Written at the end of 2025: Code, Power, and Stablecoins

Author: Stepan | squads.xyz

Original Title: Random end of year shower musings on the state of the stablecoin economy and its participants

Translation and Editing: BitpushNews

The year 2025 has clearly demonstrated one thing: stablecoins have established a foothold, and their underlying infrastructure will become the cornerstone of financial services construction in the next decade.

As this year draws to a close, I have been reflecting on the stage we are at, the insights gained in 2025, and the future trajectory. Below are my observations on the stablecoin economy as we step into 2026.

A few preliminary notes:

• Claude and Deni also contributed to this article.
• Squads is a fintech company, not a bank or digital asset custodian.
• The content of this article does not constitute financial advice.
• Charts and images in this article are generated by Nano Banana, inspired by the aesthetic of Tom Sachs, which I greatly admire.

Data Overview

In 2025, the stablecoin market size surpassed $300 billion, up from just $205 billion at the beginning of the year. In less than twelve months, new supply approached $100 billion.

By comparison: the total supply growth in 2024 was $70 billion, and in 2023 there was actually a decline.

These forecast figures reflect strong institutional confidence. JPMorgan projects stablecoin market cap will reach $500-750 billion in the coming years. Citibank’s baseline forecast is $1.9 trillion by 2030. Standard Chartered predicts it will reach $2 trillion by 2028. Today, stablecoin issuers are among the top ten holders of US Treasuries globally.

This is no longer mainly a cryptocurrency story. It is a story about money. And the infrastructure, services, and products capturing this growth will become some of the most valuable assets built over the next ten years.

What We Learned from the Synapse Incident

Part of what is driving this shift is that more and more people recognize that stablecoin infrastructure provides fundamentally different trust assumptions. This is not only because building on stablecoins is cheaper and faster (which is true), but more importantly, because you trust mathematics and code rather than centralized entities’ “trust me, your money is safe” promises.

To understand why this matters, look at what happened with Synapse.

Synapse Financial Technologies was once a model BaaS (Banking-as-a-Service) company. It received support from top investors, connecting over 100 fintech partners with FDIC-insured banks, serving about 10 million end users. Its slogan was very clever: fintech companies could access banking services without becoming banks; banks could gain distribution channels without developing apps; consumers could enjoy modern experiences with traditional protections.

In April 2024, Synapse filed for Chapter 11 bankruptcy. Over 100,000 people lost access to their funds. The court-appointed trustee discovered a shortfall of $65-96 million between what customers were owed and what the bank actually held. At a December 2024 hearing, the trustee (former FDIC chair) compared this situation to her father’s deposits going to zero during the disintegration of Yugoslavia.

The root cause was a failure in the middleware layer’s accounting and reconciliation. Synapse was responsible for recording asset ownership between fintech firms and banks. When this system failed, there was no traceable “truth.” Banks blamed each other. Fintechs had no direct relationship with customer funds. Ordinary people watched helplessly as their savings vanished into bureaucratic uncertainty.

The crypto space has also experienced catastrophic failures: FTX, Celsius, Terra/Luna. But these failures stemmed from centralized custodians making high-risk bets with deposit assets. Their failures are similar to Synapse’s: opaque systems that only reveal the truth when it’s too late.

The lesson from traditional fintech failures and crypto failures is the same: when you cannot see where the money is, you cannot know if it is safe.

Self-Custody and Insurance Issues

Self-custody stablecoin accounts have, in some ways, changed the risk model, making FDIC insurance less necessary in many scenarios.

Traditional banking operates on a fractional reserve system. When you deposit funds, banks lend out most of it, keeping only a small reserve. Your “balance” is just an IOU. If enough people demand withdrawals simultaneously, or if loans turn bad, the money may not be there. FDIC insurance is designed to prevent this failure mode. It insures against mismanagement of your funds by the bank.

Self-custody stablecoin accounts work differently. Assets are stored in smart contracts. At any moment, anyone can verify whether the funds are there. Not as IOUs, not as claims on reserves, but as actual assets under user control. There is no counterparty risk from bank lending decisions.

But this argument often overlooks one point: stablecoins themselves carry issuer risk. A smart contract filled with USDC is useless if its issuer, Circle, faces regulatory crises or reserve runs. Holding USDT is essentially a bet on Tether’s reserve management. Self-custody eliminates intermediary risk but does not eliminate issuer risk.

The difference is that issuer risk is monitorable. You can check reserve attestations. You can observe on-chain fund flows. You can diversify across different issuers. Traditional bank risk, however, is hidden in a black box until a disaster strikes.

This does not mean self-custody is suitable for everyone. Large institutions may still need regulatory frameworks and insurance products. But for many use cases, a self-custody model with monitorable issuer risk is preferable to opaque institutional trust models that require insurance as a safety net.

Global Reach and the Last Mile Challenge

Stablecoins offer something that traditional fintech cannot: true global reach from day one.

A wallet can be used anywhere. Smart contracts do not care about jurisdiction. Transactions between stablecoins are inherently borderless. For businesses paying remote contractors, managing cross-entity funds, or settling with suppliers accepting stablecoins, this infrastructure can operate instantly and globally.

Compare this to traditional international expansion: you need local banking partners, local licenses (which often require different permits for different activities), local compliance teams, local legal entities. Each country is essentially a new startup. That’s why most digital banks operate only domestically or take years to expand to a few markets.

Revolut has been working for nearly a decade and still isn’t fully global.

The bottleneck in stablecoin infrastructure is the “last mile”: connecting to fiat currency. Fiat on/off ramps still require local licenses and local partners. You cannot fully escape this.

But there is a huge difference between “we need to solve fiat connectivity in this market” and “we need to rebuild the entire banking tech stack in this market.” The “last mile” is modular. You can partner with local fiat on/off ramp providers for currency exchange without rebuilding core infrastructure from scratch. You can reach most of the world via stablecoin channels and then gradually onboard fiat partners where needed.

Traditional fintech cannot deploy services without building a complete tech stack in each market. Native stablecoin companies are inherently global from inception and then address the last mile gradually. This is a fundamentally different expansion approach.

Blockchain Wars for Purpose-Built Chains

Several well-funded teams are building new blockchains specifically for stablecoin payments. The core idea: existing blockchains are optimized for trading, not payments, and purpose-built infrastructure can offer higher throughput, lower latency, and compliance tools tailored for payments.

It’s a reasonable idea, proposed by smart people. Stripe and Paradigm are building Tempo, Circle is building Arc.

But there is a counterpoint worth considering.

Building a new Layer 1 from scratch means trust must be rebuilt from zero. Blockchains are trust machines, and trust is accumulated through operation. It comes from years of failure-free operation, safeguarding billions in funds without vulnerabilities, a developer ecosystem that deeply understands edge cases, and code that has withstood attacks. This is the Lindy effect applied to infrastructure.

Established chains have this accumulated trust. Solana has processed trillions of dollars in transactions, with mature tools, wallets, bridges, and integrations. Ethereum’s operational history is even longer. The question is whether the gap between what these chains currently offer for payments and the trust gap that new chains must fill is greater than the trust they have already accumulated.

There is also the consideration of neutrality. Chains controlled by large payment companies, regardless of how “neutral” they claim to be, embed the interests of those companies into their architecture. Building on truly neutral public infrastructure can provide different guarantees.

Agentic Finance

Today, when people talk about Agentic Finance, they often imagine intelligent agents managing your financial life: making investment decisions, managing your portfolio, optimizing your entire financial existence.

But that’s not the real opportunity, at least not yet.

The real opportunity lies in the boring and mundane parts. Automating routine financial processes that currently require manual intervention: monitoring invoices, matching them with purchase orders, initiating payments, handling reimbursements, executing periodic trades. Not replacing human judgment on key decisions, but automating tedious, operationally resistant tasks.

The question is: how do agents actually move funds?

Traditional payment channels are designed for humans. They assume the initiator is a person with credentials. Giving an agent bank login credentials is a security nightmare and a compliance violation. Agents might hallucinate, be manipulated, or make errors at machine speed.

This is where stablecoin channels and smart contracts become truly important. Agents do not get credentials; they get a set of restricted permissions encoded in smart contracts: each transaction can move up to X dollars, only to pre-approved addresses, only at certain times or for specific purposes. These constraints are enforced by code. Agents cannot overreach because permissions are baked into the architecture.

The trust assumptions provided by blockchain—verifiable, bounded, transparent—are exactly what is needed when software autonomously moves funds. Traditional systems require you to trust that agents will not misbehave. Smart contract systems, by design, make misbehavior within defined constraints impossible.

This does not eliminate all issues. What happens if an agent makes a mistake within its permissions? Who is responsible if an agent approves an invoice that is technically compliant but fraudulent? These questions need answers.

But the starting point—permissions enforced by architecture—is something native to blockchain systems, and very hard to retrofit onto traditional channels. Autonomous finance will arrive, and its secure infrastructure will inevitably be native stablecoin.

Reflection on Security

The gold rush in stablecoins is attracting teams with very different security philosophies. For some of these teams (unfortunately including their clients), the outcome will not be good.

A pattern is emerging: rapid deployment, gaining users, then solving security issues later. These teams use vague “self-custody” definitions to hide their actual trust models. They rush to integrate without proper security and vendor vetting. They take shortcuts in key management. They treat operational security as a cost center.

Part of this is understandable. The market is rapidly evolving. Competition is fierce. Spending a few extra months on security might mean losing market share to competitors.

This trade-off makes sense in most industries. But not in financial infrastructure.

Building a bank or similar institution means establishing trust over decades, not quarters. It means even aggressive growth strategies must manage risks conservatively. It means creating systems that can handle unforeseen edge cases.

Teams that succeed in 2026 and beyond will be those with genuine expertise and a security-first mindset.

Privacy Challenges

My non-mainstream view is that, so far, privacy in crypto has largely been a checkbox concern. For trading, DeFi, and speculation, lack of substantial privacy has not been a barrier. The ecosystem has mostly functioned well with pseudonymous addresses and transparent transaction histories.

But as stablecoin infrastructure brings real business activity and productive economic activity on-chain, this will change.

When real companies use stablecoin channels for fund operations, privacy becomes critical. Competitive intelligence leaks are a real concern: your suppliers, customers, cash flows are all visible to anyone willing to look. No serious company wants its financial operations exposed to competitors, and no CFO would want to move significant funds through channels where every transaction can be publicly analyzed.

This is a problem we need to address today to prevent it becoming a bottleneck for future adoption.

The good news is that stablecoin privacy models do not require the full realization of a cryptopunk utopia. We do not need complete anonymity. What we need is selective disclosure—a fundamentally different goal.

Selective disclosure means: proving what needs to be proven without revealing everything else. Proving you have sufficient funds without showing your balance; proving a transaction is compliant without exposing counterparty details; proving your identity meets requirements without submitting documents. The owner can see everything, the system can verify compliance, and others only see what is deliberately disclosed.

We have the technology to solve this. I have spoken with many outstanding teams building privacy infrastructure.

The problem is that this technology is still in early stages. These codebases are large, hard to audit, difficult to formally verify, and untested in real-world scenarios. They require trust and security assumptions very different from what we have built so far. The crypto ecosystem has spent years strengthening core protocols, accumulating operational trust only after surviving attacks and edge cases. Adding new, unproven privacy layers risks undermining this foundation.

The real challenge is how to add privacy features without compromising security. This might mean embedding privacy deeper into the first-layer protocol or finding ways to avoid relying on large-scale trust in new cryptographic systems.

Looking Ahead

The growth story of stablecoins in 2025 mainly revolves around migrating existing fintech functions onto better infrastructure: payments, yields, consumption, card services. Like a globalized Mercury or on-chain Revolut. That’s good. It’s faster, cheaper, and can reach markets that traditional fintech takes years to penetrate.

But what stablecoin channels unlock is much bigger than just doing the same things more efficiently. You gain programmable money. You connect to internet capital markets where new financial primitives are built daily. You enable agents to manage funds under truly secure conditions, not just trusting they won’t do harm.

This is our opportunity to rethink what financial services should truly be.

I have yet to see enough teams pursue this vision. The opportunity is right in front of us, yet most industry participants are still running the same 2015 fintech playbook on a new track. I hope to see this change by 2026.