Investing

Anthropic

AI can move every KPI a business tracks. Companies pay the most for models they can trust with real work, and that is exactly what Anthropic sells. I build on these models every day, help people use them, and see the value they deliver.

SpaceX

For centuries people have pushed past what was called impossible. Space is where I hold that conviction hardest: in my lifetime we will reach places that were never in reach before. SpaceX made launch routine, and that is what opens the rest.

What it taught meA long lockup is a feature. It stopped me from selling into the first rerating, and the rerating I would have sold into was the small one.

Blue Origin

The second mover reaches the same market without paying for every first. Bezos funds it, and NASA and the Space Force anchor it. It runs leaner than SpaceX, which matters once the question shifts from whether this is possible to whether it can be repeated. It trades at a real discount to SpaceX on sales, which is why I own both.

Perplexity

Search is being rebuilt around answers instead of links. The winner is whoever people open out of habit. I use it daily and watched it become a habit for the people around me.

OpenAI

The leader sets the pace everyone else runs at. OpenAI has the distribution, Microsoft behind it and an IPO ahead. Part of my holding was sold out from under me by the platform, at roughly double what I paid. The rest I still hold.

What it taught meTrimming a winner is fine when it lets me hold the rest without anxiety. Position size drives behaviour more than conviction does.

Moonshot AI

Moonshot takes on the best frontier labs with a leaner team, and it is set on going further. With this much capability and adoption this wide, I believe each frontier lab will become the expert in its own domain. I want to be there for the one where Moonshot becomes the pioneer.

Revolut

Banking keeps changing its technology: paper money, then ATMs, then core banking, now the app. The job stays the same, making money easy for the customer. Revolut builds around that job from the core, and that is why I back it.

Polymarket

Betting markets and polls have existed for centuries. Polymarket combines polls, betting and crypto into an insight engine anyone can use, and puts a number on subjective judgement at scale.

SandboxAQ

AI is moving from generalist to specialist. AI for drugs, chemicals and materials is what actually moves humanity forward. SandboxAQ builds specialised large quantitative models (LQMs) to change these core fields.

Quantic

As the world speeds up, leaders need to step up with it. Quantic lets them do that without leaving their role, learning from anywhere through a new way of teaching. I am glad to be one of its graduates.

NSE

India is the fifth largest economy and grows faster than any other large one. Capital markets are best placed to capture that growth. I plan to hold through the IPO and well after it.

Lumapps

Lumapps leads the market in workplace collaboration. Picture one operating system that connects documents, wiki, chat, tasks and shifts, and uses AI to help everyone in the company. That is what Lumapps is.

Beekeeper

Frontline workers make up about 80% of the global workforce. Beekeeper connects them to their colleagues and employers. Six years of engineering there, then my first exit, is the best reward I could ask for.

PDRL

In the drone stack, the software keeps the margin. AeroGCS runs ground control for about half of India’s drone makers. I spoke to the founder before buying.

Garuda Aerospace

Drones are becoming ordinary equipment in farming, defence and industry. The makers who reach scale first take that market. I built autonomous drones at an EPFL spin-off, so I can judge the hardware myself.

Positron

When I invested, the bottleneck for AI growth was infrastructure, not intelligence. Picture a world where AI keeps getting better while tokens get cheaper and power use drops. Positron offers that today, at scale.

Indian Gas Exchange

India is switching to gas, and traded volume follows the switch. The venue collects on every trade whatever the price does. Third exchange in the record, same logic.

Nailinit

Beauty done at home is growing fast in India. Press-on nails ride that shift from a tiny base. Early, thin on cash, and a test of whether the fund can pick beauty.

Noklu

Functional drinks are arriving in India the way they arrived everywhere else. A studio-backed team gives it a better start than most. Added recently, and I have no read yet beyond the first update.

Laadle

Pet ownership in India is only now becoming a market. Dehydrated food is where that market starts. Same studio as Noklu, added at the same time.

NCDEX

Farm income formalises as India grows, and commodity trade moves onto exchanges. Same toll booth, different market. Smaller and slower than NSE, held for the same reason.

OYO

Indian travel demand grew straight through the company’s own hype cycle. OYO survived it and turned toward profit. A pre-IPO bet that the demand outlasts the story.

Fur Jaden

Indian buyers are trading up to branded luggage. Distribution decides this market, not design. Cash is tight here and the next round decides it.

Bummer

Innerwear is a big, dull, brand-led market. Bummer reached its first profitable quarter after cuts. The odds are against it, and I keep it listed because losers belong in the record too.

HDFC Bank

India’s borrowing cycle runs through its private banks. HDFC has the best return on equity and assets in that market, and it runs on process rather than on one person. I bought when it traded near its 2008 multiples with nothing wrong in the fundamentals. I hold it directly and through funds.

Epic Games

Game engines became the toolchain for film and simulation too. Unreal sits under all of it, and Fortnite is a platform rather than a game. I bought secondary shares when employees wanted liquidity.

Alphabet

Search pays for the research, and the research is now the product. Owning both the data and the models funds the race indefinitely. My rule is simple: buy below four times sales, start trimming between six and eight.

Berkshire Hathaway

Some companies are worth valuing on what they own, not what they promise. Berkshire is a holding company, so book value is the honest measure. Buy near book, sell near twice it, and let the float work in between.

Microsoft

A business that funds its own expansion never needs permission to grow. Return on capital is the proof. Buy zone is six to eight times sales, and past twelve I trim rather than add.

Nvidia

Every model trained or served rents the same silicon. In a build-out the tool seller gets paid first. I hold it knowing the growth is real and the spending is someone else’s, which is the part that can turn.

Bharti Airtel

India’s data bill only goes up from here. Two operators are left to collect it. I did not pick this one, but the fund holds the toll on Indian data and I am glad to own that.

Maruti Suzuki

The first car a new middle class buys is the one everyone else sells against. Maruti owns that slot and the service network that keeps it. It came through the fund, and being late to electric is the risk I accept.

Groq

Fast inference was the obvious bottleneck for agents. Chips built only for low-latency serving looked like the answer. I took the position, the company was bought out, and I took the exit.

What it taught meA quick exit on an AI hardware name taught me that in this sector the buyer often shows up before the product proves itself. Take the exit.