Graham would land in the $50–100B range using net asset value and GAAP earnings discipline. The $4.9B net loss, $41B accumulated deficit, and xAI burning $30B+ annually in capex would result in no margin of safety calculation even getting close to the IPO price. He’d value Starlink’s earnings power at maybe $60–70B and net out the rest as liabilities.
Buffett would be more generous on Starlink’s moat — real competitive barriers in orbital slots, regulatory approvals, and first-mover scale — but would penalize heavily for the dual-class governance structure and Musk’s unchecked control. He’d likely isolate Starlink as a $150–200B standalone business and treat everything else as either zero or a liability, landing around $150–250B total.
Lynch would be the most forgiving, crediting Starlink’s 50% revenue growth and doubling subscriber base as legitimate “fast grower” dynamics. But even he caps PEG at roughly 2x, and at $1.75T you’re pricing in 20 years of flawless execution. He’d call it a $200–350B company at fair value — a buy at one-fifth the IPO price.