The Race for Data Centers in Space | Philip Johnston on Starcloud and the Future of Compute скачать в хорошем качестве

The Race for Data Centers in Space | Philip Johnston on Starcloud and the Future of Compute

AI is running out of energy. Earth’s data centers already consume more power than some countries, and the demand for GPUs is accelerating faster than grids can grow. The next step, says aerospace engineer and Starcloud founder Philip Johnston, is radical: put data centers in space. In this episode, Mark and Jeremy Think On Paper with Philip, who explains why orbit might be the only scalable frontier for computing — where cooling is radiative, solar power is constant, and latency can be cut by placing clusters directly above population centers. Philip breaks down the real physics and economics: how sun-synchronous orbits keep servers in permanent daylight, why radiative cooling works better than water, and how modular GPU clusters could make orbital infrastructure as routine as terrestrial cloud services. What begins as a discussion about data centers and launch costs becomes a deeper conversation about ambition, sustainability, and what it means when the cloud itself leaves Earth. Will space-based compute solve the energy crisis — or just move it beyond the atmosphere? If you care about AI infrastructure, energy use, or the future of computing in space, this episode belongs on your list. Please enjoy the show. And remember: Stay curious. Be disruptive. Keep Thinking on Paper. Cheers, Mark & Jeremy -- Timestamps 00:00 — The environmental impact of Earth-based computing 03:09 — Why build data centers in space? 08:36 — Conquering latency and communication limits in orbit 10:32 — Designing modular, reusable space infrastructure 16:03 — The economics: how much do space-based data centers cost? 19:46 — Manufacturing beyond Earth’s boundaries 26:00 — Reusability, recycling, and the problem of space junk 26:15 — GPUs in orbit: high-performance computing off-world 28:52 — Future tech rapid-fire: what’s next in orbital compute 29:55 — Five billion humanoids and the future of demand 32:35 — Addressing skepticism about data centers in space 33:41 — Quantum computing’s orbital advantage 36:00 — The humanoid difference: merging AI and biology 37:20 — Where are all the aliens? The Fermi paradox revisited 38:35 — Closing reflections and what comes next -- Links and resources Join us; www.thinkingonpaper.xyz Starcloud: https://www.starcloud.com/ Philip Johnston:   / philipjohnst   Space elevators: https://www.scientificamerican.com/ar... Proception humanoids: https://www.proception.ai/ --