Everyone is obsessed with model intelligence.
I’m increasingly obsessed with model economics.
A $1B operating profit quarter for Anthropic validates that AI isn’t just growing, it’s becoming highly profitable.
The companies building the independent AI compute layer stand to benefit enormously.
$NBIS $CRWV $IREN $CIFR
“Power is the biggest constraint for growth.”Dion Harris, NVIDIA, speaking at the #RAISE summit.
That quote perfectly captures where the AI race is today.
For years the constraint was better models.
Today, it’s access to power.
Every new AI data centre is ultimately competing for the same scarce resource: reliable electricity. GPUs can be manufactured, data centres can be built, capital can be raised, but without power, none of it matters.
I believe this is one of the most underappreciated investment themes of the next decade.
It’s why I’m so excited about the convergence of:
⚡ Power generation
🔋 Grid infrastructure
🏗️ AI data centres & neoclouds
🤖 Agentic AI
The winners won’t simply have the best models. They’ll have the ability to deliver AI at scale because they secured the one resource everyone else is fighting for.
That’s exactly why we’re building one of Africa’s largest AI data centres, and why I’m increasingly bullish on the companies enabling the AI power buildout.
$VST $CEG $TLN $NRG $OKLO $TE $BE
AI is no longer just a software story.
It’s an energy story. And we’re only at the beginning. 🚀
One of my favourite slides from NVIDIA’s presentation at #RAISE
It perfectly captures where AI is heading.
The biggest unlock isn’t just better models. It’s dramatically lowering the cost of serving AI agents at scale.
NVIDIA is showing up to 35x higher throughput for high-interactivity agentic workloads with Vera Rubin NVL72 + Groq LPX.
That changes the economics entirely.
When inference gets cheaper:
• More AI agents become economically viable.
• Companies can deploy AI to every employee, not just power users.
• Entirely new software categories emerge.
• Demand for compute increases rather than decreases (Jevons Paradox in action).
This is why I remain incredibly bullish on AI infrastructure.
The winners won’t just be the model providers. They’ll be the companies building the compute layer that powers the agent economy.
$NVDA $NBIS $CRWV $CIFR $IREN
We’re still in the early innings of the AI infrastructure buildout. The agent economy is only just getting started. 🚀
China’s reported move to restrict overseas access to its leading AI models may seem bearish for AI at first glance. I actually think it’s bullish for the AI infrastructure trade, and particularly for the neoclouds building the compute layer, such as $NBIS $CRWV $CIFR $IREN.
Over the last 18 months, one of the biggest tailwinds for neoclouds has been the explosion of open-weight models. Companies no longer need to rely exclusively on the frontier labs. They can run DeepSeek, Qwen, Kimi and other state-of-the-art models on their own infrastructure, driving demand for GPU capacity outside the hyperscalers.
If access to China’s best models becomes restricted, it doesn’t reduce demand for inference. It fragments it.
Enterprises will increasingly need trusted, sovereign AI infrastructure with access to multiple model ecosystems. That shifts value further down the stack, from the model itself to the compute, networking and infrastructure that delivers it.
AI is starting to look a lot like semiconductors:
The models become strategic assets.
Access becomes geopolitical.
Sovereign AI becomes a competitive advantage.
I think this is another step towards a world where every region wants its own AI infrastructure, not just access to someone else’s.
That’s a very constructive backdrop for the neocloud thesis, and one of the reasons we’re so excited about what we’re building with Alpha Ticker and the upcoming Nucleus Fund. 🚀
Every major AI release expands what’s possible.
GPT-5.6 isn’t just another model upgrade—it’s another step toward AI systems that can reason, research and make increasingly sophisticated decisions.
That’s exactly why we’re building the Nucleus Frontier Fund.
Not to invest in AI hype, but to identify the companies creating the infrastructure, platforms and technologies that will define the next decade.
The pace of innovation is accelerating.
Excited for what’s coming.
Congrats to the OpenAI team 👏
#AI #Investing #FrontierTech #NucleusFund
Spent today at MACHINA in Paris exploring the future of humanoid robots.
We’re entering a new era where AI doesn’t just generate text - it interacts with and changes the physical world.
As we build the Nucleus Frontier Fund, we’re spending time on the ground understanding the technologies that could define the next decade.
The future is closer than most people think. 🤖
🚨 MARKET ALERT: The "Great Divergence" is here. 📉📈
As 2025 closes, we face a clash of titans: The Federal Reserve pumps liquidity just as the Bank of Japan prepares to hike.
Volatility is GUARANTEED. Here's your strategic roadmap for Dec. 🧵
⚠️ *DYOR - Not financial advice. Consult a professional.*
#StockMarket #Fed #BoJ #Trading
6/ **Wildcard: Supreme Court** ⚖️
Legal battle over Fed independence. If Exec can fire Fed officials → bond yields spike, Dollar hurt, Yen unwinds faster.