Last month, enterprise software giant SAP suspended most of its corporate travel and hiring. The reason wasn't a sudden revenue miss or a broader macroeconomic downturn. According to an internal email obtained by 404 Media, SAP froze these budgets to offset the soaring, uncontrolled costs of its own artificial intelligence investments.
The internal memo mandated that the company must "be disciplined in how we spend." The only exceptions to the freeze were travel and hiring directly related to AI. This cost-cutting scramble coincides with SAP rolling out a newly created, company-wide AI tool to its global workforce—a deployment that, as one employee noted at a recent global meeting, is massively driving up expenses.
The Enterprise AI Reality Check
This development from SAP shatters a pervasive narrative in the tech sector: the assumption that AI immediately acts as a massive, frictionless cost saver. The reality playing out inside major enterprises is starkly different.
The compute power, licensing fees, and development costs required to give an entire workforce open-ended access to generative AI are immense. As 404 Media highlights, rather than instantly shrinking budgets, AI costs are spiraling out of control for many organizations. To cope, some companies are now actively throttling their employees' AI usage, while giants like SAP are forced to ground their staff and halt hiring just to balance the books.
For businesses looking at the AI landscape, this is a profound reality check. Unmetered, company-wide AI rollouts are incredibly expensive.
What This Means for SMBs
If a multi-billion-dollar software giant has to freeze hiring and travel to afford its AI bills, small and mid-sized businesses need to pay close attention. The primary takeaway here is that SMBs cannot afford to blindly mimic enterprise AI strategies.
When deploying AI for administration and operations, SMBs should adopt a highly surgical approach to avoid the cash-burn trap SAP has found itself in.
Avoid the "Blank Check" Rollout Giving every employee an enterprise license for a general-purpose AI chatbot might sound like modernization, but it is often a fast track to spiraling software costs. Without clear guardrails, employees may use expensive models for basic search queries or non-essential tasks, driving up token usage and subscription costs with zero measurable return on investment.
Automate Workflows, Don't Just Buy Software Instead of company-wide software deployments, SMBs should focus on targeted AI automation. This means identifying specific, high-friction administrative and operational workflows—like invoice processing, CRM data entry, or customer intake—and building automations specifically for those tasks. When AI is tied to a specific workflow via API, the costs are predictable and strictly tied to the volume of work being processed. You can calculate the exact cost per transaction and weigh it directly against the human hours saved.
Meter and Measure The fact that enterprise companies are now having to "throttle" employee AI usage proves that unmonitored AI access is financially dangerous. SMBs need to ensure that their AI tools are measurable. Working with a fractional automation partner allows businesses to build backend systems that execute tasks efficiently behind the scenes, rather than relying on unpredictable, front-end usage by dozens of employees.
Artificial intelligence is a powerful lever for operational efficiency, but it is not magic. It requires compute power, and compute power costs money. By ignoring the hype and focusing strictly on targeted, measurable automations, small and mid-sized businesses can reap the operational benefits of AI without having to ground their own teams to pay for it.