AI, Restacked: How the Frontier-Model Race is Rewiring the Software Landscape (Stifel)

Summary

• Frontier labs, AI-native vendors, and incumbents are competing directly for the same user interface, workflow, and pricing logic.p.2
• ChatGPT has reached 900M+ weekly users as software stack architectures shift toward a model-first, headless environment.p.4
• AI-native private software vendors command valuations of 30x + EV/Sales.p.4
• Incumbent software providers are trading in low single digits (LSD) EV/Sales and are down 30-50% YTD.p.4
• AI-native software represents a structural re-platforming where human workflows are passed off to agents.p.5
• The next decade of software compounders is expected to be born within the next 24 months.p.5

Key insights

1. Artificial intelligence is expanding the total software market rather than replacing it.p.2
2. ChatGPT active adoption has scaled past 900M+ weekly users.p.4
3. AI-native software vendors trade at valuation multiples exceeding 30x + EV/Sales.p.4
4. Incumbent software valuations have declined 30-50% YTD down to low single-digit multiples.p.4
5. Software workflows are structurally shifting from human execution to autonomous AI agents.p.5
6. The next 24 months will define the next decade of leading software compounders.p.5
7. Public markets have already adjusted valuation multiples to account for AI-native disruption.p.5
8. A rapidly broadening competitive landscape puts nearly all incumbent software vendors in direct risk.p.5