> ## Content Index
> Fetch the complete content index at: https://wire.fourthweb.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# Companies Are Sitting on $50B in Unused AI Licenses
- URL: https://wire.fourthweb.ai/companies-are-sitting-on-50b-in-unused-ai-licenses/
- Published: 2026-10-05T21:31:31.000Z
- Updated: 2026-10-05T21:31:33.000Z
- Description: The gap between what AI can do and what companies actually do with it just became the biggest opportunity in enterprise software.
- Author: Travis Wright
- Tags: AI Agent Economy, AI Infrastructure, OpenAI, Anthropic

**The gap between what AI can do and what companies actually do with it just became the biggest opportunity in enterprise software.**

### The Summary

- [NEA partner Tiffany Luck argues that even if frontier AI development slows, enterprise adoption still has years of runway](https://www.bloomberg.com/news/videos/2026-10-05/ai-pacing-doesn-t-mean-slower-adoption-video?ref=wire.fourthweb.ai) because most companies haven't deployed capabilities that already exist
- Current models from [OpenAI](https://wire.fourthweb.ai/tag/openai/) and [Anthropic](https://wire.fourthweb.ai/tag/anthropic/) are already more powerful than what 90% of enterprises are using in production
- The value capture is shifting from model builders to companies that can translate model capabilities into actual workflows with ROI

### The Signal

Tiffany Luck from NEA is naming something most people feel but can't articulate: the frontier and the field are operating on different clocks. While everyone watches the race between OpenAI, Anthropic, and Google for the next GPT-5 or Claude-4, the actual economy is still trying to figure out what to do with GPT-4 level intelligence.

This isn't a technology problem. It's a deployment problem. Most companies aren't constrained by model capability, they're constrained by integration, change management, and the ability to measure whether an AI investment actually moved the needle. [Luck's point](https://www.bloomberg.com/news/videos/2026-10-05/ai-pacing-doesn-t-mean-slower-adoption-video?ref=wire.fourthweb.ai) is that even if frontier labs hit a wall tomorrow, there's a multi-year window where the returns come from companies that can close that gap.

> "The biggest AI opportunity isn't building smarter models. It's making existing intelligence actually work inside real companies."

What this means practically:

- Vertical AI companies with deep domain expertise suddenly have more runway than pure infrastructure plays
- The "AI wrapper" critique misses the point — wrapping is exactly the value-add when enterprises can't deploy raw capability
- Companies solving for data integration, workflow redesign, and ROI measurement become more valuable than companies chasing benchmark scores

The real tell here is what happens to venture funding. If Luck is right, the next wave of AI investment shifts from research labs burning $100M training runs to B2B software companies with 80% gross margins selling AI that already exists into industries that haven't touched it yet. That's a different risk profile, a different timeline, and potentially better unit economics.

Think about industries still running on Excel and email: logistics, construction, legal services, healthcare administration. They don't need GPT-6\. They need someone to take GPT-4 and make it work inside their procurement system without requiring a PhD to operate it.

### The Implication

If you're building in AI, this is your strategic fork in the road. You can chase the frontier and compete with labs that have billion-dollar budgets, or you can turn around and sell what already exists to the 95% of the economy that hasn't bought it yet. The second path is less sexy. It's also probably more profitable.

For enterprises, the message is simpler: the excuse that AI isn't ready yet just expired. The models are ready. The question is whether your organization is.

### Sources

[Bloomberg Tech](https://www.bloomberg.com/news/videos/2026-10-05/ai-pacing-doesn-t-mean-slower-adoption-video?ref=wire.fourthweb.ai)