2025 was a test of discipline. Our Global Tech 15 returned +28.02% vs +16.35% for the S&P 500, but the bigger lesson was this: markets are becoming less forgiving. Valuation and execution matter again. I published our first Annual Review for accountability, not promotion. 👉

China’s first major LLM company just went public. Zhipu AI (Knowledge Atlas Technology): • $558m IPO in Hong Kong • Shares up ~15% on debut • Valued at ~HK$4.3bn • 70% of proceeds going to AI R&D One of China’s “AI tigers” is now listed.

Anthropic is reportedly raising $10B at a $350B valuation. That’s nearly double its value in three months. This isn’t hype capital. It’s long-duration money betting that AI leadership now comes down to compute, scale, and execution, not just models. The gap is widening.

AMD unveiled its Ryzen AI 400 processors at CES 2026. • First Copilot+ desktop chips • Up to 60 TOPS NPUs • Modest CPU and gaming gains • Focus still on “AI PC” positioning $AMD

Gold has always been money

Nvidia’s real advantage isn’t GPUs anymore. It’s full-stack control: • compute • networking • storage • software • systems design Competitors can ship chips. Few can ship platform economics. #AI #Semiconductors

Rubin is optimised for agentic and reasoning models, not just raw training. That matters because the bottleneck in AI is shifting from building models to running them efficiently at scale. Inference is becoming the real cost centre.

Nvidia’s Rubin platform isn’t just a faster chip cycle. It’s an attempt to reset the economics of AI itself. 4x fewer GPUs for training. 10x lower inference token costs. That’s not performance. That’s margin defense. #NVDA

Investors are tolerating weaker vehicle numbers because the bet has shifted. Autonomy. Robotics. Energy storage. That changes the risk entirely. If those bets slip, car sales won’t be enough to carry the valuation.

Tesla $TSLA losing its EV sales crown isn’t the real story. The market already knows car sales are slowing. What matters is that Tesla is no longer being valued primarily as a car company.

Nvidia’s #NVDA reported talks to buy another AI lab point to a deeper trend. The real bottleneck in AI isn’t chips anymore. It’s people who know how to build at scale. Capital buys speed. Talent buys advantage.

Wall Street is unusually aligned: most expect US stocks to rise again in 2026. After years of resilience, pessimism has vanished. That alone is worth paying attention to. Markets don’t break on bad news. They break on crowded beliefs.

Me in three people when you tell me that tech stocks are "risky"

I break down what’s really driving this move and why it matters for 2026, full piece here:

Gold keeps rising while mega-cap tech and crypto pause. That isn’t panic. It’s capital moving toward reliability. Less excitement. More balance. That shift usually matters more than headlines.

This is less about chips and more about long-term control. Full analysis here:

Nvidia $NVDA doesn’t need AI hype to win. It needs customers to stay locked into its ecosystem. Recent deals suggest that’s exactly the strategy.

Learn more here: https://t.co/nhoSi7qPYy

“Pay yourself first” is less about money and more about discipline. Systems beat willpower. Consistency beats intensity. Over time, that’s what compounds.