Salesforce blames its Claude addiction for denting profit margin guidance
- ID
- 21080
- Status
- summarized
- Published
- 03 Sep 2026, 10:44 PM
- Fetched
- 03 Sep 2026, 11:36 PM
- Provider
- The Register
- Category
- technology
- Original URL
- https://www.theregister.com/ai-and-ml/2026/09/03/salesforce-blames-its-claude-addiction-for-denting-profit-margin-guidance/5294219
- Source URL
- https://www.theregister.com/headlines.atom
Summary
- Score
- 7.5
- Created
- 03 Sep 2026, 11:37 PM
- Tags
- Audience
- developersai_agent_userssaas_founders
What happened
Salesforce deputy CFO Mike Spencer told the Deutsche Bank Technology Conference that spending on Claude tokens prevented the company from raising its full-year operating margin guidance (20.1% vs Q2's 20.5%), after the company 'unleashed Claude in its R&D cycle' roughly six months ago. Salesforce is now shifting to 'refinement mode,' prescribing model choice per task rather than defaulting to the latest model, and is evaluating OpenAI, Cursor, Claude, and Grok across different use cases.
Why it matters
If a company of Salesforce's scale can't absorb token costs without denting margins, smaller builders should model AI spend carefully from day one. The concrete takeaway: don't default to frontier models for every task—Spencer says the 'large majority' of dev and stack work is fine with second- or third-generation models, reserving frontier models only where they're genuinely necessary.
Discussion angle
What's your actual cost-per-task breakdown across model tiers, and which tasks in your pipeline could move to cheaper models today without quality loss?