
Tech • IA • Crypto
Launch costs have drastically declined over the past 50 years, driven largely by advances in rocket reusability, and further cost reductions alongside innovations in power generation, cooling, and computing in space are opening new opportunities in orbital data centers, AI workloads, and potentially Bitcoin mining—though that remains a longer-term prospect constrained by latency and economics.
Dramatic Reduction in Launch Costs Historically, launch costs have fallen from approximately $50,000 per kilogram during the Space Shuttle era to about $1,000 per kilogram today, primarily due to innovations from SpaceX. Expert projections anticipate a further drop to around $100 per kilogram within 3 to 5 years, driven by increased rocket reusability and larger payload capacities, enabling a tenfold reduction from current levels. This reduction fuels commercial space markets and is a core impetus behind SpaceX’s interest in going public.
Reusability as the Key Cost Driver The next major cost compression is attributed to “massive reusability” of rockets—meaning rockets reused dozens of times rather than single-use launch vehicles. Larger reusable rockets increase payload size, reducing cost per kilogram by expanding the denominator and spreading fixed costs. This transition shifts space launch economics from a model akin to “single-use cars” to one resembling commercial air travel where vehicles are employed repeatedly.
Power Generation and Transmission in Space Solar panels in space operate at 5 to 8 times the efficiency of terrestrial panels due to the absence of atmospheric attenuation. Companies like Overview Energy are experimenting with converting space-generated solar energy into infrared light and beaming that power down to Earth-based data centers. Other approaches focus on storing and utilizing power in orbit. This abundance of space-generated power is pivotal for future space-based computing tasks.
Cooling Challenges and Solutions Unlike Earth-based cooling, which relies heavily on convection (air movement), space cooling depends on radiative cooling—dissipating heat as infrared radiation into the vacuum. While this is slower and requires extensive surface area, it is a solvable engineering problem rather than a fundamental physical limitation. Radiative cooling innovations are expected to reduce power consumption wasted on cooling, enhancing overall system efficiency.
Bandwidth and Latency Constraints Bandwidth in space is limited, especially for downlinking large volumes of data, creating bottlenecks for applications such as Earth observation and high-throughput satellite internet (e.g., Starlink). Starlink alone handles capacities approaching 1,000 petabytes per second. However, latency—the time taken for data to travel at the speed of light—poses a fundamental lower bound, with delays measured in seconds for lunar distances and minutes for Mars transmissions. This unavoidable latency influences which applications can effectively operate at different orbital or deep space ranges.
Emerging Applications: AI Workloads Lead AI inference and training workloads are the primary envisioned uses for orbital data centers due to their immense and growing computation demand. AI inference is predicted to dominate because of its near-infinite demand and revenue potential. The strategy involves processing large raw data locally in space (such as Earth observation images), then sending smaller refined data back to Earth, easing downlink bandwidth constraints.
Bitcoin Mining in Space: A Longer-Term Prospect Currently, Bitcoin mining is not economically viable in space due to high launch costs, bandwidth restrictions, and power considerations compared to more lucrative AI workloads. Estimates suggest either Bitcoin prices need to be roughly 100 times higher or launch costs need to drop to tens of dollars per kilogram—about a 100-fold improvement—for space-based Bitcoin mining to be competitive. The technological and economic landscape may shift in the next decade or two to make it feasible.
Latency as a Fundamental Limit and Rule Latency, rooted in the finite speed of light, remains the inescapable boundary for space computing applications. This physical constraint results in a layered or "onion shell" model where low-latency applications cluster near Earth and higher-latency applications handle computation farther out in the solar system. This segmentation will dictate how different space-based computing tasks evolve and coexist. An informal principle—proposed as the “Burnett rule”—describes this intrinsic latency-driven stratification of space computing.
Significant advances in rocket reusability have transformed launch economics, enabling ambitious new space-based computing architectures centered on AI workloads with power beamed or generated in orbit. Despite the promise, Bitcoin mining in space remains a future possibility limited by economics and fundamental speed-of-light latency. The evolving space economy will be shaped by converging technologies and the immutable physics of communication delays.