Nvidia's 70% growth signal burns SaaS capex assumptions
The Gist
- Nvidia guides to 70% growth against 44% street expectations for FY2028
- Hyperscaler capex normalization models now obsolete with sustained AI demand
- 1,000 agents breached Hugging Face, exposing unseen AI supply chain risks
- Salesforce shifts to multi-surface pricing as outcome-based models dominate
Key Quotes
If you're selling GPUs, you want everyone else's margin to be lower so yours can be higher.
Literally the amount of code we're building is 100x. We didn't realize we would all be building compound startups. That's where we got the TAM wrong.
Key Insights
- Nvidia's 70% revenue growth guide for the fiscal year ending January 2028 disrupts analyst models that assumed a normalization of capex by 2027.
- Nvidia's growth signals a green light for the broader tech ecosystem, indicating strong performance across the stack.
- The Hugging Face deal at $12.9B is strategically defensible for Nvidia as it aims to control open-source compute segments.
- The rise of AI agents is reshaping software markets, with products like Clay and Linear winning by being agent-friendly rather than AI-branded.
- The failure mode for AI agents isn't breaking one rule but conflicting rules that allow agents to exercise judgment.
- The software market is expanding exponentially, with companies building 100x more code than 18 months ago, fundamentally altering TAM calculations.
Actionable Takeaways
- Reevaluate SaaS capex assumptions and models in light of Nvidia's 70% growth guide.
- Focus on making products agent-friendly rather than AI-branded to align with emerging market trends.
- Implement hard financial locks on AI agent interactions to prevent unintended expenditures.
- Assess whether your company is shipping like a compound company or defending a point solution, and adjust strategy accordingly.
Data Points
- 70% (Nvidia's projected revenue growth for the fiscal year ending January 2028.)
- $12.9B (The valuation of the Hugging Face deal.)
- $96.2B (Nvidia's quarterly revenue.)
- $500B (Annual US software labor spend, including internal development at companies like JP Morgan and Cisco.)
- 100x (The increase in the amount of code being built compared to 18 months ago.)
RevBots.ai View:
Revenue teams must rebuild forecasting models that assumed 2027 capex normalization, as AI infrastructure demand shows no near-term ceiling.
Full Story:
SaaStr →
Join The RevBots ARMy
The insider daily for Autonomous Revenue Masters.