Logo
Logo

Agentic Economy #11: Rewriting the Money Grid: Cloudflare's AI Checkout, the UK's Tokenization Push, and Sony's Self-Sufficient IP Empire

July 15, 2026

Cobo Agentic Economy

Welcome to issue 11 of Agentic Economy.

Stablecoins are turning into the base currency for direct settlement, whether that’s platform to platform or, soon enough, machine to machine. Banks, tech platforms, and crypto native firms are all racing to control the settlement rails underneath it all, because whoever owns the rails gets to set the fees and write the rules. This issue tracks that fight for control, from bank settlement all the way out to consumer platforms and direct machine payments.

Stablecoins proved something the old system swore couldn’t be done cheaply, money moving instantly around the world without a bank sitting in the middle of every step. That’s sparked a scramble, and it’s splitting into two very different strategies.

Banks and governments are racing to build that same capability inside systems they already control. UK banks are buying back a majority stake in Vocalink from Mastercard, pushed hard by the UK government and the Bank of England, part of a shift toward tokenized deposits meant to cut Visa and Mastercard out of everyday domestic settlement. SWIFT shows just how seriously incumbents are taking the threat. An organization famous for taking years to reach consensus shipped a shared ledger in nine months, handing banks back the round the clock instant settlement that used to be a stablecoin exclusive. Banks are racing the clock to keep instant global settlement as a card they hold themselves.

Crypto native firms and new entrants are pushing straight into the banks’ own territory, using efficiency as the wedge. OUSD’s zero fee settlement rail goes head to head with the Fed’s Fedwire system and SWIFT’s multi trillion dollar correspondent banking network. Circle already holds a federal trust charter from the OCC, and now it wants a Fed master account too, direct access to the central bank that used to be reserved for licensed banks alone.

Consumer platforms are playing the same game with a different face. Stablecoins let a platform become the bank inside its own economy, pulling back fee revenue and control that used to flow to Visa, Mastercard, and outside banks. Sony just got approved to set up a US trust bank. Pair that charter with its own stablecoin and the Soneium blockchain, and Sony is building a closed loop for PlayStation and Crunchyroll where no outside institution takes a cut and no third party has to sign off on anything.

Robinhood put tokenized stocks and perpetual futures right on the home screen for nearly 28 million retail users, stripping out the middlemen that used to sit between a trader and a trade. Kraken bought Reap to get into cross border B2B trade, using stablecoins to fix a weekend settlement delay that’s been a headache for global supply chains for years.

Every scenario above still needs a human somewhere in the loop. Agentic commerce doesn’t, and that changes what a payment system actually has to do.

Machines don’t watch ads. They’re not going to register an account just to call an API once. What fits them is paying per request, per data point, or per result. HTTP 402 sat dormant in the protocol for thirty years because charging a fraction of a cent used to get swallowed whole by card network fees. Stablecoins and near zero cost networks like Base solved that math problem, which is exactly why Cloudflare’s Monetization Gateway brought status code 402 back to life. AI agents can now pay in seconds, no account, no identity check, and the whole transfer costs less than a penny. The internet used to make money off human attention. Now it’s starting to make money off machine intelligence, and most of these payments never need a single human click.

Issue 10 of Agentic Economy covered OUSD in depth, a stablecoin backed by a coalition of more than 140 financial and tech giants. The whole design rests on redistributing where the money goes. Zero minting and redemption fees kill the friction, interest on reserves flows back to distribution partners, and custody sits across multiple institutions instead of one. That shakeup hit Circle’s stock hard on launch day, and the press wasted no time calling OUSD a USDC killer.

Treating OUSD as a direct rival to USDC might be the single biggest misread of the whole launch.

Misread One: Judging OUSD by Crypto Industry Logic

The sharpest criticism comes from inside crypto itself. Critics argue OUSD is stuck with a liquidity lock problem, since USDC and USDT already dominate pricing, DeFi collateral, and trading habits, and shifting that kind of market share to OUSD looks close to impossible.

Fair point, but it only holds inside the existing crypto market, where stablecoins fight each other across exchanges, perp order books, and DeFi lending pools. OUSD was built for a completely different arena.

The Fed’s Fedwire, SWIFT’s decades old bank messaging network, and the correspondent banking system built around both of them, that’s the actual battlefield. Visa and Stripe don’t lose sleep over liquidity depth on Binance. They care about settling merchant flows and cross border payments with near zero friction in seconds, and keeping that running through the weekend when banks are closed. OUSD is going after trillions in traditional capital flow that’s never touched a blockchain. The USDC killer label put it in the wrong fight from day one.

Misread Two: Zero Fees Will Invite Spam and Congestion

Here’s the other worry. Drop minting and redemption fees to zero on a public chain, and won’t spam and arbitrage bots flood the system, the classic tragedy of the commons? That risk is real in decentralized DeFi. Institutional capital markets don’t work the same way.

DeFi leans on steep gas fees as its anti spam defense against anonymous bots. OUSD’s participants aren’t anonymous bots. Every one of them is a KYC and KYB verified financial institution, names like BlackRock, Citi, and Visa, and their money moves in ways that are real and predictable.

The compliance process itself is the filter. Only so many institutions can clear KYC and KYB, so there’s no need for fees to screen counterparties on top of that. Zero fees were never a security risk for OUSD. Traditional finance has wanted exactly this for years, lower costs alongside higher efficiency.

None of that means smooth sailing ahead. A clearing channel built jointly by Visa, Stripe, and BlackRock looks like an antitrust minefield waiting to go off. Nobody’s tested whether governance across that many parties can move fast in an actual crisis. And 140 signed partners doesn’t mean 140 active, high frequency users. Getting institutions to abandon clearing rails they’ve run for decades, and migrate this deep, still takes years of system integration and rebuilding trust from scratch.

Once a stablecoin gets cleared to work as a deep settlement asset, it stops being just a payment rail and turns into a bank grade tool. Push settlement speed to its limit, and the edge comes down to liquidity management and capital efficiency. Settlement efficiency itself is becoming the thing everyone’s actually competing over.

Cloudflare just launched Monetization Gateway, built around a status code that’s been sitting dormant since the 90s, HTTP 402, Payment Required. Cloudflare has switched it back on, and now any web page, API, or data endpoint running on its edge network can charge in real time for every single request.

When an AI agent requests a protected resource, it gets back a 402 response with a price and a wallet address. The agent pays in an onchain stablecoin, attaches proof of payment, and gets the resource. No account, no API key, no credit card on file anywhere. The wallet address works as identity, and the payment itself works as proof. Developers just set one routing rule in the dashboard, the edge network handles everything else, and the origin server doesn’t even know a payment happened.

That frictionless setup solves a problem that’s been around for a while. Content creators and API developers facing AI crawlers used to have exactly two options, let them scrape for free or block them outright. Now that AI agents are fast becoming the internet’s main visitors, that all or nothing choice doesn’t work anymore.

AI agents don’t browse ads, and they’re not going to register an account, hand over a card, and pay 15 bucks a month just to check the weather API once. For non human traffic, charging per request obviously makes more sense than charging per seat or per month.

None of this works unless transaction costs stay tiny, and that’s exactly where things are right now. Stablecoins running on cheap networks like Base have pushed per transaction costs below a penny. Compare that to an API call worth half a cent, where credit card processing fees used to eat the entire margin. Crypto gets dismissed as pure speculation a lot, but here’s an actual working micropayment use case. Cloudflare’s Pay Per Crawl last year only targeted AI crawlers. Monetization Gateway stretches that same logic across the whole web.

Will x402 become the default protocol for machine payments? Cloudflare handles roughly a fifth of the world’s internet traffic, which makes it hard to route around. Whether the protocol actually takes hold depends on whether system level AI agents like OpenAI, Anthropic, and Apple decide to support it. Skip that, and this stays a Cloudflare only feature. Land it with the major agents, and it could become the default payment layer for how machines access the internet.

If this logic catches on, the internet reshapes itself. Sellers become APIs that respond to requests automatically, buyers become AI agents carrying digital wallets on behalf of users, and there’s no platform cut, no ad auctions, just clearly priced content and data. Ads and subscriptions ran the internet for thirty years because that was the only option the technology allowed. This time, there’s another one on the table.

Circle just landed a national trust charter from the OCC, setting up Circle National Trust and putting USDC’s underlying infrastructure under direct federal oversight for the first time. It’s a narrower charter than a full banking license though, no deposits from the public, no lending, no FDIC coverage, limited to digital asset custody for Circle and its affiliates.

Here’s the catch. The charter leaves out the one thing that actually makes money, reserve management. That’s filed under future capability, meaning Circle has to apply for it separately down the line. So what Circle actually built is a two tier structure. USDC issuance stays with the NYDFS regulated trust company, while the new OCC federal trust only handles custody. The most profitable part, earning interest on reserves, gets pushed to phase two.

The OCC has handed out a batch of crypto trust charters since late last year, Ripple, Paxos, BitGo, Fidelity, Bridge, and Sony all picked one up, mostly to bring these firms under the GENIUS Act’s regulatory umbrella. Getting the charter is just step one though. Staying compliant long term is the harder part. Anchorage got its charter in 2021 and landed a consent order over anti money laundering violations by 2022. Circle didn’t dodge this either. On the same day it got its charter, Wisconsin prosecutors charged the company with obstructing a search warrant, hardly the headline Circle wanted, and a sign of how rough regulatory scrutiny can get once it starts.

Two things to watch from here. First, Circle pushing the OCC to unlock reserve management authority, proving it can handle more than 70 billion dollars in cash and treasuries on its own, no third party commercial bank needed. Second, an even bigger ask, a Fed master account. Get that approved, and Circle can park funds directly on the Fed’s books, zero counterparty risk, round the clock settlement, and no middleman left at all.

Mastercard is looking to sell its 51 percent stake in Vocalink back to a consortium of UK banks for around 400 million pounds. Vocalink processes 90 percent of the UK’s payroll and Faster Payments traffic, the actual lifeline of the country’s payment clearing system. The UK government and the Bank of England are pushing hard to take back sovereignty over clearing, rebuilding the next generation of the system around direct account to account transfers and tokenized deposits, cutting the traditional card networks out entirely.

The UK is essentially redefining payment clearing as sovereign level financial infrastructure. Down the line, a consumer’s payment moves straight from a bank account to a merchant account, no international card network in between. Tokenized sterling deposits turn bank deposits into programmable, self clearing digital assets on top of that. What the UK wants is a new clearing layer, run by domestic banks, built to handle both A2A payments and tokenized deposits.

This might not even be a bad outcome for Mastercard. Domestic retail clearing is heading toward thin margins and disintermediation anyway. Vocalink’s growth is limited, political scrutiny keeps piling up, and holding onto it could actually get in Mastercard’s way as the UK builds its next generation payment infrastructure. Rather than hold onto a low growth clearing asset under heavy domestic regulatory pressure, Mastercard would rather put its resources and strategic focus into BVNK, the stablecoin platform it acquired on March 17, chasing higher margin, borderless global B2B settlement and Web3 money networks instead.

Worth watching from here, whether the UK can actually turn Vocalink into tokenized deposit infrastructure, and whether Mastercard shifts its full weight toward borderless stablecoin payments and money orchestration through BVNK.

It took SWIFT just nine months to move its Hyperledger Besu based shared ledger from white paper to an actual live network. The first pilot has 17 banks on board, Citi, HSBC, UBS, BNY, Standard Chartered, and Lloyds among them, spanning six continents. That’s an unusual pace for an organization that’s been running for 53 years and connects more than 11,500 financial institutions.

What this actually fixes is moving tokenized deposits across banks. Up until now, a tokenized deposit could mostly only move inside a single bank. HSBC to HSBC works fine at midnight on a Sunday, but the moment money needs to leave HSBC and land at another bank, it hits the old settlement wall again. SWIFT’s shared ledger is trying to open that wall up, letting tokenized deposits issued by different banks move across institutions.

Banks are essentially taking back the selling point stablecoins have owned for the past few years, round the clock, instant, global settlement. SWIFT just went with a path that draws less regulatory friction. Onchain confirmation can happen outside business hours, but the legally final settlement still runs through the existing RTGS system. SWIFT doesn’t issue its own token, and it doesn’t push every bank onto one central chain. Each bank keeps issuing tokenized deposits on its own private ledger, and SWIFT just connects, translates, and coordinates between them.

For corporate treasury teams, this is an easier sell than a non bank stablecoin. A bank issued tokenized deposit doesn’t need to be reframed as some new crypto asset. On regulation, redemption, accounting, and internal risk controls, it looks a lot more like the bank deposits they already hold.

That said, big banks aren’t putting all their chips on one system. Citi, BNY, and Wells Fargo are in SWIFT’s pilot and also core members of The Clearing House’s tokenized deposit network, due to launch in 2027. JPMorgan isn’t on SWIFT’s list at all, having spent years building its own moat through Kinexys, formerly JPM Coin. None of these banks picked a single winner. They’re running with multiple networks at once, which says global financial infrastructure won’t end up on just one chain either. SWIFT’s shared ledger, The Clearing House, Kinexys, and a handful of regional stablecoins are going to coexist for a long stretch. SWIFT just helped banks break out of their own closed off silos, which is a real short term win. But who eventually connects all these disconnected systems together is the question that actually decides how this ends, and right now, nobody has an answer.

Kraken’s parent company Payward closed its acquisition of Hong Kong stablecoin payments firm Reap Technologies for around 600 million dollars. It’s Payward’s biggest acquisition in Asia, and it hands Kraken the hardest parts of stablecoin payments to build from scratch, compliance licenses, card issuing credentials, and enterprise payment rails. Reap holds Visa Principal Issuer status in both Hong Kong and Mexico, putting it among a small handful of stablecoin payments firms with card issuing networks spanning Asia and the Americas at once. In Singapore, it holds a Major Payment Institution license from the MAS. For Kraken, this deal buys a full compliance corridor that plugs straight into cross border payments, corporate spend, and stablecoin settlement, and Reap’s API is just one piece of that corridor.

This does more than fill a gap in Kraken’s B2B payments and card issuing lineup. It marks stablecoins actually entering real trade settlement. Hong Kong brings the regulatory standing, Reap brings the API and card network connections, and Kraken brings the liquidity. Put the three together, and a business can convert trading assets into stablecoins, then use Reap’s interface to settle cross border payments, corporate spend, or card transactions, all running 24/7 without being boxed in by bank hours.

Stablecoin value is moving out of exchanges and onchain finance into much more concrete business use, cross border trade, corporate payments, supplier settlement, local currency conversion. When SWIFT and traditional bank networks go dark on weekends, that inconvenience can turn into real financial losses. In parts of Latin America, produce farmers have lost up to 20 percent of their crop to spoilage because weekend payments couldn’t clear in time. Under this new model, funds convert into local currency in seconds without ever leaving the chain. The biggest winners are merchants trading across Latin America and Africa, where these settlement channels are time sensitive enough that waiting until Monday just isn’t an option.

Three things worth following from here. Whether Hong Kong’s stablecoin licensing framework, which issued its first batch back in April, pulls in more deals like this one. How Reap gets folded into NinjaTrader, Bitnomial, and Backed, the other assets Payward has picked up recently. And with stablecoin transaction volume already at 2.3 trillion dollars in 2025 according to Artemis Analytics and still climbing, how much corporate settlement flow ends up leaving traditional bank clearing networks for this kind of always on, API driven payment infrastructure.

Robinhood launched three new products in London. Robinhood Earn offers eligible US users a floating yield around 7 percent. Stock Tokens, available in more than 120 countries outside the US, trade 24 hours a day, though these are essentially bond like instruments tracking share price with no voting or shareholder rights attached. And perpetual futures for European users now cover gold, oil, and EUR/USD, with leverage up to 10x. Lloyd’s of London and RELM underwrite the whole lineup against smart contract and hacking risk.

Line up all three and the picture gets clearer. Earn lends out USDG through Morpho for spread. Stock Tokens hand out tradeable, pledgeable onchain assets. Perps absorb the demand for leveraged trading. Yield, collateral, credit, and trading all loop through the same system now, which is basically prime brokerage, just onchain, with the entry barrier removed. Coinbase buries its perps deep in a professional grade interface. Robinhood puts them right on the home screen. And as a licensed broker with years of securities infrastructure behind it, Robinhood can push a tightly regulated product like Stock Tokens through compliance channels that pure crypto platforms simply can’t reach yet, stitching traditional brokerage, trading, and DeFi lending into a single onchain gateway for everyday users.

Whether that 7 percent on Earn holds up is still the biggest variable in this whole business. Robinhood itself calls it a floating estimate, not a promise to maintain it long term. About half of that yield right now comes straight from marketing subsidies, which looks a lot more like customer acquisition cost in the short run. Robinhood’s bet is that as USDG scales up, float income off the reserve assets gradually takes over from the subsidy and becomes the real backbone of the yield. Pull that switch off successfully, and Earn turns from an acquisition cost into an actual sustainable business. Fail to, and it’s still just a customer acquisition cost dressed up as a yield.

Sony’s US subsidiary Connectia Trust just picked up preliminary conditional approval from the OCC to set up a national trust bank. Initial capital is set at 40 million dollars, with operations targeted for 2027. The restrictions match every other charter of this kind, no deposits, no lending, no FDIC coverage. What it can do is issue and manage dollar stablecoins, hold cash and treasuries directly as reserves, and provide digital asset custody.

What makes this different is who’s gotten these charters before. Every stablecoin trust charter the OCC has issued so far went to a crypto native firm or a traditional financial institution, Circle, Ripple, Fidelity. Sony is the first purely commercial tech giant to get one, arguably the first ecosystem bank ever built by a commercial conglomerate. The Bank Policy Institute pushed back hard, saying the application ran straight into the decades old separation between banking and commerce. The OCC approved it anyway.

So what happens once Sony actually puts this charter to work? Look at what Sony has already built.

Sony launched the Soneium Ethereum Layer 2 network back in 2025, runs a licensed exchange called S.BLOX in Japan, and has a yen stablecoin pilot already running. A dollar stablecoin is the piece that ties all of it together.

Sony doesn’t need years hunting for a use case the way early USDT or USDC did. PlayStation players worldwide can use it for cross border in game item trades. Crunchyroll can run merchandise sales and cross border payments through it. Independent creators under Sony Music and Sony Pictures can get paid globally without touching a traditional bank at all. Any single one of these use cases alone could carry a massive amount of money flow.

Most stablecoins get created first and then spend years hunting for somewhere to be useful. Sony flipped that order. It already has a massive, self sufficient economy sitting there. All it needed was the currency to make that economy run efficiently.

View more

Get started with Cobo Portal

Secure your digital assets for free