There's a thesis floating around VC circles — attributed to Sequoia Capital — that goes something like this:
"The next trillion-dollar company will be a software company masquerading as a services firm."
At first, that sounds backwards. Aren't we supposed to be building scalable software? Isn't "services" a dirty word in startup land?
But sit with it for a minute. It's one of the sharpest observations about where AI is actually heading.
The $1 vs $6 Problem
Here's the economics that most builders get wrong:
For every $1 spent on software, businesses spend roughly $6 on services — the consultants, agencies, and freelancers who actually use that software to produce work.
Salesforce sells CRM software for $25/user/month. But companies spend 3-6x that on Salesforce consultants, admins, and implementation partners.
Adobe sells Creative Cloud for $55/month. But companies spend thousands on designers who use it.
The software is the $1. The work is the $6.
AI is about to collapse that ratio — and whoever captures the $6 wins.
Why "Sell the Tool" Is a Trap
The default instinct for technical founders (myself included) is to build tools. An AI writing assistant. A prompt library. A workflow builder. A dashboard.
Tools are comfortable because they're scalable. Build once, sell forever.
But here's what's actually happening in the market:
- Tools commoditize fast. There are 9,000+ AI tools on Product Hunt. Users are drowning in options.
- Customers don't want tools. They want the output the tool produces. Nobody wakes up wanting "an AI-powered content optimization platform." They want blog posts that rank.
- The switching cost is zero. If your tool costs $29/month and a competitor launches at $19, you're in a race to the bottom.
Meanwhile, the person selling finished deliverables — the actual work — has pricing power, stickiness, and a moat that's hard to replicate.
Tools vs. Work: A Quick Comparison
| Selling the Tool | Selling the Work |
|---|---|
| "Here's an AI writing assistant" | "Here's 50 blog posts optimized for your niche" |
| "Here's a prompt template pack" | "Here's your entire Q3 content calendar, written" |
| "Here's a logo generator" | "Here's your brand identity, done" |
| "Here's an SEO analysis tool" | "Here's your SEO audit with fixes applied" |
See the pattern? The left column competes on features. The right column competes on outcomes.
And outcomes always win.
The Hybrid Play
Now, I don't think it's purely binary. The smartest approach I'm seeing is a hybrid model: give people tools and finished work.
This is something we've been experimenting with at MidasTools. We have free AI tools people can use directly — text generators, formatters, utilities — things that deliver immediate value with zero friction.
But the bigger insight was packaging bundles of finished outputs. Instead of saying "here's a prompt template you can customize," the bundle approach delivers the actual artifacts people need — ready to deploy, not ready to tinker with.
The free tools build trust. The finished work captures the $6.
What This Means for Builders
If you're building in the AI space right now, ask yourself:
1. Am I selling the pickaxe or the gold?
During the gold rush, the pickaxe sellers did fine. But the people who sold claims to proven gold deposits did better. Your AI tool is a pickaxe. The output it produces is the gold. Sell the gold.
2. Can I productize the service?
The magic of AI is that it lets you deliver service-level outputs at software-level margins. You don't need a team of 50 consultants. You need a well-built system that produces consistent, high-quality work.
3. Where's the $6 in my market?
Whatever tool you're building, look at what people are hiring humans to do with that tool. That's your real market. It's 6x bigger than the tool market, and AI just made it accessible to small teams.
The Uncomfortable Truth
Most of us in the dev community are tool builders by nature. We love elegant APIs, clean abstractions, and scalable architectures.
But the market doesn't care about your architecture. It cares about the problem being solved — completely, not partially.
The Sequoia thesis isn't saying "don't build software." It's saying: build software that delivers the end result, not software that helps someone get to the end result.
That's a subtle but trillion-dollar distinction.
What do you think — are we heading toward a world where "AI tools" become invisible infrastructure, and the real product is always the output? I'd love to hear how others are thinking about this.
Top comments (0)