BlockBeats News, June 16th - In the latest episode of the BG2 podcast, Fox and Brad Gerstner discussed the economics of deploying AI computation power in space, viewing it as a potential long-term option value within SpaceX's valuation.

BG2 is a tech investment podcast hosted by Altimeter co-founder Brad Gerstner and former Benchmark general partner Bill Gurley, highly recognized in the Silicon Valley investor community. Gerstner is the founder and CEO of Altimeter Capital; Andrew Fox (often nicknamed Foxy) is an analyst at Atreides Management.

Fox stated that, by his calculations, each Starship launch could correspond to approximately 5 terawatts of computing power capacity. Extrapolating further, the capital expenditure to deploy relevant equipment in space could be around $5 billion per gigawatt.

He compared this figure to ground-based data centers, mentioning that the capital expenditure on switching equipment, generators, transformers, buildings, power acquisition, and cooling systems in ground data centers could reach $20-25 billion per gigawatt. Therefore, if the orbital computation power plan materializes, some non-GPU infrastructure costs could be significantly lower than those on the ground.

Gerstner added from the perspective of overall computation construction costs, noting that today, deploying 1 gigawatt of computation power on the ground may require around $60 billion, with about $35 billion for GPUs and chips and around $25 billion for land, buildings, power, and cooling. He believes that these ground elements may also face inflationary pressures in the future, while in space, space, power, and cooling are nearly free in the economic model.

However, Fox also made it clear that orbital computation power is not a necessary condition for establishing SpaceX's IPO valuation. He believes this business is more like significant upside potential rather than a core premise of the current valuation.