The Great Inversion
The Margin Flip
Software margins are compressing. Hardware margins are expanding. The crossover is happening now. Why applied AI + hardware will outperform pure software over the next decade.
Pure Software
Features commoditize, open source eats value
AI + Hardware
Physics creates moats, AI adds 10-100x value
Value is migrating from abundant (software) to scarce (proprietary atoms + data). The companies that win will understand physics as well as algorithms.
Why This Matters
Proprietary sensor data from deployed hardware creates defensibility software alone can't match
Manufacturing learning curves mean unit economics improve dramatically with scale
LLMs can replicate software in weeks, but replicating AI + hardware takes 3-7 years
When Is Technology 'Good Enough' to Invest?
The hardest call in deep-tech investing is betting on companies where the technology works in proof-of-concept but isn't commercially viable yet. The gap between technically proven and economically deployable kills more companies than anything else.
Wait for Commercial Proof
Lab demos are not equivalent to production reality. There are too many variables (manufacturing, cost, reliability, scale) and it's better to pay 3x valuation for 10x less risk.
Bet on the Inflection
If you wait for proof, you are not doing venture. Returns come from seeing what others can't and first-mover advantages compound.
Both are right sometimes. The real question isn't "does it work?" It's whether the team can bridge the gap with available capital and time.
There's no formula. The best investors assess specific risks, evaluate if the team can retire them, and make a conviction-based probabilistic bet. Sometimes you override logic for exceptional founders or market timing. But mostly? If it's "maybe," it's no.
Ajay Jain
Founder & Managing Partner, SilverX Fund
22+ years in venture capital, sales/GTM, and deep-tech investing. Previously at Xseed Partners, T-Hub, Intel, AMD.