The Only 9 Themes I Need for the Next Decade
Most retail investors track 50 stocks.
They watch CNBC. They check X and Stocktwits.
They have alerts on 20 tickers.
Their watchlist is a graveyard of names they bought, ones their friends mentioned, and ones some podcast hyped six months ago.
This is how you lose.
Wealth doesn’t get built by tracking more. It gets built by tracking less, and tracking the right things.
I track nine disruptive themes. Every day.
Everything else is noise.
1. AI Infrastructure.
The GPU compute build-out.
The most obvious theme on the board. Also the most underestimated.
Every AI model needs compute. Every compute cluster needs data centers.
Every data center needs power.
The companies building, leasing, and operating that infrastructure are absorbing capital faster than they can deploy it.
Hyperscalers spent $364 billion on data center buildouts in 2025.
They’re spending more in 2026. The neocloud operators are booked two years forward.
$NBIS (Nebius) – Microsoft-backed neocloud with a $19.4B Microsoft commitment.
$CIFR (Cipher Mining) – Bitcoin miner pivoting to AI compute. Cash-flowing transition story.
$IREN (IREN Limited) – $9.7B Microsoft cloud deal. $9.7B Microsoft cloud deal. 150,000 GPU fleet target by end-2026, GB300 Blackwell systems deploying through Childress.
Simple way to think about it: Sell the picks during the gold rush. Each of these names is selling a different kind of pick.
2. Photonics
AI's bandwidth bottleneck.
Compute scales. Bandwidth doesn’t.
As AI clusters grow, the wiring between GPUs becomes the bottleneck.
Photonics solves that.
Optical interconnects are the only way to move data fast enough inside modern AI training racks.
$LITE (Lumentum) – Largest pure-play optical components company.
$AAOI (Applied Optoelectronics) – Hyperscaler optical transceivers. The AI tailwind hit them hardest in 2025.
$AXTI (AXT Inc) – Indium phosphide substrates. Every optical chip starts here.
$CIEN (Ciena) – The networking layer that connects data centers across geographies.
Simple way to think about it: Without photonics, there’s no AI cluster. These companies make the wires that light moves through.
3. Space
The launch and satellite broadband economy.
Launch costs have dropped 95% in the last decade.
Satellite broadband is becoming the new global utility.
Government contracts are exploding. The space economy is real, not science fiction.
$RKLB (Rocket Lab) – The next SpaceX. Neutron launch vehicle plus growing space systems business.
$ASTS (AST SpaceMobile) – Direct-to-cell satellite broadband. AT&T and Verizon partnerships.
$PL (Planet Labs) – Earth observation. The Google Maps of satellite imagery.
$LUNR (Intuitive Machines) – Lunar economy infrastructure. NASA partnership.
Simple way to think about it: The next economy isn’t only on Earth. A growing chunk of it sits 200 miles above the surface.
4. Robotics
The physical AI sensing layer.
AI is moving out of the data center and into the physical world.
Humanoid robots. Autonomous vehicles. Industrial automation. All of them need to see and feel.
The sensing layer is where the asymmetric upside lives, not the headline-grabbing robot company.
$OUST (Ouster) – LiDAR pure-play. Solid-state digital LiDAR. The cheapest scale-ready sensor.
$AMBA (Ambarella) – Edge AI vision processors. The brain behind every smart camera.
$VPG (Vishay Precision Group) – Force sensing for cobots. The least-known, most asymmetric robotics name.
Simple way to think about it: Every robot needs to see and feel. These companies build the eyes and the skin.
