Elon Musk announced on March 11th that he’s going to replicate every software company on earth. It sounds absurd. But it’s hard to ignore entirely.
His new project, Macrohard, uses superintelligent AI and Digital Optimus to replicate existing software function by function. If AI can replicate software across every industry, competitive advantage needs to be redefined.
So what’s left? Everyone says data. But turning that into an actual strategy is anything but simple.
The data you’ve already collected may not be a moat
Large manufacturers sit on staggering amounts of data — production logs, driving records, maintenance histories, sensor feeds.
In this, executives in these companies would ask “are we sitting on an undiscovered gold mine?” But no one could agree on what to do with it. One camp said: understand how deeply the data is tangled first, then refine and structure it. Without that precedent, we just put garbage in, getting garbage out. The other pushed back: run experiments now, learn fast, move quickly rather than waiting for perfect structure.
Both were reasonable — and shared the same assumption: data without direction isn’t an asset. It’s dead weight.
Tesla didn’t collect data, but designed for it
Tesla’s early FSD was rule-based. Then, with FSD v12 in early 2024, Tesla switched to end-to-end AI — effectively discarding everything built before.
Musk appears to have been planning for this from the beginning. As far back as 2016, Tesla ran “shadow mode” across its global fleet, quietly collecting real-world driving data from every car on the road. By the time v12 arrived, it had billions of video frames to learn from.
Tesla’s moat isn’t its technology. It’s the data architecture it designed years before anyone else knew what to design for.
A forest takes years to grow. You have to plant now.
A gold mine is already there. You find it, you dig. A forest is different. If you don’t plant today, there are no trees in ten years.
Many companies assume the data sitting in their systems will yield gold if properly extracted. But data accumulated without a clear sense of what value it’s meant to create doesn’t transform through effort alone. That’s not a strategy. That’s alchemy.
Work backwards instead. What customer value are you building toward? What AI application gets you there? What data does it need to improve? What seeds do you need to plant right now?
A tree you never planted can’t grow, no matter how much you invest later. Are you planting the right seeds today — or still looking for a gold mine?