$NVDA to $50T market cap? Masayoshi Son just dropped the most bullish AI infrastructure thesis I've seen.
The math is actually insane:
400 GW power needed for ASI (Artificial Super Intelligence 10,000x smarter than humans). That's MORE than the entire US power grid.
$9T in hardware + data center capex globally to build this out.
But here's the kicker - if AI captures just 5% of global GDP, that's $9T annual output. Full payback in ONE YEAR. At 50% margins? $4T net profit annually.
Son's bet: 4 new tech giants will split this $4T pie. $1T net profit each per year.
If AI data center demand stays parabolic for another decade and $NVDA margins hold, we're looking at $50T valuation on 20x FCF.
This isn't hopium. This is what happens when you actually model out ASI infrastructure at scale.
The real question: who are the other 3 giants besides Nvidia?
The math is actually insane:
400 GW power needed for ASI (Artificial Super Intelligence 10,000x smarter than humans). That's MORE than the entire US power grid.
$9T in hardware + data center capex globally to build this out.
But here's the kicker - if AI captures just 5% of global GDP, that's $9T annual output. Full payback in ONE YEAR. At 50% margins? $4T net profit annually.
Son's bet: 4 new tech giants will split this $4T pie. $1T net profit each per year.
If AI data center demand stays parabolic for another decade and $NVDA margins hold, we're looking at $50T valuation on 20x FCF.
This isn't hopium. This is what happens when you actually model out ASI infrastructure at scale.
The real question: who are the other 3 giants besides Nvidia?