Every time I refresh LinkedIn these days, I see at least two posts about data centers. AI data centers. Hyperscale data centers. 100 MW data centers. Gigawatt data centers. Lots of numbers. Very few pictures. So here’s the mental picture I use when I try to explain it to someone outside the industry: A data center is basically a giant,

Almost every services or engineering company we evaluate has a slide called something like: Products. Accelerators. IP. Sometimes it’s a very impressive slide. Twenty products. Fifteen accelerators. A long list of proprietary frameworks and platforms. After looking at enough of them, I’ve stopped reading the list. I read what comes after it. Who is actually using it? How many customers?

When I took a long break from my career, one of the happiest things I did was rewatch Star Wars with my dad. He watched it in a theatre when it first came out, which I still think is a flex. 💪 Somehow, the obsession has now made its way to the next generation. I’ve been slowly turning my nephew

Why Cooling a Computer in Space Is So Hard Everyone assumes cooling a computer in space is easy. Space is cold, right? It’s actually the opposite. On Earth, a data centre cools itself with air and water carrying heat away. In space there’s no air to carry anything. A chip can only get rid of heat by radiating it out,

You Don’t Have a Capability Gap. You Have a Packaging Gap. I say this in almost every strategy conversation we run at ProdZen, and it’s still the hardest thing for founders and leadership teams to hear. Most companies don’t have a capability gap. They have a positioning and packaging gap. The question everyone asks: “what do we need to build

Beyond the Data Center Boom: Building Better, Not Just More I’ve been researching #DataCenters a lot these days. And the data center boom is honestly outstanding. Every other report seems to have another trillion-dollar number, a massive demand projection, or a 2050 scenario. I’ve probably read 15+ reports since last week on data centers, #SovereignAI, grid buildouts, AI infrastructure, and

Rent, Buy or Build? A Housing Lesson in Corporate Strategy When I moved back to India after 18 years in the US, I made three very different housing decisions over time. (2 are done and one is a dream) First, we rented. I needed speed and flexibility, and I didn’t want to lock up heavy capital before knowing where I

Grafting as a Growth Strategy: Build, Buy or Partner? A grafted branch can start fruiting much sooner than a tree grown from seed. I’ve known that for years. Just never thought of it as a business idea. My grandfather worked in the forest department, so trees, budding and grafting were always part of conversations at home. When I was a

1 GW by 2031: Karnataka’s Bet on Sustainable Data Centres.    I’ve been doing a lot of research on data centres lately, and somehow even our lunch conversations are now about them. Yesterday @Sahil Harwani asked me, “How’s the research going?” I said, “My ethical side keeps questioning it, but the economy and market around data centres is huge.” Then he

If you’re new to the data-center industry, start with these 10 terms. I’ve been researching this space a lot lately. Shoutout to @Kaustubh, whose review-call questions have basically become my data-center vocabulary list. Here’s the shortlist I’d hand to anyone starting from zero, split roughly between the business/infrastructure side and the engineering side. MW / GW – Megawatt / Gigawatt

When a smaller fund loses an investor to a bigger, more established one, it’s easy to assume it’s just about brand or scale. Often, the real moment it’s decided is quieter: an investor asks to see exactly how decisions get made, and a strong story doesn’t carry the same weight as something you can actually show. This isn’t about bigger

When a fund launches, it usually decides upfront how much of its total money to set aside – or “reserve” – for investing more in its strongest companies down the line. That number is a genuine, informed guess made before there’s any real evidence to go on. The tricky part isn’t deciding it – it’s finding time, later, to check