November 16, 2026

The five functions of money
Money has five jobs. They don't change. The rails do.
Movement. Storage. Growth. Assurance. Credit. Five functions that defined the fiat world are migrating to programmable, always-on rails and the disciplines that picked winners on the old rails will pick them on the new ones.
I learned what money is for in Nairobi, not in a textbook. In 2016, while at Google, I was working on mobile financial services app launches with the Kenyan operators, and what struck me was how little the technology mattered to the customer. Nobody in Kibera cared about USSD or SIM toolkits. They cared that they could send their mother money before the shops closed, keep it somewhere it wouldn't be stolen, and borrow against next week's wages. Movement, storage, credit — on a Nokia. The functions were as old as money; the rail was new, and the rail changed who could perform them.
Later, at Wise and then Revolut, I ran cross-border products and saw the same thing from the other side: a modern payments company spending most of its ingenuity not on moving money but on pre-funding it, reconciling it, and explaining to regulators in the odd 20+ countries where it had been. That's the fiat world in one sentence. The functions of money are fixed. The cost of performing them is a function of the rail. And every few decades the rail changes.
It is changing again now.
{{5-funcs-timeline}}
Why a framework?
Most tokenization narratives decide nothing. This one is built to.
Most tokenization decks fail in one of two ways. They lapse into technology tourism—chains, wallets, oracles—which flatters the audience but decides nothing. Or they collapse into a market map of logos, which reads as coverage but confers no point of view.
QED's 'Five Functions' framework reframes tokenization not as a novel asset class but as a set of upgrades to a very old operating system, the one fintech has been rebuilding, piece by piece, for two decades.
Its power comes from a single assertion: the functions money performs are constant, but the rails change with every technology cycle: cash to cards, cards to mobile, mobile to PIX and UPI.
Tokenization is simply the next rail, the most consequential in a generation, because for the first time all five functions can be reconstituted on a single programmable, globally accessible, 24/7 settlement layer.
{{5-funcs-timeline-today}}
The five functions
Operator language, not economist language.
Every fintech balance sheet in the world resolves to some combination of five verbs. A neobank does storage and movement. A wealth platform does growth. An insurtech does assurance. A card issuer does credit and movement. The jobs do not change: the cost, reach, speed and composability of doing them changes profoundly.
{{5-functions-of-money}}
What tokenization changes
Not one change, but four physical properties, resetting every unit economic
{{5-funcs-insight}}
In the fiat world, five functions meant five institutions and five balances — a current account, a savings account, a brokerage, an insurer, a lender — and value leaked at every hand-off between all these players. A tokenized dollar balance performs all five at once: it is the payment instrument, the deposit and, through tokenized treasuries, the yield-bearing asset. The functions do not change; the boundaries between them collapse. Whoever owns the balance owns all five jobs. That is why distribution, not infrastructure, is where we expect the economics to consolidate and it is why we are happy to back the company that owns the customer, even when it doesn't own the token it distributes.
Where the money goes today is not a mystery. The largest pool of value in stablecoins is float income earned by the issuer on reserves, and the GENIUS Act protects it by barring issuers from paying yield to holders. The result is that consumer and business applications compete for distribution rebates rather than deposit spread. Movement, storage and growth are all shaped by that single fact, and it is why we are also happy to back companies that won the consumer, even if they don't won the underlying yield-bearing token they are distributing.
Function by function
Five jobs, one discipline: underwrite where the value actually accrues
{{5-funcs-msgac}}
Movement
I · Payments · Pay · Send · Collect
The job: Get value from one party to another, ideally instantly and at zero cost.
The fiat-world state
A fragmented stack of card networks, ACH, SWIFT and country-specific real-time schemes. Domestic payments have improved dramatically; cross-border has barely moved — remittance corridors still average 6 percent cost and multi-day settlement. A $200 trillion annual flow with structurally poor unit economics for the customer.
What tokenization does
Stablecoin settlement compresses cost to under 1 percent and time to seconds. 2025 volume of roughly $33 trillion (+72 percent YoY) already rivals the card networks, yet only ~1 percent of it is true end-user payments. That is not the ceiling; it is the floor. Consensus (including our recent annual state of fintech report with McKinsey) points to $2–4 trillion of stablecoin payment volume by 2030.
When I ran SMB Growth cross-border at Wise, the hard cost of a corridor was almost never the technology. It was the liquidity, the capital parked in local-currency accounts, so a payment could land before the sender's money had actually arrived and the compliance overhead of a dozen national regulators, each asking a slightly different question about the same dollar. Stablecoin settlement doesn't remove the compliance. It removes the pre-funding.
When the money and the message move together, in seconds, the working capital a corridor business needs collapses, and with it the moat that protected incumbents with big balance sheets. That is the change that matters, and it is why the winners in movement will be the companies that already understood corridors, not the ones that understand chains.
That is the demand curve behind the $310 billion of dollar stablecoins outstanding. It is not speculation. It is people in Lagos, Buenos Aires and Istanbul solving the storage problem their banking system cannot solve for them, and doing it with a phone. The function is ancient. The rail is new, and it makes a dollar account portable to anyone with an internet connection. For QED, whose footprint is deepest exactly where this problem is worst, that is not a trend to observe. It is the business.
Where durable value accrues
Corridor specialists: Last-mile fiat on/off-ramps in specific geographies. Félix Pago moves U.S.→LatAm remittances over WhatsApp on USDC settlement the sender never sees.
B2B orchestrators: Enterprise-grade compliance, treasury and accounting wrapped around stablecoin rails — Cedar Money and Cobre are QED bets here.
{{5-funcs-logo-movement}}
Storage
II · Savings & Custody · Save · Hold · Safeguard
The job: Store value where it will not be lost to theft, fraud, inflation or institutional failure.
The fiat-world state
The neobank era, think Nubank, Chime, Revolut, was a storage business: trust built through licensing, deposit insurance and superior digital experience. In emerging markets, well over a billion adults have no easy access to a stable-currency account. For a Lagos or Buenos Aires saver, the dollar is the unit of account long before it is the medium of exchange; a stablecoin simply makes that legal tender in the pocket.
What tokenization does
A regulated digital dollar account—globally portable, always on, holdable by anyone with a smartphone—is a genuinely new financial primitive. The $310 billion of stablecoin float outstanding, 99 percent dollar-denominated, is best read as a global demand curve for digital dollar savings.
I grew up in Nigeria, and for most of my adult life the dollar has been Nigeria's real unit of account. Rents, school fees, a car — quoted in dollars, paid in naira at whatever the rate was that morning. Between 2023 and 2025, the naira lost roughly 70percent of its value against the dollar, which means a saver who did everything right — earned, saved, avoided debt was essentially “punished” for holding the wrong currency. Storage was the function that failed. The domestic bank account was doing its job, safeguarding the naira. The naira was not. This is also true for Argentina, Venezuela and many other countries across the world.
Where durable value accrues
Stablecoin-native neobanks: KAST is QED's lead bet: a licensed spend-and-save experience denominated in stablecoins. Raenest does the analogous work for African dollar accounts.
Wallet & custody infrastructure: The picks-and-shovels layer, key management outsourced by enterprises and applications (Privy, Fireblocks, Utila, Dfns).
Identity, KYC & fraud: These functions don't disappear on-chain; they get rebuilt. Footprint (QED portfolio) sits exactly on this seam.
{{5-funcs-logo-security}}
Growth
III · Investing & Yield · Invest · Yield
The job: Compound value over time through productive assets and yield-bearing instruments.
The fiat-world state
Retail investing is a mature, distribution-dominated business. Whoever owns the customer relationship—Schwab, Robinhood, Nubank Invest, XP, Zerodha—captures most of the economics; asset issuers are largely commoditized.
What tokenization does
Tokenized real-world assets are moving on-chain at industrial scale: tokenized U.S. Treasuries went from under $1 billion in early 2024 to ~$13 billion by mid-2026, a 15x move in two years, anchored by BlackRock's BUIDL, Franklin Templeton and Circle's USYC. Stablecoin balances become natively yield-bearing, and collateral becomes a 24/7 mobile property.
Where durable value accrues
EM retail investment platforms: Midas, QED's Turkey/MENA bet, owns the retail investment pipe in a dollar-hungry emerging market: the distribution channel for tokenized U.S. assets.
The suppliers: Ondo, Securitize and peers convert traditional issuances into on chain form. Distributors buy from them; we don't expect the margin to sit there.
Yield-bearing dollar products: Yield from tokenized money market funds passed into spending accounts. In the U.S., GENIUS bars issuers from paying yield on the stablecoin itself, so this routes through tokenized funds; outside the U.S. it can be more direct. That is the reason KAST matters beyond storage.
{{5-funcs-logo-growth}}
Assurance
IV · Insurance · Insure · Hedge
The job: Transfer the risk of a loss event to a counterparty willing to bear it.
The fiat-world state
Insurance is where financial services is most broken and most protected by incumbency at once. Underwriting has always been a data business dressed up as an actuarial one; the insurtech generation proved data and distribution can rewire it.
What tokenization does
Tokenization doesn't do much for underwriting. Insurance is a data business wrapped in paper, and putting the paper on a chain doesn't change the data. Where tokenization may be more useful is around claims execution and collateral. A policy can pay out automatically on an on-chain verified trigger, which collapses days of claims administration into seconds; and a tokenized reserve can be seen, audited and moved in real time rather than attested every quarter.
One of the applications we are monitoring here is parametric: weather and crop cover from Arbol, flight and event cover from Etherisc and others, where the trigger is observable and the payout is mechanical. They are good add-ons. They do not yet touch the central risks an insurer carries, which still sit with whoever has the best loss data, on chain or off. Meanwhile, the life insurer denominated in bitcoin, is a different bet again: a change of denomination rather than of rail, and worth watching for what it says about demand for long dated savings outside fiat.
We are not forcing a bet here. Tokenization will also create new categories of risk: custody failure, smart contract exploits, depeg, issuer failure. Someone will underwrite them well. When they do, they will win the way insurers have always won, on data rather than the tech itself. Until then this is white space we should all monitor.
{{5-funcs-logo-assurance}}
Credit
V · Lending · Borrow · Lend
The job: Bring future purchasing power into the present against expected repayment.
The fiat-world state
Credit is QED's founding DNA—leadership forged at Capital One, arguably the most important credit underwriting institution of the past 30 years. The modern portfolio spans consumer, SMB, cards, BNPL, trade finance and specialty credit. Real-time loan performance is also the risk. There is no quarterly cushion when marks move.
What tokenization does
Digital-asset-backed lending collapses the operational cost of secured credit (Figure has pushed HELOCs on-chain at material volume). Stablecoin-settled trade finance eliminates settlement drag from cross-border SMB credit. On-chain private credit replaces quarterly LP letters with real-time loan-performance transparency.
A caution, because this is the function most likely to be pitched to us badly. Tokenized rails lower the cost of originating, settling and servicing a loan. They do not lower the cost of being wrong about who to lend to, at what rate and for how long, and they do nothing for a business whose lending economics are a whipsaw of the macro cycle. The lenders that have compounded are hybrid: Capital One, Nubank, the best of our own portfolio.
An always-on relationship where they own the customer and usually have a balance or a flow that the credit sits inside. That is the same principle as the rest of this piece. Whoever owns the customer or the balance owns the five jobs, and credit is the job you earn the right to do last, not the one you start with. Within the QED portfolio, Finkargo and OatFi qualify because they sit inside an import or a payables flow.
Also on the credit side, there is a capital markets angle. One of the biggest opportunities here may be away from the borrower and in the collateral. Wholesale secured lending, such as repo, securities lending and margin, runs on collateral that moves slowly and settles in batches. Each custodian keeps its own private ledger, so nobody can see the whole chain end to end. A tokenized treasury or fund share can be pledged and recalled in minutes, at any hour, with the terms of its reuse written into the asset, essentially a "smart contract." That puts idle collateral to work. It also brings assets that have never been practical to lend against, like private credit, fund interests and receivables, into secured lending. For secured credit this partly answers the caution above: when collateral is marked and enforceable in real time, the cost of being wrong falls.
There is a catch. Faster reuse of the same collateral is how leverage builds unseen. It is why clients of Lehman and MF Global found their collateral had been reused and couldn’t get it back, and it helped sink a string of crypto lenders in 2022. So the opportunity is collateral that moves fast with limits built in. It is our view that while incumbents will build much of this by upgrading their own plumbing, they will still need infrastructure they don’t build themselves, and lenders willing to lend against assets that weren’t lendable before. That is where we are actively hunting.
Where durable value accrues
Trade & working-capital finance: Finkargo (LatAm import finance) and OatFi (B2B working-capital infrastructure) sit directly on flows whose logical settlement layer is stablecoin.
Collateralized digital-asset lending: Tokenized Treasuries, equities and stablecoin holdings as instant, programmable collateral.
On-chain private credit: Maple, Centrifuge and emerging institutional protocols, credit funds migrating to programmable rails.
{{5-funcs-logo-credit}}
Cross-cutting implications
Three structural truths cut across all five functions
{{5-funcs-insight-stack}}
What we rule out
A discipline is only as strong as what it disqualifies
{{5-funcs-exclusions}}
If a company is not doing movement, storage, growth, assurance or credit in a way that gets better on tokenized rails, we default to a pass.
Why QED is advantaged
Payments has been QED's densest investment area for two decades. Picking corridor winners, understanding interchange economics and evaluating fraud at real payment volumes is a QED specialty. Every neobank ultimately becomes a storage business, too. Nubank started in credit and earned its way into deposits. Revolut started with travel and is earning its way into deposits.
QED has underwritten investment distribution repeatedly across LatAm and beyond. We know the difference between a distribution moat and a product feature. Underwriting risk on data is the specific muscle QED's insurtech portfolio has built for years, and when on-chain assurance matures, operators who price risk on data, not ideology, win. That is a fight QED is built for. Finally, credit is QED's craft: distinguishing lending businesses from fintechs pretending to be one, charge-offs earned from charge-offs deferred. On-chain credit is a straight-line extension, not a stretch.
Three bets
1. By 2030, the majority of cross-border B2B trade between Africa and China, and between Latin America and Asia, will settle in stablecoins rather than through correspondent banks. Both corridors have the worst fiat economics in the world: multi day settlement, rationed dollar liquidity, FX spreads that would embarrass a bureau de change. It is where the pre funding problem bites hardest that the rail flips first.
We see the same stablecoin volumes and the same density of fintech activity on both sides of the Atlantic. Brazil has a direct settlement arrangement between the real and the yuan with Beijing, and importers still route through stablecoins, because the bank leg is slow and the dollar leg is rationed. Brazil also shows what the finished product looks like: PIX settles the domestic leg in real time and stablecoins settle the international leg in real time, so for the first time the whole journey is instant rather than one fast leg bolted to one slow one. That combination is the product. If most Africa to China or Brazil to China trade is still settling through New York in 2030, I was wrong.
2. By 2030, in at least three of the five largest dollar hungry economies (Nigeria, Argentina, Turkey, Egypt and Pakistan) the largest single custodian of household dollar savings will be a licensed stablecoin wallet or neobank, not a traditional domestic bank. The pattern is structural: wherever a weak currency, capital controls and a smartphone coexist, savers hold dollars by whatever route is open to them. Argentines already hold more dollar stablecoins per capita than almost anyone on earth; Turks moved into them through the lira collapse of 2022 and 2023; Nigerians did the same as the naira fell.
The central risk to this bet is not technology. It is sovereignty. No central bank will give up control of its monetary perimeter willingly, and several have already pushed back. Nigeria cut banks off from crypto rails in 2021 before licensing the sector three years later; Argentina, Turkey and India have each tightened. Some countries will fight this and some will win, at least for a while. That is why the bet is about licensed entities, and why QED only backs companies that start from "how do we get licensed" and build backwards. We are not in the business of helping anyone evade the law of the country they live in. We are in the business of backing the companies that regulators eventually license because the customers are already there. It is the mobile money story in East Africa, replayed: tolerate, then licence, then claim credit. If in 2030 the biggest retail dollar balances in those countries still sit in traditional domestic banks, this framework has overstated the migration and I will say so.
3. By 2030, stablecoin issuers will retain less than half of the yield on their reserves. Today the issuer keeps the float and the GENIUS Act forbids paying it to holders. That is a statutory margin, and statutory margins get competed away; through distribution rebates to the wallets that own the customer, and through tokenized money market wrappers that pay yield legally. It has already started. Coinbase takes all of the reserve income on USDC held on its platform and roughly half of the rest. The value in stablecoins moves from whoever mints the dollar to whoever holds the relationship. That is the neobank story again, and it is the one QED has been underwriting for 19 years.
The takeaway
{{5-functions-of-money-takeaway}}
Money has five jobs. They don't change. The rails do.
Movement. Storage. Growth. Assurance. Credit. Five functions that defined the fiat world are migrating to programmable, always-on rails and the disciplines that picked winners on the old rails will pick them on the new ones.
I learned what money is for in Nairobi, not in a textbook. In 2016, while at Google, I was working on mobile financial services app launches with the Kenyan operators, and what struck me was how little the technology mattered to the customer. Nobody in Kibera cared about USSD or SIM toolkits. They cared that they could send their mother money before the shops closed, keep it somewhere it wouldn't be stolen, and borrow against next week's wages. Movement, storage, credit — on a Nokia. The functions were as old as money; the rail was new, and the rail changed who could perform them.
Later, at Wise and then Revolut, I ran cross-border products and saw the same thing from the other side: a modern payments company spending most of its ingenuity not on moving money but on pre-funding it, reconciling it, and explaining to regulators in the odd 20+ countries where it had been. That's the fiat world in one sentence. The functions of money are fixed. The cost of performing them is a function of the rail. And every few decades the rail changes.
It is changing again now.
{{5-funcs-timeline}}
Why a framework?
Most tokenization narratives decide nothing. This one is built to.
Most tokenization decks fail in one of two ways. They lapse into technology tourism—chains, wallets, oracles—which flatters the audience but decides nothing. Or they collapse into a market map of logos, which reads as coverage but confers no point of view.
QED's 'Five Functions' framework reframes tokenization not as a novel asset class but as a set of upgrades to a very old operating system, the one fintech has been rebuilding, piece by piece, for two decades.
Its power comes from a single assertion: the functions money performs are constant, but the rails change with every technology cycle: cash to cards, cards to mobile, mobile to PIX and UPI.
Tokenization is simply the next rail, the most consequential in a generation, because for the first time all five functions can be reconstituted on a single programmable, globally accessible, 24/7 settlement layer.
{{5-funcs-timeline-today}}
The five functions
Operator language, not economist language.
Every fintech balance sheet in the world resolves to some combination of five verbs. A neobank does storage and movement. A wealth platform does growth. An insurtech does assurance. A card issuer does credit and movement. The jobs do not change: the cost, reach, speed and composability of doing them changes profoundly.
{{5-functions-of-money}}
What tokenization changes
Not one change, but four physical properties, resetting every unit economic
{{5-funcs-insight}}
In the fiat world, five functions meant five institutions and five balances — a current account, a savings account, a brokerage, an insurer, a lender — and value leaked at every hand-off between all these players. A tokenized dollar balance performs all five at once: it is the payment instrument, the deposit and, through tokenized treasuries, the yield-bearing asset. The functions do not change; the boundaries between them collapse. Whoever owns the balance owns all five jobs. That is why distribution, not infrastructure, is where we expect the economics to consolidate and it is why we are happy to back the company that owns the customer, even when it doesn't own the token it distributes.
Where the money goes today is not a mystery. The largest pool of value in stablecoins is float income earned by the issuer on reserves, and the GENIUS Act protects it by barring issuers from paying yield to holders. The result is that consumer and business applications compete for distribution rebates rather than deposit spread. Movement, storage and growth are all shaped by that single fact, and it is why we are also happy to back companies that won the consumer, even if they don't won the underlying yield-bearing token they are distributing.
Function by function
Five jobs, one discipline: underwrite where the value actually accrues
{{5-funcs-msgac}}
Movement
I · Payments · Pay · Send · Collect
The job: Get value from one party to another, ideally instantly and at zero cost.
The fiat-world state
A fragmented stack of card networks, ACH, SWIFT and country-specific real-time schemes. Domestic payments have improved dramatically; cross-border has barely moved — remittance corridors still average 6 percent cost and multi-day settlement. A $200 trillion annual flow with structurally poor unit economics for the customer.
What tokenization does
Stablecoin settlement compresses cost to under 1 percent and time to seconds. 2025 volume of roughly $33 trillion (+72 percent YoY) already rivals the card networks, yet only ~1 percent of it is true end-user payments. That is not the ceiling; it is the floor. Consensus (including our recent annual state of fintech report with McKinsey) points to $2–4 trillion of stablecoin payment volume by 2030.
When I ran SMB Growth cross-border at Wise, the hard cost of a corridor was almost never the technology. It was the liquidity, the capital parked in local-currency accounts, so a payment could land before the sender's money had actually arrived and the compliance overhead of a dozen national regulators, each asking a slightly different question about the same dollar. Stablecoin settlement doesn't remove the compliance. It removes the pre-funding.
When the money and the message move together, in seconds, the working capital a corridor business needs collapses, and with it the moat that protected incumbents with big balance sheets. That is the change that matters, and it is why the winners in movement will be the companies that already understood corridors, not the ones that understand chains.
That is the demand curve behind the $310 billion of dollar stablecoins outstanding. It is not speculation. It is people in Lagos, Buenos Aires and Istanbul solving the storage problem their banking system cannot solve for them, and doing it with a phone. The function is ancient. The rail is new, and it makes a dollar account portable to anyone with an internet connection. For QED, whose footprint is deepest exactly where this problem is worst, that is not a trend to observe. It is the business.
Where durable value accrues
Corridor specialists: Last-mile fiat on/off-ramps in specific geographies. Félix Pago moves U.S.→LatAm remittances over WhatsApp on USDC settlement the sender never sees.
B2B orchestrators: Enterprise-grade compliance, treasury and accounting wrapped around stablecoin rails — Cedar Money and Cobre are QED bets here.
{{5-funcs-logo-movement}}
Storage
II · Savings & Custody · Save · Hold · Safeguard
The job: Store value where it will not be lost to theft, fraud, inflation or institutional failure.
The fiat-world state
The neobank era, think Nubank, Chime, Revolut, was a storage business: trust built through licensing, deposit insurance and superior digital experience. In emerging markets, well over a billion adults have no easy access to a stable-currency account. For a Lagos or Buenos Aires saver, the dollar is the unit of account long before it is the medium of exchange; a stablecoin simply makes that legal tender in the pocket.
What tokenization does
A regulated digital dollar account—globally portable, always on, holdable by anyone with a smartphone—is a genuinely new financial primitive. The $310 billion of stablecoin float outstanding, 99 percent dollar-denominated, is best read as a global demand curve for digital dollar savings.
I grew up in Nigeria, and for most of my adult life the dollar has been Nigeria's real unit of account. Rents, school fees, a car — quoted in dollars, paid in naira at whatever the rate was that morning. Between 2023 and 2025, the naira lost roughly 70percent of its value against the dollar, which means a saver who did everything right — earned, saved, avoided debt was essentially “punished” for holding the wrong currency. Storage was the function that failed. The domestic bank account was doing its job, safeguarding the naira. The naira was not. This is also true for Argentina, Venezuela and many other countries across the world.
Where durable value accrues
Stablecoin-native neobanks: KAST is QED's lead bet: a licensed spend-and-save experience denominated in stablecoins. Raenest does the analogous work for African dollar accounts.
Wallet & custody infrastructure: The picks-and-shovels layer, key management outsourced by enterprises and applications (Privy, Fireblocks, Utila, Dfns).
Identity, KYC & fraud: These functions don't disappear on-chain; they get rebuilt. Footprint (QED portfolio) sits exactly on this seam.
{{5-funcs-logo-security}}
Growth
III · Investing & Yield · Invest · Yield
The job: Compound value over time through productive assets and yield-bearing instruments.
The fiat-world state
Retail investing is a mature, distribution-dominated business. Whoever owns the customer relationship—Schwab, Robinhood, Nubank Invest, XP, Zerodha—captures most of the economics; asset issuers are largely commoditized.
What tokenization does
Tokenized real-world assets are moving on-chain at industrial scale: tokenized U.S. Treasuries went from under $1 billion in early 2024 to ~$13 billion by mid-2026, a 15x move in two years, anchored by BlackRock's BUIDL, Franklin Templeton and Circle's USYC. Stablecoin balances become natively yield-bearing, and collateral becomes a 24/7 mobile property.
Where durable value accrues
EM retail investment platforms: Midas, QED's Turkey/MENA bet, owns the retail investment pipe in a dollar-hungry emerging market: the distribution channel for tokenized U.S. assets.
The suppliers: Ondo, Securitize and peers convert traditional issuances into on chain form. Distributors buy from them; we don't expect the margin to sit there.
Yield-bearing dollar products: Yield from tokenized money market funds passed into spending accounts. In the U.S., GENIUS bars issuers from paying yield on the stablecoin itself, so this routes through tokenized funds; outside the U.S. it can be more direct. That is the reason KAST matters beyond storage.
{{5-funcs-logo-growth}}
Assurance
IV · Insurance · Insure · Hedge
The job: Transfer the risk of a loss event to a counterparty willing to bear it.
The fiat-world state
Insurance is where financial services is most broken and most protected by incumbency at once. Underwriting has always been a data business dressed up as an actuarial one; the insurtech generation proved data and distribution can rewire it.
What tokenization does
Tokenization doesn't do much for underwriting. Insurance is a data business wrapped in paper, and putting the paper on a chain doesn't change the data. Where tokenization may be more useful is around claims execution and collateral. A policy can pay out automatically on an on-chain verified trigger, which collapses days of claims administration into seconds; and a tokenized reserve can be seen, audited and moved in real time rather than attested every quarter.
One of the applications we are monitoring here is parametric: weather and crop cover from Arbol, flight and event cover from Etherisc and others, where the trigger is observable and the payout is mechanical. They are good add-ons. They do not yet touch the central risks an insurer carries, which still sit with whoever has the best loss data, on chain or off. Meanwhile, the life insurer denominated in bitcoin, is a different bet again: a change of denomination rather than of rail, and worth watching for what it says about demand for long dated savings outside fiat.
We are not forcing a bet here. Tokenization will also create new categories of risk: custody failure, smart contract exploits, depeg, issuer failure. Someone will underwrite them well. When they do, they will win the way insurers have always won, on data rather than the tech itself. Until then this is white space we should all monitor.
{{5-funcs-logo-assurance}}
Credit
V · Lending · Borrow · Lend
The job: Bring future purchasing power into the present against expected repayment.
The fiat-world state
Credit is QED's founding DNA—leadership forged at Capital One, arguably the most important credit underwriting institution of the past 30 years. The modern portfolio spans consumer, SMB, cards, BNPL, trade finance and specialty credit. Real-time loan performance is also the risk. There is no quarterly cushion when marks move.
What tokenization does
Digital-asset-backed lending collapses the operational cost of secured credit (Figure has pushed HELOCs on-chain at material volume). Stablecoin-settled trade finance eliminates settlement drag from cross-border SMB credit. On-chain private credit replaces quarterly LP letters with real-time loan-performance transparency.
A caution, because this is the function most likely to be pitched to us badly. Tokenized rails lower the cost of originating, settling and servicing a loan. They do not lower the cost of being wrong about who to lend to, at what rate and for how long, and they do nothing for a business whose lending economics are a whipsaw of the macro cycle. The lenders that have compounded are hybrid: Capital One, Nubank, the best of our own portfolio.
An always-on relationship where they own the customer and usually have a balance or a flow that the credit sits inside. That is the same principle as the rest of this piece. Whoever owns the customer or the balance owns the five jobs, and credit is the job you earn the right to do last, not the one you start with. Within the QED portfolio, Finkargo and OatFi qualify because they sit inside an import or a payables flow.
Also on the credit side, there is a capital markets angle. One of the biggest opportunities here may be away from the borrower and in the collateral. Wholesale secured lending, such as repo, securities lending and margin, runs on collateral that moves slowly and settles in batches. Each custodian keeps its own private ledger, so nobody can see the whole chain end to end. A tokenized treasury or fund share can be pledged and recalled in minutes, at any hour, with the terms of its reuse written into the asset, essentially a "smart contract." That puts idle collateral to work. It also brings assets that have never been practical to lend against, like private credit, fund interests and receivables, into secured lending. For secured credit this partly answers the caution above: when collateral is marked and enforceable in real time, the cost of being wrong falls.
There is a catch. Faster reuse of the same collateral is how leverage builds unseen. It is why clients of Lehman and MF Global found their collateral had been reused and couldn’t get it back, and it helped sink a string of crypto lenders in 2022. So the opportunity is collateral that moves fast with limits built in. It is our view that while incumbents will build much of this by upgrading their own plumbing, they will still need infrastructure they don’t build themselves, and lenders willing to lend against assets that weren’t lendable before. That is where we are actively hunting.
Where durable value accrues
Trade & working-capital finance: Finkargo (LatAm import finance) and OatFi (B2B working-capital infrastructure) sit directly on flows whose logical settlement layer is stablecoin.
Collateralized digital-asset lending: Tokenized Treasuries, equities and stablecoin holdings as instant, programmable collateral.
On-chain private credit: Maple, Centrifuge and emerging institutional protocols, credit funds migrating to programmable rails.
{{5-funcs-logo-credit}}
Cross-cutting implications
Three structural truths cut across all five functions
{{5-funcs-insight-stack}}
What we rule out
A discipline is only as strong as what it disqualifies
{{5-funcs-exclusions}}
If a company is not doing movement, storage, growth, assurance or credit in a way that gets better on tokenized rails, we default to a pass.
Why QED is advantaged
Payments has been QED's densest investment area for two decades. Picking corridor winners, understanding interchange economics and evaluating fraud at real payment volumes is a QED specialty. Every neobank ultimately becomes a storage business, too. Nubank started in credit and earned its way into deposits. Revolut started with travel and is earning its way into deposits.
QED has underwritten investment distribution repeatedly across LatAm and beyond. We know the difference between a distribution moat and a product feature. Underwriting risk on data is the specific muscle QED's insurtech portfolio has built for years, and when on-chain assurance matures, operators who price risk on data, not ideology, win. That is a fight QED is built for. Finally, credit is QED's craft: distinguishing lending businesses from fintechs pretending to be one, charge-offs earned from charge-offs deferred. On-chain credit is a straight-line extension, not a stretch.
Three bets
1. By 2030, the majority of cross-border B2B trade between Africa and China, and between Latin America and Asia, will settle in stablecoins rather than through correspondent banks. Both corridors have the worst fiat economics in the world: multi day settlement, rationed dollar liquidity, FX spreads that would embarrass a bureau de change. It is where the pre funding problem bites hardest that the rail flips first.
We see the same stablecoin volumes and the same density of fintech activity on both sides of the Atlantic. Brazil has a direct settlement arrangement between the real and the yuan with Beijing, and importers still route through stablecoins, because the bank leg is slow and the dollar leg is rationed. Brazil also shows what the finished product looks like: PIX settles the domestic leg in real time and stablecoins settle the international leg in real time, so for the first time the whole journey is instant rather than one fast leg bolted to one slow one. That combination is the product. If most Africa to China or Brazil to China trade is still settling through New York in 2030, I was wrong.
2. By 2030, in at least three of the five largest dollar hungry economies (Nigeria, Argentina, Turkey, Egypt and Pakistan) the largest single custodian of household dollar savings will be a licensed stablecoin wallet or neobank, not a traditional domestic bank. The pattern is structural: wherever a weak currency, capital controls and a smartphone coexist, savers hold dollars by whatever route is open to them. Argentines already hold more dollar stablecoins per capita than almost anyone on earth; Turks moved into them through the lira collapse of 2022 and 2023; Nigerians did the same as the naira fell.
The central risk to this bet is not technology. It is sovereignty. No central bank will give up control of its monetary perimeter willingly, and several have already pushed back. Nigeria cut banks off from crypto rails in 2021 before licensing the sector three years later; Argentina, Turkey and India have each tightened. Some countries will fight this and some will win, at least for a while. That is why the bet is about licensed entities, and why QED only backs companies that start from "how do we get licensed" and build backwards. We are not in the business of helping anyone evade the law of the country they live in. We are in the business of backing the companies that regulators eventually license because the customers are already there. It is the mobile money story in East Africa, replayed: tolerate, then licence, then claim credit. If in 2030 the biggest retail dollar balances in those countries still sit in traditional domestic banks, this framework has overstated the migration and I will say so.
3. By 2030, stablecoin issuers will retain less than half of the yield on their reserves. Today the issuer keeps the float and the GENIUS Act forbids paying it to holders. That is a statutory margin, and statutory margins get competed away; through distribution rebates to the wallets that own the customer, and through tokenized money market wrappers that pay yield legally. It has already started. Coinbase takes all of the reserve income on USDC held on its platform and roughly half of the rest. The value in stablecoins moves from whoever mints the dollar to whoever holds the relationship. That is the neobank story again, and it is the one QED has been underwriting for 19 years.
The takeaway
{{5-functions-of-money-takeaway}}


















