A simulation of the US economy projects that AI will break the per-seat pricing model for business software and cut entry-level desk hiring in every scenario it tests. These are the projections of World 2030, a macroeconomic study site that runs the economy from late 2026 to late 2030. They are model output, not observed data.
The study models nine actors. Some are obvious (the AI labs, the Fed, governments, workers), others less so: utilities, universities, startups, big employers, and the banks and insurers that finance and underwrite everyone else. It advances them in six-month turns through three paths. One is a shock-free Baseline. The other two add a 2028 shock: Automated Assurance, and a Credit Crunch with no bailout. The site says the goal is to find the invariants, the things that hold in every run, and the boundaries of what can change.
The software result is the starkest. From 2027, the study projects, AI startups undercut business-to-business contracts while large employers force deep discounts at renewal. The result is a cut of 50 to 67 percent in the value of older per-seat software firms, in every branch. Its Legacy Seat-SaaS Index (2026 equals 100) ends 2030 at 52.6 in Baseline, 32.0 in Assurance, and 40.4 in Crunch.
The jobs result runs in parallel. Corporate buyers, the study argues, cut software bills and entry-level payrolls well before they clean up old databases. Campus recruiting in the US freezes, down about 40 percent, while code review moves to offshore centers in India and Latin America, or to automated provers in the Assurance branch. US entry-level desk hiring lands at 61.4 percent of its 2026 level in Baseline, 58.0 percent in Assurance, and 63.6 percent in Crunch. Insurers’ human sign-off rules do not protect these domestic jobs.
Cheap AI does not rescue household budgets, according to the study. Low-cost inference and open-weight models drag the consumer digital slice of CPI, about 5 percent of the basket, to annual deflation of roughly 4 to 5 percent by 2030. The other 95 percent, physical services such as hospital nursing and housing, stays near 3 percent a year. Headline inflation holds around 2.5 percent, so the Fed stays at 3.75 percent in Baseline and Assurance and stops cutting at 3.00 percent in Crunch.
Power is the third constraint. World 2030 projects that angry voters push state regulators to cap what households pay for electricity. Utilities then ask data-center developers for long take-or-pay contracts and cash escrows paid in advance for high-voltage transformers. Developers short on cash lose their place in line. Rich cloud companies take over, paying early for turbines at former coal sites, solar parks bought ahead of time, and leases on military airfields. Guaranteed AI power roughly doubles from 18 GW to between 32 and 41 GW, and the early shortfall disappears in every run.
The two shocks produce the study’s counterintuitive finding. In Assurance, insurers, regulators, and courts accept cryptographic execution proofs in early 2028. Migration costs fall 80 percent, 60 percent of legacy corporate IT gets modernized, and tech equities finish 2030 at 109.2, the only branch above 2026. In Crunch, gates on private-credit withdrawals plus Congress’s No-Bailout Act push bad debts into Chapter 11 auctions at once, and tech stocks end at 84.8. Baseline, where lenders convert shaky software loans into equity stakes in 2029, ends at 77.8. A fast wipeout beats a slow bleed.
Real GDP growth in 2030 reads 2.34 percent in Baseline, 3.05 percent in Assurance, and 2.12 percent in Crunch. The labor share of income falls from 56.5 percent in late 2026 to 55.5 percent in two branches and 53.5 percent in Assurance.
Treat all of this with the caution any scenario model earns. The study’s outputs follow from its own assumptions about how the nine groups behave, and the page names no outside validation. Still, the software and hiring results hold across every branch the authors built, which makes them the part most worth testing against your own plans. A software vendor still pricing by seat should model a renewal discount of 50 to 67 percent, and a company counting on a junior hiring pipeline should ask what replaces it.
Source: World 2030, the macroeconomic simulation study at world-2030.pages.dev, which lists no named authors or publication date and covers late 2026 to late 2030.