Coinbase CEO Brian Armstrong said on X on July 13 that Base’s content-coin strategy failed after running for more than a year. “It didn’t work, we pivoted earlier this year. We blew it, time to move on,” he wrote.
The comment is the clearest public admission so far that Base’s creator-token experiment did not deliver. ZORA, the token tied to the infrastructure behind that effort, has fallen about 95% from its all-time high in August last year, while its market capitalization dropped from roughly $550 million to about $30 million.
Base made content tokens a core product feature
Base began heavily promoting content coins through Zora in 2025 and built them into its own wallet product as a core feature. For a period, that helped push Base to the top among Layer 2 networks for new token issuance.
In July 2025, Coinbase renamed Coinbase Wallet to Base App and added a social feed, chat, payments, trading and app discovery. When the app expanded to more than 140 countries in December that year, Base was still describing it as a combined social, trading and payments product.
Zora supplied the tooling behind the creator economy model. Each content coin was tied to a specific post. When a user published an image, video or text post, the system would automatically create a freely tradable ERC-20 token. Each content coin had a fixed supply of 1 billion tokens, with creators receiving 1%, or 10 million tokens, at launch.
Those tokens started without a price. A market quote only emerged after other users began buying, and later trades determined market capitalization and holder gains or losses. Buyers received a token linked to a post and could sell it at any time, but the token did not grant copyright in the post, equity in the creator, future income rights or profit sharing.
Zora’s terms of service limited the purpose of the tokens to entertainment, use and consumption, and required users to confirm that the purchase was not for equity or profit-sharing purposes. In practice, buyer returns mainly depended on whether someone else would later pay more.
Creator coins and content coins were designed to feed each other
Zora also offered creator coins tied to an entire account. Each Zora account could have only one creator coin, also with a supply of 1 billion tokens. Half went to the public market, and the other half vested linearly to the creator over five years.
Content coins published by that account were linked back to the creator coin. Zora’s idea was that popular content would raise demand for the creator coin. Creators could sell their allocated tokens and also receive a share of trading fees.
Base had argued at the time that the structure could bypass advertising, brand deals and follower thresholds, turning attention directly into trading income.
Token issuance drove activity, but not durable users
Cheap token creation quickly lifted Base’s headline activity metrics. In August 2025, after the relaunch of Base App, Zora hit record activity: more than 1.6 million creator coins were minted, nearly 3 million independent traders took part, and total trading volume exceeded $470 million. Zora’s token price also rose nearly fivefold in one month.
That surge did not hold.
In April 2025, Base’s official account posted “Base is for everyone” through Zora, which automatically generated a token with the same name. The token jumped shortly after launch, then fell about 95% within hours. Base said it had not sold the token and had not issued it as an official project, but the line between a post, a token launch and official endorsement was hard for ordinary users to distinguish.
A later example involving creator Nick Shirley made the problem more visible. Shirley’s investigative videos drew more than 100 million views on social media, and Armstrong had publicly promoted him. Shirley’s creator coin at one point reached a market capitalization of $15 million before falling back quickly. Viral attention brought short-term buying, but it did not create lasting demand.
The partnership between Base and Zora also drew complaints from developers in the ecosystem. Some argued that Base devoted too much visibility and too many resources to Zora and creator coins, without building a durable user moat, while reducing exposure for other Base projects. A community member who questioned Armstrong on July 13 also said many participants suffered losses as token prices fell.
Armstrong defended the model in January, then Base pulled back in February
Armstrong was still defending content coins in January. Responding to a former Coinbase engineer who criticized the model as zero-sum, he said purchases of content coins would create economic value and demand for creator coins.
Roughly a month later, Base App said it would end Creator Rewards and remove its Farcaster-powered social feed, while shifting the product focus toward tradable assets.
In March, Armstrong said on a podcast for the first time that Base App’s SocialFi features had not worked especially well. Base’s 2026 strategy later put trading and stablecoin payments at the center.
The company said Base processed more than $17 trillion in stablecoin volume in 2025, covering 26 local currencies and 17 countries. Those figures gave a clearer commercial basis for its shift back toward financial infrastructure.
Base says trading still gets most of the resources
In the same X post, Armstrong also pushed back on another criticism. @smileyXBT argued that Base’s current focus on AI agents amounted to chasing another hype cycle. Armstrong replied that Base’s roadmap has consistently centered on three priorities: trading, payments and AI agents. He added that most resources are currently going to trading.
From Base’s own tokenized post in April 2025 to Armstrong’s “we blew it” comment in July 2026, the cycle ran for 15 months. Over that stretch, ZORA lost nearly $500 million in market value.
According to the article, Coinbase kept leaning into the strategy throughout that period, integrating Zora into its wallet product, encouraging funds to build creator-token indexes, and giving platform-level exposure to insider-linked tokens. The company may now classify the period as a closed product experiment. Losses in user accounts, however, did not disappear with the strategy shift.