Technology Cycles Split Winners and Losers the Same Way

The technology changes each decade. What separates the minority that wins is sequencing: they test the commercial hypothesis while being wrong is still cheap.

By Harinderpal Hanspal on January 2026. Updated October 2026

Better technology is not what separates the two groups. The failed majority's technology usually worked.

Both groups build the same things; the minority confirms buyer, price and channel first Heavy capital flows to technology that works and splits into two lanes. A small minority confirms buyer, price and channel, then builds the sales motion, channel and product. Most build first, find the wrong buyer, price or channel, and pay for the answer. SMALL MINORITY MOST Heavy capital goes to technology that works Confirm buyer, price and channel Build the sales motion, channel and product Build the sales motion, channel and product Find the wrong buyer, price or channel Pay for the answer
Both groups build the same things; the minority confirms buyer, price and channel first

Each transition drew heavy capital to technology that worked. A small minority confirmed buyer, price and channel before they built a sales motion, contracted a channel or tooled a product. Most found the wrong buyer, price or channel afterward and paid for the answer.

The early numbers for this cycle have the same shape. Gartner expects more than 40% of agentic AI projects to be canceled by the end of 2027. MIT's Project NANDA found, in a preliminary report, that about 5% of integrated generative AI pilots were extracting millions in value. CB Insights' March 2026 review of venture-backed shutdowns found poor product-market fit behind 43% of the failures with identifiable reasons, close to the 42% it found for no market need in 2014. The methods differ, so these do not form a precise series.

Industrial work raises the price of getting the order wrong. A mistaken bet there spends fleet capital, long procurement relationships, operational technology (OT) integration work and frontline trust, and each is slow to rebuild.

Go deeper: Too successful to stop, too unproven to scale sets out the pattern and the sources, and when an initiative is not delivering covers what to do once it shows.