Hello Betamax, It's easy to forget when you work in or even around tech that the software we use has constraints that another funding round can't cure. It's physics, stupid. The AI models we prompt need GPUs, the cloud servers our websites are hosted on need data centers, our ecommerce deliveries need drivers - the list goes on. As much as Big Tech and some VCs might want to pretend otherwise, tech will always be limited by the physical realities we live in.  Putting data centers in space or covering the Earth in servers sound like ideas out of sci-fi, because they are. Digital businesses still run on very real, very finite resources. Reading about Indian quick commerce player Zepto's retreat from "10-minute deliveries" is a case in point - even if it can still get orders to customers' doorsteps absurdly fast. The company dropped the claim from its branding under scrutiny, but groceries still arrive in five to 10 minutes in many areas. Zepto has raised hefty sums to sustain its time-bending service, but an IPO is approaching, and it faces physical realities. To make the numbers work, the company is shifting toward owning inventory - a capital-heavy, operationally messy, and decidedly "old economy" move. This is what "it's physics, stupid" looks like in practice, as every tech company has to deal with atoms and neutrons as well as dollars and cents. Warehouses cost money. Power grids have limits. Drivers get tired. Margins don't bend forever. For years, abundant capital made it possible to ignore these constraints, but that era is ending. Whether it's AI firms racing to secure compute, delivery startups struggling with unit economics, or cloud providers grappling with energy costs, more companies are discovering the same uncomfortable truth: Software scales easily, but physical systems don't. That doesn't mean ambition is misplaced. It means innovation works best when it's built on solid ground and on goals that reality can support. Peter Cowan, engagement editor |