Base's growth dilemma: Doing everything right, but users still leaving

Original author: Thejaswini M A

Original compilation: Chopper, Foresight News

A few days ago, I came across a concept in Japanese philosophy: basho (場). A rough translation would be “place,” but the meaning Nishida Kitaro gave it goes far beyond a mere geographic location—it’s more like a situation: a field in which everything can become itself. In other words, you don’t show up somewhere by accident; the place you’re in shapes you. Today, I’m going to use this framework to interpret Base.

Last month, its number of active addresses fell to a new 18-month low. Reflecting on this, I realized: Base has only built a location—it has never created the conditions that let things grow and take shape.

When Coinbase launched Base in 2023, the crypto-native community rarely produced a kind of faith. People believed it could finally solve Ethereum’s oldest problem: infrastructure everywhere, yet no real users. With Coinbase’s 100 million users and unrivaled distribution power, it had a unique advantage. Once the gate opened, users were already waiting outside.

For a while, this confidence seemed to be validated. Base’s growth rate outpaced all prior Layer2s. In October 2025, its total value locked (TVL) reached $5.6 billion, and its fee revenue was unmatched across the entire L2 space. Then in September 2025, Base confirmed the issuance of a token, as if predicting a destined-to-succeed experiment. Yes—somewhere was turning into a basho.

Then the users left.

Looking at the data is even more direct: Base’s active addresses returned to the level of July 2024. The token issuance expectation just happened to meet the needs of airdrop hunters—get the last payout, then go.

Base also bet on the creator economy in 2025, but it didn’t work. At its core is the Zora protocol, which defaults to tokenizing content. By the end of the year, Base had issued 6.52 million creator and content tokens through Zora, but throughout the year only 17,800 remained continuously active—just 0.3%. The remaining 99.7% has gone untouched.

Base’s daily active addresses peaked in June 2025 at 1.72 million. By March 2026, it was down to 458,000—a 73% plunge from the high. After Armstrong announced in September 2025 that Base was considering issuing a token, within just six months active addresses dropped by 54%, meaning speculative capital completely exited.

Sociologist Ray Oldenburg studied what makes people repeatedly return to a place without regard to compensation. He called it the third place—such as bars, barbershops, and city squares. They aren’t efficient production spaces, yet they give people a reason to return that has nothing to do with incentives. The key is this: the desire to come back can’t be manufactured by humans; it grows naturally out of the long-term possibilities that the place provides. The crypto industry designs places in order to extract users—then wonders why nobody stays.

This is what it looks like to have no basho: people pass through, take what they need, and leave, because leaving costs nothing. There’s no identity formed, no ability that can’t be replicated elsewhere within three weeks, and nothing that makes leaving a loss. Is there any unique relationship built on this chain? We never build things with that kind of thinking, do we?

You can’t build a basho with financial incentives. Incentives can certainly lure people through the door, but they can’t make people want to stay. The desire to stay must come from the possibilities that the place nurtures over time. Nishida Kitaro called it “the logic of place” (basho logic), referring to how a field of relationships shapes the things that emerge within it. The crypto industry designed fields for extraction—and then is shocked to find that what emerges is only extraction.

Brian Armstrong has publicly stated that the Base App is now focused on becoming Coinbase’s self-custody and trading version.

The social and creator-economy vision that once aimed to build social stickiness—so users could establish identities on-chain worth protecting—has disappeared. According to the data, this was a rational decision, but it also admits: that vision was never truly formed. Base has a location; it now focuses only on serving past users, because that’s what it can provide.

One chain, one track

Base is the most striking microcosm of the entire L2 model.

Since June 2025, the usage rates of small and mid-sized L2s overall have declined by 61%. Most chains beyond the top three have turned into zombie chains: active enough not to be shut down, yet so quiet they’re essentially irrelevant. The ratio of daily active usage relative to L1 dropped from 15x in mid-2024 to today’s 10–11x. Most new L2s see usage collapse outright once the incentive period ends. The entire L2 ecosystem is cooling down—not just Base.

The rollup-centered roadmap used to be a theory about user adoption: lower participation costs → users flood in → the ecosystem forms → compounding growth. This year, the Ethereum Foundation released a 38-page vision document outlining the future direction of Ethereum. Meanwhile, the largest L2 by active usage bottomed out and left the OP Stack, and the second-largest L2’s growth stalled.

Lowering the cost of entry doesn’t mean you’ve created the conditions for things to take shape. The industry solved the “getting in” problem, then assumed “belonging” would follow automatically. It won’t appear on its own, because belonging isn’t a feature you can ship.

Farcaster is the product closest to building a basho in the crypto world. Because a group of specific people built a specific culture on it: developers share their work, discuss Ethereum, and over the course of months develop their opinions of one another. This takes time, and competitors can’t replicate it with higher rewards. Friend.tech tried to do the same using incentive mechanisms: it hit the top in a week and vanished within a month. Same mechanism, but no culture formed. The difference isn’t the product—it’s whether someone stays long enough for something to truly take shape.

What can keep people?

A chain that retains users through the winter doesn’t rely on richer incentives.

Arbitrum’s daily active addresses reached a peak of 740,000 in June 2024; now it’s 157,000—also a 79% crash. Both chains are declining, but the underlying logic is completely different.

Base’s users came online to trade; when trading volume falls, they leave. Arbitrum’s users, by contrast, are largely unaffected by fee levels—there is almost no correlation between user numbers and fee revenue. Base attracts tourists, while Arbitrum, somehow, manages to keep users.

Hyperliquid holds steady because its trading experience is uniquely good, and a community forms a sense of identity no other place has. Token incentives matter almost not at all; being inside it becomes part of how they behave and who they are. Things shape users, and users in turn shape things.

The crypto industry is still optimizing “how to get people to come,” while “how to create a situation” is only remembered after the data collapses—never considered at the start of chain design.

I believe Base had the strongest distribution power ever, and could have solved this problem better than any chain.

Now it’s a trading app. That’s a reasonable product direction, but it’s also something already done by more than 40 similar products. A trading app can’t produce a basho—it can only produce sessions: users come in when they have trading needs, complete the transaction, and then leave.

To truly become a successful app, you need to build an ongoing connection. You need users to form a relationship between each visit, so the next visit feels like a return—not just an arrival.

Armstrong’s pivot is, to a large extent, based on lessons Base learned from its data. The social layer, the creator economy, on-chain identity—these should have turned Base from “being used” into “being inhabited,” but they all require patience, and the system doesn’t reward patience.

The Ethereum ecosystem needs Base to be more than just a trading venue. The foundation of the entire L2 narrative is that a chain can become infrastructure around which people build their lives. If the L2 with the strongest distribution in crypto history ultimately settles for being a faster Coinbase, then this narrative itself can’t hold up.

Nishida Kitaro believes the deepest basho is where the boundary between self and place begins to dissolve. You can’t completely separate “who you are” from “how a place shapes you.” That sounds abstract, but on a public chain it means: a user can’t imagine a financial life after leaving that chain; all of a developer’s toolkits are built on a particular ecosystem; and their identity can scarcely exist anywhere else.

As far as I know, nothing like this has ever been built on any L2. It may be impossible to build it under incentive programs in the first place.

Even if you sit on 100 million potential users, as long as there’s nothing worth staying for, the end result will still be an empty crowd. Base understands that now.