Three years ago, we argued that the traditional software-as-a-service (SaaS) model was dying. The culprit then was capability: as artificial intelligence began to execute the work it had merely organized, the logic of renting tools to humans cracked. That was the first death, a death by disintermediation, where value slid from the software layer into the hardware doing the real thinking.
Now, a second death is unfolding, driven not by what AI can do but by what it costs. Software companies, long accustomed to near-100% gross margins, suddenly face meaningful and fast-rising variable costs tied to AI compute. This is not a fringe theory; it is the invisible gravity warping the global economy, rewriting policy, and keeping politicians from Tokyo to New Delhi awake at night.
The Barrel and the Refinery: Software as Hardware's Downstream
The ongoing Silicon Shock has multiple parallels with the 1970s Oil Shock. In that era, crude producers—the oil majors—were largely insulated and frequently enriched. The converters downstream, refiners and petrochemical makers, took the blow. They had to buy a feedstock they did not control, convert it into something finished, and discover in a crisis that they had inherited all the volatility and almost none of the pricing power.
Software now sits in precisely that downstream seat. The token is the barrel. The model labs and semiconductor manufacturers are the producers, raising prices and signing capacity contracts with the serenity of people who own the well. For software companies, the situation is worse because of their history: the decades-long miracle of the flat subscription was only possible because the feedstock—compute—was effectively free. One cannot overstate the shock of moving to a world of meaningful variable costs.
This shift is creating ripples through the application layer. Companies from the largest to the smallest are grappling not only with costs but also with how to attract customers with the right pricing, manage rising customer anguish, and navigate a competitive landscape changing fast due to innovations in pricing models. The changes in how software is priced and sold will likely remain unsettled for a while. For investors, the implications extend beyond revenues and profits to questions of survival, as sudden pricing-related business decisions can waylay even established players.
The defensive habits that followed the 1970s oil shock—hedging desks, feedstock pass-through clauses, flexible plants that could switch inputs, take-or-pay contracts—were not strategy in any ambitious sense. They were the scar tissue of an industry that had been taught it no longer owned its own cost base. Software companies will need to develop similar scar tissue, but the path is uncertain. The buffet of fixed-price software is closing, and the industry must learn to live with the crack spread of AI compute costs.
This second death of SaaS has a larger body count than the first. The turmoil caused by a new cost item is not just about technology companies making more or less money; it is about its impact on everything macro globally for a prolonged period. As we noted in our analysis of AI supremacy's new frontier, the implications extend from silicon to electricity grids, reshaping industries across the Indo-Pacific.
For now, the market remains fixated on whether the AI trade is already priced in. That is a fine question for a screen but a poor question for an economy. We are studying a technological earthquake the way a gambler studies a scoreboard, fixated on the ticker while the ground itself reorganizes underneath the table. The real issues are less glamorous and more useful: what happens when software, the great fixed-price machine of the last 30 years, starts burning a variable-cost fuel every time it thinks?
