Google sends AI chips into space, Earth’s data center backlash, did the plumbing lift SPCX, and Jonathan Rose on what comes after the landing

The intersection of artificial intelligence and aerospace technology reached a significant milestone this week as Google launched its first experimental AI-processing satellite, marking a potential shift in how the world’s most resource-intensive computing tasks are managed. This technological leap occurs against a backdrop of volatile market conditions on Wall Street, characterized by dramatic fluctuations in semiconductor stocks and a notable surge in Treasury yields that briefly touched levels not seen in over two decades. As terrestrial data centers face increasing regulatory and environmental headwinds, the aerospace sector is positioning itself as the next frontier for the massive "compute" requirements of generative AI models.

The MVP Mission: Testing Google’s AI Architecture in Orbit

On Thursday at 2:32 p.m. ET, a SpaceX Falcon 9 rocket lifted off from Vandenberg Space Force Base in California, carrying a payload that could redefine the infrastructure of the internet. Tucked within the satellite, internally designated as "MVP" by Google, are four specialized AI chips—the same hardware architecture utilized to power Google’s Gemini chatbot. This mission represents the first real-world stress test of whether high-performance AI chips, typically housed in climate-controlled terrestrial facilities, can survive the rigors of space.

Before the launch, technicians at a specialized laboratory in San Francisco subjected the satellite to extreme vibration testing, simulating the intense G-forces and "bumpy" atmospheric exit of a rocket launch. While the current hardware possesses the computing power of roughly one server and consumes power equivalent to a standard household hair dryer, the objective is proof of concept rather than immediate scale. Google engineers are seeking to determine if these chips can maintain operational integrity under the harsh radiation and extreme temperature fluctuations of low Earth orbit (LEO).

The Terrestrial Crisis: Why AI is Moving Off-Planet

The push toward orbital computing is driven by a growing crisis on the ground. Terrestrial data centers, the backbone of the AI revolution, are becoming increasingly difficult and expensive to build. These facilities require immense quantities of electricity to run processors and millions of gallons of water for cooling systems. According to recent industry data, the energy consumption of data centers is expected to double by 2026, putting immense strain on aging power grids.

Beyond technical requirements, social and regulatory opposition is mounting. Data Center Watch, an organization that tracks infrastructure projects, reports that at least 75 data center developments valued at approximately $130 billion were blocked or significantly delayed in the first quarter of 2025 alone. This trend continued into the second quarter, with another 45 projects worth $68 billion facing similar hurdles. Local communities have increasingly protested these projects over concerns regarding rising utility bills, noise pollution, and the depletion of local water tables.

Space offers a theoretical solution to these constraints. In specific orbital paths, solar panels can collect up to eight times more energy than those on Earth due to the lack of atmospheric interference and more consistent exposure to the sun. Furthermore, the vacuum of space provides a unique environment for thermal management, allowing chips to shed heat into the cold void without the need for massive water-intensive cooling towers.

Market Volatility: Micron and the 10-Year Treasury Yield

The technological optimism of the Google launch was tempered by a "whipsaw" session on Wall Street. Semiconductor giant Micron (MU) released blockbuster earnings, reporting that data-center revenue had jumped 11-fold year-over-year. Despite crushing estimates, investors initially sold off the stock, reflecting deep-seated anxieties about the cyclical nature of the memory market. Micron’s announcement of a massive increase in capital expenditure for new "fab" (fabrication) and clean room construction led some analysts to worry about future overcapacity once the current AI-driven demand peak subsides.

Compounding the pressure on the technology sector, the 10-year Treasury yield surged to 5.34% earlier in the session, its highest mark since 2002. Higher yields generally compress the valuations of growth stocks, as they increase the discount rate applied to future earnings. However, the market showed resilience; by early afternoon, the 10-year yield retreated to 5.21%, and stocks, including Micron, recovered much of their initial losses to trade in positive territory.

SpaceX and the Mechanics of the "Plumbing" Trade

The aerospace giant SpaceX (SPCX) has been a central figure in this market narrative, not just as a launch provider but as a burgeoning AI infrastructure player. Following its merger with Elon Musk’s AI venture, xAI, SpaceX now controls the Colossus data centers in Memphis, Tennessee. These facilities house hundreds of thousands of Nvidia (NVDA) chips, which SpaceX leases to major tech firms including Anthropic and Google. This transition from a pure-play rocket company to a "landlord of compute" has significantly altered its financial profile, with the company reporting a 92% year-over-year revenue increase to $7.8 billion in its first quarterly report as a public entity.

However, the recent price action in SPCX has been driven as much by "index plumbing" as by fundamentals. As lockup periods expired and more shares became available for trade, the company’s weighting in the Nasdaq-100 was adjusted. During the September rebalance, SpaceX’s weighting more than doubled from 1.28% to 2.82%, triggering an estimated $15.5 billion to $22 billion in forced buying from index-tracking funds.

Despite a 10% gain for investors who entered after the initial weighting shift, the company’s valuation remains a point of contention. Trading at approximately 106 times sales—compared to the S&P 500 average of 3.7—SpaceX is priced for perfection. Market observers are closely watching the upcoming lockup expirations on October 9 and October 24, as well as the full lockup expiration on December 8, which could introduce significant selling pressure as early employees and insiders gain the ability to liquidate their holdings.

The Lunar Economy: From Landers to Utilities

While Google and SpaceX focus on orbital data centers, a secondary "space race" is developing around lunar infrastructure. The United States and China are currently competing to establish permanent bases on the Moon, a project that requires a complex web of private-sector support.

Analysts are increasingly looking beyond the companies that build the rockets to those that will provide the "utilities" for a lunar colony. NASA’s Artemis program relies heavily on private contractors for communications, power, and logistics. Key players in this emerging sub-sector include:

1. Intuitive Machines Inc. (LUNR): Recently awarded a $180.4 million NASA contract for a mission to the lunar south pole, the company holds a $1.8 billion backlog. Analysts view LUNR not merely as a lander company but as a potential "utility company for the Moon," providing essential navigation and communication services that remain active long after a landing is completed.
2. Firefly Aerospace Inc. (FLY): Following the successful landing of its Blue Ghost spacecraft, Firefly has secured contracts for six additional lunar missions. The company is focused on building a dependable, repeatable delivery service to the lunar surface.
3. L3Harris Technologies Inc. (LHX): Representing a lower-volatility entry point into the space economy, L3Harris provides the RS-25 engines for NASA’s Artemis missions. As an established defense and aerospace contractor, it serves as a "pick and shovel" play that benefits regardless of which specific startup wins the race to build the first lunar base.

Strategic Implications and Future Outlook

The transition of AI compute into orbit is no longer a matter of pure speculation. Elon Musk has predicted that orbital compute will be the only viable way to scale AI by 2029 due to terrestrial power and permitting constraints. While Google’s more conservative research suggests that space-based AI may not reach full economic parity until 2035—contingent on launch costs falling to approximately $200 per kilogram—the trajectory is clear.

For the broader market, these developments signal a new era of infrastructure investment. The "Space Economy" is evolving from a government-funded exploration endeavor into a commercially driven extension of the global technology stack. Investors are now tasked with distinguishing between companies benefiting from temporary index mechanics and those building the long-term foundations of orbital and lunar industry.

As Google’s "MVP" satellite continues its orbit, the data it sends back will provide the first definitive evidence of whether the future of artificial intelligence lies within the atmosphere or among the stars. In the interim, the volatility seen in Micron and the Treasury markets serves as a reminder that even as technology reaches for the heavens, it remains tethered to the fundamental realities of Earthly finance.