To build on Ashwath’s excellent post around countercyclicality investing, where he details how investors have been throwing money at infra projects for the last two years, I wanted to explain why I think infra and L1s are over-emphasized as a play and how I think crypto as a whole needs to refocus. 

Ashwath is correct that most investors have recently poured money into infra projects – which makes sense. Ethereum, the most popular smart contract chain, has shown its inability to handle high transactional load multiple times in the last few years. NFT mints would spike transaction fees into the hundreds of dollars, causing the chain to have liveliness failures. As such, pouring money into infrastructure makes total sense. Investors are betting that the chain with the best infra will capture most of the market because it can handle all the users that show up. The last two years have given us a ton of infra improvement – DA layers, inscriptions, roll-ups, app chains, liquid staking, etc. Which is excellent; chains are probably more ready than ever to handle more users. 

However, beyond the fact that infra plays are over-saturated, there are also two more significant issues in my estimation. First, crypto has no ‘killer app.’ Secondly, I think building infra before use cases are clear is going about it backward.

First, crypto is building massive infrastructure for what amounts to a ghost town. There are barely any users in crypto at the moment. Some will say that Covid restrictions opening up, higher interest rates, lack of liquidity, or whatever else is causing the lack of users. But to me, the cause is much simpler. There is nothing to do. There is no killer app. When I wake up, I sign into the same collection of websites every day to check various things. Want to know what I don’t do? Sign into metamask or check a Dapp. Crypto needs apps that people check every day. 

Looking at Ethereum, we see a slow bleed of users in 2021. Ethereum users peaked in 2021 at almost 800K but have since bled to around 400K in 2023. Of course, the L2s have taken users from Ethereum. However, regarding L2s, we also see a decline in users. Arbitrum peaked at 1.37M users and has since declined to 682K. Optimism meanwhile peaked at 582K and has since fallen to 372K. Solana, meanwhile, peaked at 446K users and is now down 85K. The amount of active users on this small cross-section of chains has fallen between 37% – 81%. In the grand scheme of the internet, sites with fewer than a million users barely register. Neopets had 25M users at its peak and still has more daily users than Solana! 

Granted, some apps and chains have carved out a niche – like Uniswap, perp platforms, Friend Tech, etc. But people are not showing up en masse to use what is already here. Why do people think users will show up in the future for the same groups of apps? If our apps had such a big draw, we would see people showing up to them regardless of the circumstances. Instead, we see the opposite. Users will climb through broken glass if something cool is on the other side, but there is currently little reason to crawl.

The lack of killer apps also creates a fundamental disconnect with infra. People are building infra unquestioningly for the use cases they are predicting. The use cases people are forecasting are what has generally already come before – DeFi, P2E, DAOs, etc. But what if these use cases disappear? What if crypto is building infra for things that are flashes in the pan? No one has made a legit DAO in my mind, DeFi is proving to be unstable and potentially regulatorily unworkable, and no one has made a sustainable P2E game. Everything we take for granted in crypto could end tomorrow – leaving us with ghost chains and infra with nothing to do. Not only may the existing group of apps and use cases die off, but the future apps may be wildly different than what we have now, and the infrastructure teams are building may be unsuitable. With so much development in the AI sphere, it seems plausible that we need some specialized AI<>crypto infra structure – hell, maybe AI makes finance obsolete, and all the composable DeFi stuff around liquid staking is useless. The future is vast, stupid well of possibilities; who knows what it will bring? I think better the app first, then the infra to suit.

None of this is to disparage the work done or the teams who have made the infrastructure innovations. But, I want to highlight that we should be wary about constantly throwing more money at infra projects before seeing the apps. The app layer will be the most important for crypto – it’s where the users are. Without knowing what that app is or what users want to use crypto for, the infra is blind. 

As such, I am focusing much of my attention now on the app layer. I want to see new apps for the chains rather than more infra and yet another L1. I think the chain that ‘wins’ will be the one with the best apps, not the one with the best infra. Of course, developers should smooth the UX as much as possible, but without a compelling app layer and killer apps, there is little reason, in my mind, to get excited about specific chains or infra.

Keeping the importance of apps in mind and how much infra is out there, I think there are some things to watch for:

1. Some clever teams and VCs are probably quietly deploying capital into apps while people focus elsewhere – proving Ashwath’s countercyclicality thesis. 
2. Watch where these apps go. As I said, the chain that ‘wins’ will have the best apps. 
3. I expect ecosystem funds and grants to realign towards apps strongly. If you work on infra or with a chain, now may be the time to pivot to apps and think through some ecosystem plans. 
4. I will be most excited about the teams/chains focusing on creating compelling apps through hackathons, grant programs, and incentives. Focusing on apps shows an understanding of their importance and reflects well on their team. And after the apps arrive, they can build infra to suit their needs.

Of course, the app-first mindset will eventually become over-saturated, just like the infra and AI plays that Ashwath pointed out. But for now, this is where I think savvy people should be looking.