In August 2023 we wrote a column entitled " Base & Liberty Reserve: More Similar Than You Think ." That argued Base was little more than a vehicle for Coinbase to offer centralized, custodial money transmission services without KYC/AML. And that vehicle was cynically disguised behind technobabble and vague aspirational claims. That framing has only become harder to dismiss following Jesse Pollak’s recent admission that the Base app has been handed back to the Coinbase mothership, with Jordan Fish, better known as Cobie, now leading it as a standalone product layer – even one that may expand beyond the Base ecosystem itself.
As Coinbase holds a wide range of licenses this was, and is, bad. Coinbase itself had admitted months earlier in February 2023 that Base was not decentralized at all at launch time. And Coinbase presented a roadmap for decentralization in 2023 . We will get into some details below but suffice it to say that roadmap was abandoned long before Coinbase ever made much progress on it. And L2Beat, a good general source for these things, has Base's current functionality at stage-0 (it will be downgraded from 1 to 0 in August 2026) with full training wheels. For those less familiar with industry jargon: a large and well respected monitor of systems like Base rates it as having no decentralization.
Right after our August 2023 column was published, Coinbase put out a commitment to decentralization with the Superchain that pretty much admitted Coinbase ran Base and promised to fix this by growing alongside Optimism and the larger Superchain effort. Again we will get to details below but that, too, was abandoned before much progress was made.
2024 had a public commitment from Vitalik to stop talking about stage-0 products alongside a heavy focus on pushing products to achieve stage-2 . Then in February of 2026 Vitalik led abandonment of the "layer-2 scaling as a solution for Ethereum scaling" strategy in the face of "progress to stage 2 [that] has been far slower and more difficult than originally expected."
We are just going to point out here that nobody had a fully fleshed out technical plan for achieving stage-2 in a secure and scaleable manner in 2023 or 2024. And the projects Vitalik is referring to there still do not have such plans. So it is not that people failed to execute on clear plans. It is that people's R&D efforts have yet to yield such plans. This is a bit like abandoning your Mars mission because you ran out of money before finishing detailed design of the rocket. The project was always a bet you could figure it out quickly and that bet did not pan out.
That is all a fairly dense summary of a few years of history. Something similar applies to pretty much every L2 out there. You can go check L2Beat to confirm the situation more broadly. Progress has been somewhere between slow-and-sporadic and nowhere.
Now we are going to look at what Base has done in a little bit more detail. And then we are going to face the big question nobody wants to talk about: at what point is running a stage-0 "full training wheels" L2 just running a custodial fund manager or money transmitter? And at what point should the lack of licenses and KYC/AML get prosecuted?
Base's History
The initial Base plan linked above had them implementing something known as permissionless fault proofs in 2023. If you do not know what that means no worries: what matters is this component was not implemented in 2023. As it was an essential step to removing some training wheels this was a problematic delay.
Then in April 2025 Coinbase finally announced success on this milestone . At that time Base's architecture was rated as a "stage-1" L2 indicating it functioned as mostly decentralized but still had backup protection mechanisms with full central control. There is of course some judgement required in evaluating exactly where these lines are. And what Base achieved by April 2025 would not count as stage-1 today; today it would be considered stage-0 with full training wheels and centralization. It is worth noting here that stage definitions have evolved over time as repeated incidents have shown that special mechanisms which are supposed to be "reserved for emergencies" show up more and more and turn out to convey more authority than the team originally let on.
This is a common theme in web3 that includes not just decentralization theater but also the occasional repurposing of a "safety" or "investor protection" mechanism to solve a different problem. Notably that problem is sometimes "how can I steal all this money?" That is what happens when a "decentralized" protocol is hacked via theft of admin keys nobody knew about. Yeah.
For years now our view has been that all these powers should be viewed in the worst possible light. Effectively we think staging evaluations should assume bad faith everywhere and take as conservative as possible an approach to the "could they steal this money?" question. Events have generally proven our view correct. And it is nice to see more people coming around to our view.
But back to the main story. There was progress on the roadmap even if it was weak and late. In February 2026 Coinbase announced it was abandoning the Optimism-centric decentralization plan for Base. Coinbase was instread moving to a unified, Base-operated stack . Prior to this change some of Base's admin controls were shared with Optimism. Following this "evolution" Coinbase placed control with a centralized security council multisig and a centralized coordinator multisig. Coinbase has a lot of money and influence and surely had some sway over Optimism. It for sure has an awful lot of influence over the individuals and small companies represented on those two multisigs.
So at this point Coinbase was years behind the 2023 plan and had straight-up abandoned it without providing any sort of real replacement plan or roadmap. And not only that, but Coinbase had reclaimed the sliver of control that briefly passed out of its hands to Optimism. Recall joint-venture partners can hold joint custody of funds. And the Coinbase-Optimism JV surely did just that. But at least it was some kind of movement in the direction provided back in the 2023 roadmap. When that roadmap got abandoned so did any legitimate claim of real progress. Noise is not progress; noise is noise.
Then in June 2026 Base experienced 2 outages in quick succession. Resolving these problems required 2 fixes which involved Coinbase modifying the code, rolling back the chain, and then instructing everyone on the network to restart with the fixes applied. So 3 and a half years in the system is stage-0 fully centralized and that central authority is exercised in practice to manage user assets. Right.
A Centralized Distributed Database
The resolution of those outages makes clear what Base is: nothing but a standard distributed database. You can buy an incredibly reliable version of this product from Oracle or SAP or IBM. You can run an incredibly reliable version for free using MySQL. You can get a cloud version of this from any cloud vendor. Distributed databases are a solved problem so long as your transaction volume sits within the capabilities of the technology you chose to deploy. They just work fine unless you are pushing super-super hard.
Does Base push distributed database technology hard? Base does a few hundred transactions per second. This is 1980s performance. Maybe early 1990s. If you read coverage from the internet bubble 20-25-30 years ago you can find large companies struggling to keep systems up at orders of magnitude more performance . That is not just before Bitcoin. It is before mobile phones were ubiquitous. Internet banking meant using a web browser on a computer and you actually had to confirm it was an https site before entering your password. Super old technology.
Go read the linked 2001 article about eBay's struggles in the late 1990s. We are not pretending these problems are trivial. We remember very well reading data from digital cameras in the mid 1990s over serial ports. We remember when, while doing medical imaging work, we could not load a full set of images – dozens of 16 megapixel black-and-white images – into memory at once on a computer that cost as much as a nice car. Now those images fit, comfortably and many times over, on the cheapest phone Samsung makes today.
That does not mean Base is trying to solve old problems. Achieving 1990s centralized server performance on a decentralized, permissionless and globally distributed network might be a genuinely difficult engineering challenge. Or not. But Base is not actually doing that. Base is not even trying. Base is now a slow, inefficient, unreliable and frankly terrible version of a database technology that was worked out decades ago. Coinbase has aspirations for Base to be much more. Back in 2023 we could have had a debate about how much leeway was warranted and which regulations should maybe be relaxed or ignored in the name of encouraging innovation. Maybe in 2024 too. By 2025 this was getting a bit tiring.
Base is now 3 and a half years old. It has a blockchain and that blockchain makes keeping the system up and running difficult. But the blockchain is not actually leveraged to build something different or new or novel. Coinbase still runs the entire thing centrally after all this time. The blockchain inside Base just forces engineers to face the "hard mode" version of a solved distributed database design problem. It is a self-imposed constraint and a self-imposed difficulty that has not, yet at least, bought decentralization. You do not deserve special treatment because you choose to build a standard product in a difficult way.
Why We Keep Saying Coinbase Not Base
We refer directly to Coinbase a lot there and attribute control to Coinbase rather than Base. We think this is justified. Go read, for example, this announcement blog post . In particular go down to the bottom where it says:
Building the next frontier of the internet is a collective effort. If you are passionate about scaling, securing, or decentralizing the future of Base, we want to hear from you. View our open roles and apply here .
If you click on the link those are Coinbase job listings. Literally. They are listings on a job posting platform called Greenhouse that clearly and explicitly state you would be working for Coinbase either as an employee or contractor. If you look in the comments on that post (and others) you can also find named Coinbase employees replying using "we" with respect to the Base team. The base.org website also has TOS and a Privacy Policy that reference Coinbase all over the place.
The logic is supposed to be that Coinbase "incubated" this thing and because it is decentralized Coinbase retains no liability. But we can see trivially that Coinbase's employees are running it. And if it is a stage-0 L2 it is not decentralized. So.....it is just Coinbase.
That is why we refer to Coinbase directly and often.
The Outlook & Legal Questions
Many people criticize many web3 projects by asking "why does this need a blockchain?" when plainly the project does not need one. For Base the original answer was "because blockchains are the only known technology that might enable solving this problem efficiently." And that may still be true. Again: or not. But after this much time, and without any visible forward progress, it is worth asking three difficult questions:
- Did Coinbase ever have a real plan?
- Given the lack of progress Base and others have made is this feasible near term?
- At what point should law enforcement stop providing room to innovate and instead start to investigate?
Coinbase had in 2023, and has today, a legal obligation not to run a KYC-free custodial money transmission platform. That stems both from Coinbase's licenses and the general principle that Coinbase needs to follow the law just like everyone else. Reasonable people can disagree as to whether a little bit of leeway is justified when a new system is launching. Engineers rightly worry about bugs and general user safety. Software is hard. But no reasonable person is going to argue that zero forward progress in 40 months means the leeway should continue indefinitely.
Coinbase also has an obligation to make truthful public statements. Especially with respect to Coinbase shareholders and business prospects and plans and all that kind of stuff. What if Coinbase never really had a plan? What if, to borrow a phrase, Coinbase only had "concepts of a plan"? There are real questions here about honesty and candor and whether the company misrepresented that it knew how to do this when it manifestly did not.
At some point these zero-progress development efforts look more like trolling than buidling. To go from stage-0 in February 2023 to stage-0 in August 2026 is literally zero progress. Work was done in the physics sense: code was written and deployed and lots of funds moved around. And we can be a bit more precise using more words from thermodynamics: we know work was done because hot air was expelled by the team. But that hot air did not accompany any progress. No mechanical work was done along the decentralization axis.
To the extent Base has more features or more complex mechanisms or more whatever else now: who cares? The point is not that the engineers building the system find the software interesting. Or even that they "worked" on things and expelled hot air. The point is supposed to be that the system does something useful and legal . It would certainly help if that something was related to Coinbase's publicly stated plans for the project.
As of now the entire exercise looks like an attempt to dress up "we did not, and do not, know how to do this correctly" as "we deserve an indefinite exemption from the law because what we want to figure out is so important." That started to smell bad years ago. And now it stinks.
Companies deserve some space to maneuver and we are not arguing every project undertaken by a public company must be perfectly planned out in advance and executed exactly on time to that exact plan. But several years of repeated failure, and a failure mode that is a listed company running an outlaw business, has to result in consequences. Otherwise why would anyone care about the rules?
