Base’s Growth Dilemma: Everything Done Right, Yet Users Still Leave

Original Compilation: Chopper, Foresight News

A few days ago, I came across a concept from Japanese philosophy: basho. A rough translation is “place,” but the philosopher Kitaro Nishida imbued it with a meaning far beyond a geographical location; it’s more like a situation: a field where all things can become themselves. In other words: people do not appear somewhere by accident but are shaped by where they are. Today, I will use this theory to interpret Base.

Last month, its number of active addresses fell to an 18-month low. Reflecting on this phenomenon, I realized: Base built only a location but never created the conditions for things to grow and take shape.

When Coinbase launched Base in 2023, the 暗号-native community developed a rare kind of faith. People believed it could finally solve Ethereum’s oldest problem: infrastructure everywhere, but no real users. Coinbase, with its 100 million users and unparalleled distribution capabilities, held a unique advantage. The door opened, and users were already waiting outside.

For a while, this confidence seemed validated. Base grew faster than any previous Layer 2. In October 2025, its Total Value Locked (TVL) reached $5.6 billion, and its fee revenue was unmatched across the entire L2 landscape. Thus, in September 2025, Base confirmed the issuance of a token, seemingly heralding an inevitably successful experiment. Yes, a place was becoming a basho.

Then, the users left.

The data tells a clearer story: Base’s active addresses have returned to levels seen in July 2024. The token issuance expectation perfectly satisfied the needs of airdrop hunters: get the final payout, then leave.

Base’s bet on the creator economy in 2025 also didn’t work. Its core was the Zora protocol, which defaults to tokenizing content. By the end of the year, 6.52 million creator and content tokens had been issued on Base via Zora. Of those, only 17,800 remained consistently active throughout the year, accounting for 0.3%. The remaining 99.7% have been abandoned.

Base’s daily active addresses peaked at 1.72 million in June 2025. By March 2026, it was down to 458,000, a staggering 73% drop from the peak. After Armstrong announced in September 2025 that Base was considering a token, active addresses fell by 54% in just six months, signaling a complete exodus of speculative capital.

Sociologist Ray Oldenburg studied what makes people return to a place repeatedly without compensation. He called them “third places,” like bars, barbershops, and city squares. They are not spaces for efficient production but offer a reason to return unrelated to incentives. The core idea is that the desire to return cannot be manufactured; it can only grow naturally from the possibilities a place provides over the long term. The crypto industry designs places to extract value from users and then wonders why no one stays.

This is a location without basho: people pass through, take what they need, and leave because leaving costs nothing. No identity is formed here, no capabilities that can’t be replicated elsewhere in three weeks, nothing that makes leaving a loss. Do unique relationships exist on this chain? We’ve never built things with that in mind, have we?

You cannot build a basho with financial incentives. Incentives can certainly pull people in, but they cannot make people want to stay. The desire to stay must come from the possibilities nurtured by the place over time. Kitaro Nishida called this the “logic of place,” referring to how a relational field shapes the things that emerge within it. The crypto industry designed fields for extraction and is now surprised that only extraction was born.

Brian Armstrong has publicly stated that the Base App is now focused on becoming a self-custody, trading version of Coinbase.

The original social and creator vision aimed at building social stickiness and allowing users to establish identities on-chain worth protecting has vanished. Judging by the data, this is a rational decision, but it also admits that such a vision never truly materialized. Base has a location; it now focuses solely on serving existing users because that’s all it can provide.

One Chain, One Narrative

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

Since June 2025, the usage of small and medium-sized L2s has collectively declined by 61%. Most chains outside the top three have become zombie chains: active enough not to shut down, yet desolate enough to be insignificant. The ratio of L2 to L1 daily active users has dropped from 15x in mid-2024 to 10–11x today. Most new L2s see their usage collapse directly after the incentive cycle ends. The entire L2 ecosystem is cooling, not just Base.

The rollup-centric roadmap was once a theory about user adoption: lower participation costs → users flood in → ecosystem forms → compound growth. The Ethereum Foundation released a 38-page vision document this year outlining Ethereum’s future direction. Meanwhile, the largest L2 by activity has bottomed out and left the OP Stack, and the second-largest L2 has stagnated.

Lowering entry costs does not equal creating the conditions for things to take shape. The industry solved the “entry” problem but took it for granted that a “sense of belonging” would follow. It doesn’t appear automatically because belonging is not a feature that can be shipped.

Farcaster is the product in crypto that comes closest to building a basho. Because a specific group of people has built a specific culture on it: developers sharing work, discussing Ethereum, forming opinions of each other over months. This takes time; competitors cannot copy it with higher rewards. Friend.tech tried to do the same with incentive mechanisms, topping the charts in a week and dying in a month. Same mechanisms, but no culture formed. The difference isn’t the product but whether people stay long enough for something to truly take shape.

What Makes People Stay?

The chains that retain users through the winter don’t rely on more generous incentives.

Arbitrum’s daily active addresses peaked at 740,000 in June 2024 and now stand at 157,000, also a 79% plunge. Both chains are declining, but the underlying logic is completely different.

Base’s users come online to trade, and they leave when trading volume drops. Arbitrum’s users, however, are unaffected by fee levels; the correlation between user count and fee revenue is almost zero. Base attracts tourists, while Arbitrum, somehow, retains users.

Hyperliquid stands firm because its trading experience is unique, and the community has formed an identity that doesn’t exist elsewhere. トークン incentives are almost irrelevant; being part of it has become part of their behavior and identity. Things shape users, and users, in turn, shape things.

The crypto industry is still optimizing “how to get people to come,” while the question of “how to create a situation” is only remembered after the data crashes, never considered at the chain’s design inception.

I believe Base, with the strongest distribution power in history, could have solved this problem better than any chain.

Now it’s a trading app. This is a reasonable product direction, but it’s also what over 40 products are already doing. Trading apps cannot generate a basho; they can only generate sessions: users come in when they have a trading need and leave when it’s done.

To truly become a successful application, a continuous connection needs to be established. Users need to build a relationship between visits, making the next visit feel like a return, not just an arrival.

Armstrong’s pivot is largely based on lessons Base learned from the data. The social layer, creator economy, on-chain identity—these were supposed to turn Base from “being used” to “being inhabited”—all require patience, and the system does not reward patience.

The Ethereum ecosystem needs Base to be more than just a trading venue. The entire L2 narrative is based on the idea that chains can become infrastructure around which people build their lives. If the L2 with the strongest distribution power in crypto history ultimately settles for being a faster Coinbase, then this narrative itself is untenable.

Kitaro Nishida believed that the deepest basho is where the boundary between the self and the place begins to dissolve. You cannot completely separate “who you are” from “where you are shaped.” This sounds abstract, but applied to a public chain, it means: a user cannot imagine their financial life after leaving a certain chain; a developer’s entire toolkit is based on a specific ecosystem; their identity barely exists elsewhere.

To my knowledge, such a thing has never been built on any L2. It might be impossible to build under incentive programs.

Even if you have 100 million potential users, if there’s nothing worth staying for, the place will eventually be empty. Base understands that now.

この記事はインターネットから得たものです。 Base’s Growth Dilemma: Everything Done Right, Yet Users Still Leave

Related: Arthur Hayes: HYPE’s Next Stop is $150

Original Author: Arthur Hayes, Co-founder of BitMEX Original Compilation: Saoirse, Foresight News We walk into the frozen forest and climb the steep volcano. It’s just another day of ski touring and meditation. Amidst the silent snowy woods, my thoughts roam freely. The creativity you can unleash when you’re fully focused, slowly ascending the volcano step by step, placing one ski in front of the other thousands of times, is astonishing. I cherish these three months of tranquility. Both body and mind crave rest and recovery. During the ski season, this means going to a real ski resort. On resort days, I remove my climbing skins, take the mechanical lift, and within minutes, I’m transported hundreds of meters into the air. Chairlifts and gondolas are great, but sometimes I have to…