newsroom
$META · Leopold Aschenbrenner - 2027 AGI, China/US super-intelligence race, and the return of history
now playing · $META
$META···bullish· mediumleopold aschenbrenner
Meta's lagged open models could accelerate broad AI diffusion
Open-sourcing models behind the frontier could preserve civilian innovation while restricting only the most dangerous dual-use capabilities.
$NVDA···bullish· high· posleopold aschenbrenner
Nvidia was the only pure AI-beta trade before the boom
Nvidia offered uniquely concentrated AI exposure: each successive increase in AI compute demand translated directly into material revenue growth, unlike suppliers whose AI busines…
$TSM···bullish· high· posleopold aschenbrenner
TSMC enters the AI sweet spot as exposure turns material
TSMC's AI revenue mix is expected to reach roughly 20% by 2027, making another demand doubling large enough to move the overall business and broaden the AI trade beyond Nvidia.
$GOOGL···bullish· medium· posleopold aschenbrenner
Google could explode once AI revenue crosses $100 billion
Alphabet becomes a compelling AI-beta investment once AI revenue is large enough to reshape the consolidated business; Aschenbrenner sees a path to $100 billion of AI revenue and…
$NVDA···bullish· highleopold aschenbrenner
Industrial AI buildout puts Nvidia on a trillion-dollar demand curve
Frontier AI requires exponentially larger clusters, power plants and eventually new semiconductor fabs rather than merely more software. If training compute continues scaling at i…
$MSFT···bullish· highleopold aschenbrenner
Microsoft Office distribution could unlock 100 billion dollars of AI revenue
Microsoft can finance frontier-scale clusters because its installed Office base offers a direct path to enormous AI monetization. Selling a 100 dollar monthly AI add-on to roughly…
$MSFT···bullish· highleopold aschenbrenner
Microsoft can finance the hundred-billion-dollar race to AGI
Microsoft's balance sheet and access to deep US capital markets give it the capacity to finance hundred-billion-dollar AI clusters, a scale private laboratories may struggle to ma…
Leopold warns OpenAI repeatedly sidelines safety and security commitments
OpenAI publicly makes major safety and security commitments but repeatedly fails to prioritize them when resources and trade-offs become real, creating material governance and exe…