Commoditize Your Complements
Make the inputs you depend on cheap so you keep more of the value you create.
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 80%
This is a mental model for defending margin when your product's value depends on a technology or input controlled by someone else. Instead of paying rents to a proprietary supplier (or a platform gatekeeper), you invest heavily to make that input freely available to the whole market — open source, open standard, or subsidized — so no single vendor can extract pricing power over you. Meta used this twice: first with the Open Compute Project (open-sourcing data center designs to drive down hardware costs), then with Llama (open-sourcing frontier AI models so no closed AI vendor could lock Meta's product roadmap or margins). The fixed cost of building and giving away the technology is treated as operating leverage that lowers future variable costs and preserves strategic freedom.
Origin
Facebook's infrastructure team noticed vendors were earning fat margins on data center hardware Facebook depended on; publishing their own designs as Open Compute Project drove those margins down industry-wide. Meta later applied the identical logic to AI by open-sourcing Llama rather than depending on closed-model vendors.
Core principles
- 01Complements are goods whose demand rises together, but that doesn't mean profits are shared evenly between them.
- 02Whoever controls the complement can tax or restrict the product built on top of it.
- 03Spending heavily to make a complement free or cheap for the whole market removes a competitor's ability to extract rent from you.
- 04This works best when you have the balance sheet to absorb the fixed cost of building what you're giving away.
- 05The payoff shows up as lower future variable costs and preserved control over the product roadmap, not direct revenue.
How to run it
- 1
Identify the chokepoint
Find the technology or input your product depends on that is controlled by a small number of vendors who could restrict or price-gouge you.
- 2
Assess your leverage
Confirm you have the scale and capital to build a competitive version of that input yourself — this only works if your version can be credible.
- 3
Build (don't just buy) the capability
Invest in R&D to produce a version of the input that is genuinely state-of-the-art, not a token effort.
Pro tip Half-hearted open source efforts don't commoditize anything — they need to be good enough that developers actually prefer them.
- 4
Give it away freely
Open source or otherwise remove the price tag on the technology, seeding it across the ecosystem so other vendors' pricing power erodes.
Watch out This can look irrational to investors in the short term since you're subsidizing the entire market.
- 5
Harvest the downstream benefit
Use the resulting lower costs and preserved control over your roadmap to protect the margins of your actual product.
In the wild
Facebook realized it was paying fat margins to data center hardware vendors, so it published its own server and networking designs for free, pulling the whole industry toward its specs.
→ Hardware costs across the industry, including Facebook's own, dropped as suppliers lost pricing power.
Rather than depend on closed AI vendors like OpenAI or Google for the models powering Meta's products, Meta poured billions into training Llama and gave it away.
→ Downward pricing pressure on closed AI providers and preserved control over Meta's future AI-driven product roadmap.
Common mistakes
Doing it without scale
A company without Meta's balance sheet can't absorb the fixed cost of building and giving away a frontier technology; the strategy requires real capital.
Treating it as charity instead of strategy
The goal isn't goodwill — it's removing a competitor's leverage over your margins and roadmap; losing sight of that can lead to under-investment.
Is it for you?
Best for
Large, capital-rich companies whose products depend on a technology controlled by a small number of external vendors.
Not ideal for
Startups or companies without the balance sheet to subsidize an entire market's access to a technology.
From the transcript
“We must ensure that we always have access to the best technology and that we're not locking into a competitor's closed ecosystem where they restrict…”
“Think about cars and gasoline. These are complements. When sales increase in one, sales increases in the other.”
From the episode
Meta