# AI Is Compressing Software; Space Is Building the Physical Economy > Source: > Published: 2026-07-30 21:03:54+00:00 For almost three decades, “learn to code” was treated as universal career advice. It was not bad advice. Software ate the world, and those who could build it captured an extraordinary share of the value. However, AI is changing the economics of digital work. Glenn Edens, inventor of the laptop, royalty in Silicon Valley, told me in 2022, “Z, AI is not going to replace plumbers; it will hit white-collar jobs hardest.” As they have been so often, Glenn’s words proved prophetic. In the first half of 2026, the technology sector announced 139,156 job cuts, [83% more](https://www.challengergray.com/blog/challenger-report-june-layoffs-cool-to-45849-down-53-from-may-ai-leads-reasons-for-fourth-consecutive-month/?utm_source=chatgpt.com) than during the same period in 2025. Across U.S. employers, AI was cited in more than 101,000 announced cuts. Not every one of those jobs was replaced by an algorithm; over-hiring, cost pressure, and ordinary restructuring matter. But the direction is clear: AI allows many organizations to produce more digital output with fewer people. That does not mean technology careers are disappearing. It means value is moving from work that exists entirely on a screen toward industries that must turn intelligence into physical capability. At some point, every value chain is anchored in the real world, passing through but never ending in cyberspace. The space economy is one of the clearest examples. The global space economy reached a record [$613 billion in 2024](https://www.spacefoundation.org/2025/07/22/the-space-report-2025-q2/?utm_source=chatgpt.com), growing 7.8%, with commercial activity accounting for 78% of the total. The World Economic Forum and McKinsey project that it could reach $1.8 trillion by 2035. Space is not immune to layoffs, failed startups, or capital cycles. Nothing is. But it is more resistant to the compression occurring in purely digital businesses because its output is physical capital. A satellite cannot be prompted into orbit. Lunar power systems, ground stations, robotic servicers, orbital factories, pharmaceutical platforms, and return vehicles must be designed, financed, fabricated, integrated, qualified, launched, and operated. AI will improve each activity and may reduce the labor required for coding, analysis, and documentation. It does not remove the need to build, test, and take responsibility for systems that must survive radiation, vibration, vacuum, and failure. This is why semiconductors belong at the center of the conversation. [View All](https://www.eetimes.com/category/sponsored-content/) AI needs compute. Compute needs semiconductors, power, cooling, networking, and resilient supply chains. For decades, better chips enabled better spacecraft. Now the relationship is becoming bidirectional. Space companies are [consuming high-performance](https://www.semiconductors.org/), radiation-tolerant, and energy-efficient electronics, while some are exploring whether the space environment can improve the materials from which future chips are made. The credible near-term case is not to lift an entire [leading-edge fab into orbit](https://www.nasa.gov/general/the-benefits-of-semiconductor-manufacturing-in-low-earth-orbit-leo-for-terrestrial-use/?utm_source=chatgpt.com). It is to use microgravity and vacuum for steps where those conditions may offer an advantage—crystal growth, deposition, and high-value substrates—and then return the materials for terrestrial processing. NASA has argued that gravity creates barriers to rapid, high-yield semiconductor production, and that low Earth orbit could become part of a future semiconductor supply chain. That argument is moving beyond white papers. Space Forge’s ForgeStar-1 [generated plasma in orbit](https://www.spaceforge.com/press-releases/space-forge-ignites-a-new-industrial-era-delivering-world-first-capability-for-orbital-semiconductor-manufacturing), showing that conditions required for gas-phase crystal growth could be created and controlled on an autonomous commercial platform. The company is focused on wide- and ultra-wide-bandgap materials used in power electronics, communications, quantum systems, defense, and high-performance computing. It has also partnered with Intuitive Machines to pair an orbital semiconductor reactor with a return vehicle, and with United Semiconductors to connect orbital deposition with terrestrial crystal growth, wafer processing, and testing. Policy is beginning to follow. The [bipartisan Semiconductor Superiority Act](https://www.budd.senate.gov/2026/06/11/budd-introduce-bipartisan-bill-to-unlock-u-s-space-based-chip-manufacturing/?utm_source=chatgpt.com), introduced in June 2026, would clarify that qualifying semiconductor manufacturing facilities in outer space can receive the advanced manufacturing investment credit created by the CHIPS and Science Act. It has not become law, but its introduction is a signal: Policymakers increasingly see space not merely as a customer for semiconductors, but as a possible part of the semiconductor industrial base. Contrast that with software. GitLab disclosed plans to eliminate approximately 350 positions—[14% of its workforce](https://www.sec.gov/Archives/edgar/data/1653482/000162828026039805/gitlab-ex99120260430fy27.htm?utm_source=chatgpt.com)—while its strategy described using AI agents to automate internal reviews, approvals, and handoffs. Meanwhile, Space Forge was proving a physical manufacturing process in orbit and assembling the partnerships needed to turn it into a supply chain. One company was using AI to compress its organization. The other was building hardware, logistics, and manufacturing capacity around the needs of the AI economy. The broader labor data point in the same direction. The Bureau of Labor Statistics projects computer-programmer employment to [decline 6% between 2024 and 2034](https://www.bls.gov/emp/skills/computers-and-information-technology.htm?utm_source=chatgpt.com). Semiconductor-processing-technician employment is projected to grow 11%, while electrical and electronics engineering is projected to grow 7%. These categories are not proxies for the whole space workforce, but they illustrate the shift: Software remains essential, yet opportunity increasingly lies in applying it to hardware, energy, manufacturing, and complex operations. Higher education is beginning to recognize what industry already knows: The next space workforce cannot be educated inside a single discipline. The University of Mississippi’s [Center for Air and Space Law](https://olemiss.edu/airandspacelaw/) addresses the legal and regulatory layer, preparing professionals to navigate the frameworks governing aviation, space, cyber, and emerging technologies. Embry-Riddle’s space operations programs focus on mission operations, safety, human factors, planning, policy, and complex space systems. The University of Central Florida’s Space MBA approaches the same economy through space entrepreneurship, governmental and commercial space finance, leadership, and the global space domain. These are not competing versions of the same degree. They address different bottlenecks: Ole Miss develops the legal architecture, Embry-Riddle develops operational capability, and UCF develops the commercial leadership required to turn missions into markets. [UCF’s Space Ideation Challenge](https://business.ucf.edu/space-ideation-challenge/), which asks participants to design market-shaping and pay-for-success policies, reflects the same change in thinking. Space innovation is no longer confined to laboratories. Regulation, finance, procurement, and market design increasingly determine whether technology becomes an industry. For someone considering a career change, the lesson is not to abandon existing expertise. It is to reposition it. A software engineer can move into autonomous spacecraft, mission planning, ground systems, or cyber-resilient operations. A semiconductor professional can work on radiation tolerance, edge computing, power electronics, or space-grown materials. Lawyers, financiers, supply chain experts, and product leaders are equally necessary because space companies must obtain licenses, secure capital and customers, manufacture repeatedly, and operate reliably. We spent thirty years teaching people how to build software. The next thirty will require teaching them how to build industries. Those are not the same task. The winners of the AI era will not simply write better algorithms. They will understand how semiconductors, energy, manufacturing, law, finance, and space come together to create physical systems and durable markets. AI is not ending technology. It is giving technology somewhere new to grow. ##### Read also: [Data Centers in Space: A Brilliant Idea or Delusional?](https://www.eetimes.com/data-centers-in-space-a-brilliant-idea-or-delusional/) [The Hidden Physics of Running Data Centers in Orbit](https://www.eetimes.com/the-hidden-physics-of-running-data-centers-in-orbit/) [Space-Station Tech Pivots to Cool AI Data Centers](https://www.eetimes.com/space-station-tech-pivots-to-cool-ai-data-centers/)