The End-State Compute Portfolio: Land, orbital and bio-organic computing as complementary layers of AI infrastructure, and a strategy for Japan

AI compute is spreading from land-based data centres into two new settings: satellites in low Earth orbit and living neural tissue grown on chips. Public debate treats these as competitors, and as successive stages that will replace one another. This paper argues instead for a portfolio view. Each layer has a distinct profile of maturity, cost, energy use, repairability, latency, legal clarity and ethical acceptance, and each is suited to different work. Using public reporting to November 2026, the paper compares the three layers on nine criteria, summarises their strengths, weaknesses, opportunities and threats, and sets out an illustrative life-cycle scenario in which land remains dominant, orbit takes a niche of roughly 10 to 15% of suitable workloads from the early 2030s, and biology arrives as a specialist co-processor around 2040, with DNA as a long-term archive. It maps the main actors by layer and region and proposes a differentiated strategy for Japan: compete and host on land, serve and supply in orbit, and lead on science and rules in biology. It closes by linking the portfolio to the human capability it requires, namely people certified to use each layer for specific tasks within defined boundaries, which connects to the author's earlier work on clean-engagement certification and certificate-to-employment pathways. Scores and scenarios are the author's illustrative assessments. Declared interests: the author is the founder of GCTI, Managing Member of E.M.B.O.A LLC and Representative Member designate of Emboa Japan GK (in formation). Discussion paper, version 1.0.