Today a16z published their 2022 State of Crypto Report. For such an intelligent and successful firm, they really seem to be struggling with the basic task of presenting logically-coherent evidence to support their claims.
In this post I’ll just focus on a couple slides, 39 and 40, because I think that’s enough to raise major questions about what a16z thinks they’re doing.
Here’s a16z’s accompanying text:
The take-rates of web2 giants are extortionate; web3 platforms offer fairer economic terms. (See slide 39 in the deck.) Compare Meta’s nearly 100% take-rates across Facebook and Instagram to NFT marketplace OpenSea’s 2.5%. As U.S. Congressman Ritchie Torres noted in a recent op-ed, “You know something is profoundly wrong with our economy when Big Tech has a higher take rate than the mafia.”
…In 2021, primary sales of Ethereum-based NFTs (ERC-721 and ERC-1155), plus the royalties paid to creators from secondary sales on OpenSea, yielded a total of $3.9 billion. That’s quadruple the $1 billion — less than 1% of revenues — that Meta has earmarked for creators through 2022.
I find it stunning to read this so-called “analysis” coming from one of the top VC firms in the world. The fact that Meta’s free ad-supported social network doesn’t pay content producers to use it means that Meta is “extortionate” and “100% commission”, and directly comparable to the 2.5% commission that OpenSea charges on NFT sales?
What are they expecting readers to conclude here? “Case closed, Web3 commissions are 40x lower than Web2 commissions. 2.5% vs 100%. There’s no hope for any social product that doesn’t use blockchain technology to close this 97.5% gap!” It’s nonsensical.
Have they tried thinking for five seconds about why the “take rates” are what they are? YouTube gets to charge 45% because they’ve built a platform that directs billions of users to watch videos they’re likely to be interested in. If you don’t need your video-hosting platform to help you acquire viewer traffic, then you don’t need to share as much of your revenue. E.g. Vimeo provides tools for monetizing hosted videos and their take rate is only 10% plus transaction fees. And of course, just making your own website with an HTML <video> tag and putting your own custom ads next to it enables you to pay 0-1% of your revenue toward your hosting platform.
So according to my own counter-analysis, Web2 has a lower take rate than Web3, because 0-1% is less than 2.5%.
Something is not right if a16z is publishing this as a serious attempt to explain their investments in Web3 to the public. They’re gaslighting us.
Slides 39 and 40 in a16z’s presentation wouldn’t be passable in a high school business class. They’re an insult to the reader’s intelligence. No one should be expected to nod along to these slides rather than spitting out their coffee and pausing to seriously question what’s going on.
I would have thought this was some kind of prank, a parody of that funny 2019 Softbank deck, if I hadn’t already seen a pattern of similar content coming from this firm.
Even if a16z’s expectations for the future of Web3 turn out to be correct, even if Web3 really is the future, it’s alarming that this is a16z’s publicly stated reasoning that’s supposedly worthy of our respect. Regardless if the conclusion is right or wrong, this analysis is logically flimsy.
For a more detailed unpacking of this “Web3 vs. Web2 take rate” nonsense, see my previous post, Chris Dixon’s Crypto Claims Are Logically Flimsy.
Or check out these other Web3 posts:
- Why Web3’s Shared Data Across Applications Doesn’t Matter
The ability for applications to read and write their data to the blockchain will be inconsequential
- Naval Ravikant and Chris Dixon Didn’t Explain Any Web3 Use Cases
Their web3-themed episode of Tim Ferriss’s Podcast didn’t provide the one thing I was hoping for
- Web3 is Vision-LARPing
Role-playing as a visionary when you don’t have a coherent vision
- Debating with Balaji Srinivasan about Blockchain Use Cases
Struggling to understand how Balaji thinks blockchain technology adds value to internet-enabled use cases
What are your thoughts? Comment here or tweet me @liron