Pat Gelsinger is gone. His vision isn’t — not entirely. But the pieces of Intel he tried to assemble into something coherent are now scattering in directions that would have seemed absurd even two years ago. The latest twist: Intel’s foundry business, newly semi-independent and desperate for relevance, is reportedly angling for satellite and space-related chip contracts. At the same time, the company’s entanglement with the U.S. government’s industrial policy ambitions has become so complex that untangling it may require a map, a forensic accountant, and possibly a therapist.
The story starts with CHIPS Act money — $7.86 billion in direct subsidies that Intel was promised under the Biden administration’s signature semiconductor initiative. That funding, designed to reshore advanced chip manufacturing to American soil, came with strings. Lots of them. And now, under a Trump administration that views government spending through a fundamentally different lens, those strings are being yanked in new directions. According to The Register, the current political environment has introduced a wild card that nobody in Santa Clara’s boardroom could have predicted: Elon Musk.
Not Musk the technologist. Musk the political operator.
His influence over the current administration’s technology priorities is, by most accounts from Washington insiders, substantial. SpaceX’s dominance in launch services and Starlink’s growing monopoly on low-earth-orbit broadband have made Musk the de facto gatekeeper for American space infrastructure. Intel, which has been quietly developing radiation-hardened chip designs and exploring contracts with defense and intelligence agencies for satellite processors, now finds itself in an uncomfortable position. Any serious push into space-grade silicon puts Intel on a collision course with Musk’s commercial interests — or, at minimum, requires his implicit blessing.
That’s a strange place for a $100 billion semiconductor company to be.
The Register’s reporting frames Intel’s space aspirations as part of a broader pattern of magical thinking that has plagued the company’s strategic planning for years. The publication argues that Intel’s leadership — both past and present — has consistently overestimated its ability to enter adjacent markets while underestimating the difficulty of fixing its core manufacturing problems. The space play, in this reading, is less a calculated expansion and more a symptom of institutional desperation. A company that can’t reliably produce leading-edge chips for laptops and data centers is now promising it can build processors tough enough to survive cosmic radiation and temperature extremes that would melt consumer electronics.
There’s a counterargument, of course. Intel’s fabrication facilities in Oregon, Arizona, and Ohio — the ones the CHIPS Act money is supposed to expand — genuinely could produce specialized chips for defense and space applications. The company has a long history with the U.S. military. Its older process nodes, the ones that aren’t competing with TSMC’s latest, are actually well-suited for the kind of hardened, reliable silicon that satellites and missiles require. The margins on these contracts can be enormous. And the national security argument for domestic production of space-grade chips is, if anything, stronger than the argument for domestic production of smartphone processors.
But here’s the problem. Intel isn’t just competing against foreign fabs for these contracts. It’s competing against a political reality in which Musk’s companies are both the primary customer for space-grade electronics and the most powerful lobbying force in American space policy. SpaceX already designs much of its own avionics hardware. Starlink satellites use custom chips. The idea that Intel could waltz into this market without navigating Musk’s preferences and priorities is, according to multiple analysts quoted by The Register, naive at best.
The financial picture makes the stakes clearer. Intel’s foundry division — now operating as a separate business unit under CEO Lip-Bu Tan’s restructuring — lost roughly $7 billion in 2024. The unit is supposed to attract outside customers, companies that would pay Intel to manufacture their chip designs, much as TSMC does for Apple, Nvidia, and Qualcomm. So far, the customer list is thin. Amazon’s AWS has committed to some test production. The U.S. Department of Defense has signaled interest. But the blockbuster commercial deals that would justify Intel’s massive capital expenditure plans haven’t materialized.
Space contracts alone won’t fill that gap. The entire global market for space-grade semiconductors is measured in single-digit billions annually. Intel’s foundry needs tens of billions in revenue to reach profitability. The math doesn’t work unless space is one small piece of a much larger puzzle — and right now, the other pieces aren’t falling into place either.
Recent reporting from Reuters has highlighted the broader challenges facing Intel’s turnaround effort. Tan, who took over as CEO in early 2025, has been praised for his operational discipline and willingness to make hard cuts. He’s slashed headcount, killed underperforming product lines, and pushed decision-making authority down to engineering teams. But he inherited a company that spent a decade underinvesting in manufacturing technology while simultaneously overinvesting in acquisitions that never paid off. Mobileye. Altera. The autonomous driving unit. Each was supposed to be the next growth engine. None delivered at the scale Intel needed.
And then there’s the AI problem. Nvidia’s dominance in training and inference chips for artificial intelligence has created a gravitational pull that’s warping the entire semiconductor industry. Intel’s Gaudi accelerators — acquired through the Habana Labs purchase — have gained minimal traction against Nvidia’s CUDA-entrenched customer base. Every quarter that passes without a credible AI hardware story makes Intel less relevant to the customers who are spending the most money on chips right now. Hyperscalers. Cloud providers. AI startups flush with venture capital. They’re all buying Nvidia. Or AMD. Or designing their own chips and having TSMC build them.
Intel’s pitch to these customers is essentially: trust us, our manufacturing will catch up, and when it does, you’ll want to be on our platform. It’s a promissory note. And the semiconductor industry has a long memory for broken promises.
The Musk dimension adds a layer of political risk that Intel’s board likely didn’t model in its scenario planning. Musk’s Department of Government Efficiency — DOGE — has already taken aim at federal spending programs, and while the CHIPS Act funds have bipartisan support, they’re not immune to scrutiny. Any suggestion that Intel is using taxpayer subsidies to compete against Musk-affiliated companies in the space sector could trigger exactly the kind of political fight that Intel can’t afford right now. The company needs Washington’s support, not its attention.
There’s also the question of whether Intel’s space chip ambitions are technically credible. Radiation-hardened processors are a niche specialty. Companies like BAE Systems, Microchip Technology, and Cobham Advanced Electronic Solutions have spent decades building expertise in this area. Intel’s experience with rad-hard design is limited. Its process technology, while advanced for commercial applications, hasn’t been extensively validated for the kind of total ionizing dose and single-event effects testing that space applications require. Building a chip that works in a data center is hard. Building one that works reliably for fifteen years in geostationary orbit, bombarded by solar particles, is a different discipline entirely.
None of this means Intel’s space play is doomed. It means it’s hard. Harder than the optimistic briefing slides probably suggest.
The Register’s characterization of a “trillion-dollar delusion” — referencing the combined market value that Intel’s leadership has at various points suggested the company could achieve through its foundry and product businesses — is pointed but not entirely unfair. Intel’s market capitalization currently sits around $90 billion, down from peaks above $250 billion. The gap between where Intel is and where it says it’s going has rarely been wider. And the credibility deficit compounds with each quarter of losses, each delayed process node, each customer that chooses TSMC instead.
Lip-Bu Tan knows this. His public statements have been notably more measured than Gelsinger’s, avoiding the kind of grandiose timelines and market-share predictions that his predecessor favored. But measured doesn’t mean modest. Tan has committed to Intel 18A — the company’s next-generation manufacturing process — as a make-or-break technology. If it works as promised, delivering competitive performance and power efficiency by late 2025 or early 2026, Intel’s foundry story becomes real. If it doesn’t, the story is over.
Space contracts, defense work, government subsidies — these are all supporting actors. The main event is whether Intel can build chips as well as TSMC. Everything else is a sideshow. An expensive, politically complicated, Elon Musk-adjacent sideshow.
For industry insiders watching this unfold, the Intel saga is a case study in what happens when industrial policy collides with market reality and political power. The CHIPS Act was designed with a specific theory of the case: that American national security requires domestic advanced semiconductor manufacturing, and that Intel is the best vehicle for achieving it. That theory hasn’t been disproven. But it’s being tested — by Intel’s execution challenges, by TSMC’s construction of its own Arizona fabs, by Musk’s expanding influence over technology policy, and by an administration that may ultimately decide the money is better spent elsewhere.
The next twelve months will be decisive. Intel 18A either works or it doesn’t. The foundry either signs major customers or it doesn’t. The CHIPS Act money either flows or it gets redirected. And Elon Musk either remains a background presence in Intel’s strategic calculations or becomes a direct obstacle.
No amount of space-chip ambition changes those fundamentals. Intel’s future will be decided in the fab, not in orbit.
WebProNews is an iEntry Publication