Place every rack
Hand-build halls GPU by GPU, node by node, cluster by cluster - in crunchy isometric pixel art.
Place the racks. Feed the gigawatts. Sign the contracts. Keep two thousand GPUs from cooking themselves in the desert sun - and getobscenely rich doing it.

Hand-build halls GPU by GPU, node by node, cluster by cluster - in crunchy isometric pixel art.
Power & cooling redundancy sets your Uptime Tier I–IV. Cut a corner, watch a contract breach in real time.
Sign compute contracts from scrappy startups to hyperscalers. Bill per GPU, per tick, around the clock.
Earn your way from Hawk H100 to Gale-Hawk GH300. Newer chips, fatter margins, louder hairdryers.
Buy land, drop a hall, fill it with GPU clusters. Land and layout matter - sprawl costs you uptime.
Wire up grid, generators, UPS and chillers. Stack redundancy to raise your Uptime Tier - or gamble.
Sign jobs that fit your tier. Hold the SLA and the money prints; breach it and the penalties bite.
Reinvest. Bigger halls, newer GPUs, tighter SLAs - until one desert plot isn't nearly enough.
→ Then do it again, but obscenely →

Title screen

Main menu

Site briefing

Build-out

Site complete

Insolvent
$6.0M of seed capital. One dusty plot. Grow it into a $40M compute empire - or get repossessed trying.