Synthesis: Starbucks and monday.com Show the Same AI-Labor-Substitution Bet From Opposite Sides of the Vendor Relationship
The connection
Two 2026-07-26/28 items, read together, show the same underlying economic shift — AI-assisted coding lowering the labor cost of building software enough to change build-vs-buy math — playing out on both sides of an enterprise software vendor relationship simultaneously:
- Starbucks builds AI-generated internal software to replace vendor systems — a buyer deciding to stop paying Microsoft Dynamics 365 and IBM TRIRIGA licensing fees and build in-house replacements instead, as part of a $2B cost-cutting push against a ~$400M/year software budget. The market reacted immediately: IBM -3%, ServiceNow -3.5%, Salesforce -4% on the report.
- monday.com Restructures to Refocus Around AI Work Platform, Cuts 20% of Workforce — a seller of exactly this kind of enterprise software cutting 630 jobs (~20% of headcount) while reaffirming revenue guidance and raising its operating-margin outlook, explicitly repositioning around "AI agents working alongside employees" rather than seat-based licensing.
Why this wasn't visible before
Starbucks is tagged business/solution-architecting and filed as a cost-cutting story about a coffee retailer; monday.com is tagged business-strategy/agents and filed as a SaaS-industry layoff story. Nothing connects a customer's build-vs-buy decision to a vendor's own headcount response to the same underlying force, even though they're the same market dynamic viewed from opposite ends of the transaction: AI coding tools are simultaneously making it cheaper for enterprises to build what they used to buy, and making it possible for SaaS vendors to serve their remaining customers with fewer people.
What this suggests
- Both companies are pricing labor at a lower marginal cost than a year ago, but the effects point in opposite directions for revenue: Starbucks's bet reduces vendor revenue directly (fewer licenses bought industry-wide); monday.com's bet tries to preserve revenue with fewer employees by reframing the product itself around agents. If Starbucks-style build-over-buy generalizes, monday.com's cost-cutting alone won't be enough — the vendor also needs the AI-native product repositioning to defend against exactly the substitution Starbucks is demonstrating.
- This is the same mechanism the token-cost-discipline synthesis tracks at the model-tiering level, but one layer up the stack: instead of "use a cheaper model for an easy task," it's "build the whole application in-house because AI coding makes that labor cheap enough to compete with a vendor's fixed licensing fee." Same underlying force (AI compressing labor cost), different unit of substitution (a token vs. an entire enterprise system).
- Worth revisiting if a clear pattern emerges over the next few quarters — other CIOs following Starbucks (per the Forbes framing, "a warning shot for every enterprise software vendor") would be the concrete evidence that this isn't a one-off, and would directly inform how to read future SaaS-vendor restructuring announcements like monday.com's as defensive rather than purely strategic.