthropic's updated IPO filing: as of September 30, annual revenue was $33.24 billion, up 1,206% year over year, while the loss over the same period was $23.29 billion, and it committed nearly $900 billion to compute and infrastructure investment. In addition, freezers at several grocery stores in Europe were hit by cyberattacks, with temperatures tampered with to about 50 degrees Celsius, exposing the security shortcomings of industrial IoT devices. Together, these three stories point to the same trend: satellite networks, AI compute, and connected refrigeration equipment are all becoming new infrastructure and new risk frontiers.