Software vendors like Salesforce are getting left behind in the AI boom, as most AI spending at this point is going toward hardware or cloud infrastructure
Context & Ripple Effects
Earlier coverage projected broad growth in AI outlays across software, hardware, and services, but this report identifies where the early budget concentration is landing: infrastructure rather than application vendors. The contrast matters because the earlier forecast for AI spending across the stack did not imply that each layer would benefit at the same time.
Cloud providers were already under pressure to serve AI demand, creating an opening for on-premises hardware suppliers as well as hyperscalers in the push to meet AI capacity demand. This report places Salesforce and peer software vendors on the less-favored side of that near-term allocation.
First-order effects
- Salesforce and other software vendors have less direct exposure to the current wave of AI spending, while hardware makers and cloud-infrastructure providers capture the immediate budgets.
- Enterprise AI investment is being directed first toward the compute and hosting layer needed to run AI workloads, delaying the spending uplift software vendors might expect from AI adoption.
Second-order effects
- Software vendors face stronger pressure to show that AI features translate into paid product demand, rather than relying on AI enthusiasm alone; cloud and hardware suppliers gain greater leverage over the pace of deployment.
- The imbalance reinforces the capacity build-out already visible among AWS, Azure, Google Cloud, and hardware providers, as cloud demand for AI workloads pulls spending toward infrastructure.
Third-order effects
- If spending continues to concentrate at the infrastructure layer, AI economics may increasingly favor firms that control compute and cloud distribution, while application vendors compete on access to those inputs and customer distribution.
- The pattern is not necessarily permanent: broader AI spending was forecast to include software and services, but software’s share depends on deployments moving from infrastructure procurement into repeatable business applications.
The trend: AI’s initial commercial cycle is infrastructure-led, with value moving toward software only as capacity investment becomes deployed applications.