Superintelligence Full-Stack Bet
Mao Geng@gengmao
−11.1%
since Oct 7
The idea
I believe AI is on track to surpass individual human intelligence within a year or two, with gains compounding hard toward 2030, and that this overshoots even bullish forecasts for compute and power demand. This basket leans into the bottlenecks first, chips, memory, and networking, then the turbines, nuclear, and grid gear feeding data centers, rounded out with the public companies best positioned to win the race itself. I keep a small cash buffer to add on volatile dips, and I'd rethink this if the pace of AI progress clearly stalls out.
What it owns
- If artificial intelligence advances faster than expected, the chips and manufacturing capacity required to train and run frontier systems will remain scarce and expensive. · Advanced Micro DevicesIt serves as the only credible alternative processor source to dominant competitors, and growth depends on winning actual orders from major artificial intelligence labs rather than design wins alone., Micron TechnologyIt manufactures high-bandwidth memory that artificial intelligence chips require to achieve performance, and supply remains constrained relative to rapid demand growth.25%
- Public companies with direct ownership stakes in or revenue contracts with frontier artificial intelligence labs capture the most value if those labs achieve major breakthroughs. · Oracle CorporationIt leases massive compute clusters to frontier artificial intelligence labs under long-term contracts, positioning it as infrastructure landlord to the industry.20%
- Data centers training frontier artificial intelligence models require electricity volumes the electrical grid was not designed to handle, creating sustained demand for new power generation and distribution infrastructure. · Constellation EnergyIt operates the largest nuclear fleet in the country and sells carbon-free power to data centers that artificial intelligence operators prefer for long-term capacity agreements., Vistra CorpIt generates power in regions where artificial intelligence data centers are driving electricity demand, converting grid strain into higher wholesale prices and revenues.18%
- Cash36%