
Sign up to save your podcasts
Or


Binance just suffered the biggest hack in crypto history — $6.3 billion gone in under an hour. Not through stolen keys, but a hijacked signing system, a weakness far more mundane than blockchain cryptography. This episode traces how it happened, puts the number in context against far larger historical cases and fraud across traditional finance, and asks why markets barely flinched. It also looks ahead: quantum computing is a slow-building risk to the cryptography underneath it all, while AI agents are already starting to pay, trade and invest on their own — raising a new question about oversight, not just security.
In this episode:
How the Binance hack actually happened — a hijacked signing system, not stolen keys
Why the real weak point sits in the web infrastructure layer, not the blockchain itself
Borg Security’s AI-agent tests exposing vulnerabilities in eight Norwegian companies
Putting $6.3 billion in context: Mt. Gox, Bitfinex, and crypto crime versus global fraud
Why the market barely panicked — Tether and Circle’s fast freezes, and Binance’s own capital
Matthew Hougan’s case that a maturing industry survived what would once have ended it
Quantum computing’s slow-building threat to blockchain cryptography — and “quantum debt”
AI agents starting to shop, trade and manage money on their own
The shift from what an AI agent can do to what it’s been given permission to do
MiCA licensing, DTCC’s tokenization launch, and traditional finance stepping into the room
Is crypto finally merging with the rest of finance? For years the story was two separate worlds — onchain and everything else. This episode argues that line is gone. Bitcoin is now a line item for wealthy investors and a listed treasury asset. Stock exchanges are tokenising their own infrastructure while crypto exchanges start trading stocks, oil and silver. Stablecoins are moving through card networks, fintechs and remittance giants. Banks are splitting into two camps on digital money. And a thirty-year-old, unused piece of internet code is quietly becoming the payment rail for AI agents.
In this episode:
Why “is finance moving onchain” finally has a real answer
Bitcoin as an accepted asset class — and the rise of listed bitcoin treasury companies
Crypto exchanges becoming “everything exchanges”: Coinbase, Hyperliquid, Binance, Kraken
DTCC, Nasdaq, NYSE and ICE tokenising the rails Wall Street runs on
Stablecoins moving through Circle, Tether, Visa, Mastercard, Stripe and Revolut
Western Union and MoneyGram taking stablecoins into remittance
Europe’s banks split: Qivalis’s euro stablecoin vs. DNB’s tokenised deposits
Pontes, the ECB’s bridge to central bank money, and Swift’s answer with Chainlink
x402 and Cloudflare: the internet’s new payment layer for AI agents
DeFi’s Nordic founders, and why agents might be the users blockchain was missing
Is finance finally going onchain? The question could have been asked more or less every year for the last five or six years. This time the answer starts in an unexpected place: the European Central Bank, which has published official articles under the title Central banks on chain.
This episode follows the rails being built at every level. In the core of the financial system: a bridge to central bank money, tokenised deposits, a bank-issued euro stablecoin and the digital euro. In the market: Circle's own blockchain, the big fintech companies and DTCC's tokenised stocks. In regulation: the SEC, the EU's pilot regime and MiCA. And in a small Nordic business, where Kaupr is exploring agentic payments itself.
In this episode:
Why a question asked every year for five or six years deserves a new answer
Central banks on chain: Pontes and the bridge to central bank money
Tokenised deposits: DNB joins CBMT, and Norges Bank talks to twelve Norwegian banks
Qivalis and a euro stablecoin from 37 European banks
The digital euro on two fronts: a pilot and the negotiations in the European Parliament
Kaupr's own experiments with agentic payments and pay-per-use
Circle's Arc and the big fintech companies offering onchain infrastructure
DTCC, stock exchanges buying into crypto exchanges, and the SEC's five-year test
Europe's pilot regime, MiCA, and two Nordic examples: Centiglobe and NBX
Blockchain has been called a solution searching for a problem, and for a long time that has been fair. The technology came first, and the search for something to point it at came afterwards. Stockholm in early September offered a different order. At a blockathon inside Nordic Tech Week, the organisers wrote the challenges first — eight of them, and only two had anything to do with finance.
This episode follows what came out of that week. Three finalist teams all ended up building verification: that aid was delivered, that a crisis message is genuine, that an influencer approved what goes out under her name. It also takes up tokenisation, which is one word covering at least three different things, each with its own domain, rules and players. And it looks at what Sweden has that makes this possible, from a regulator willing to sit in the room to eight political parties answering questions about blockchain ahead of the election.
In this episode:
Why blockchain has spent fifteen years looking for a problem to solve
The blockathon where six of eight challenges had nothing to do with finance
Problem statements from the World Food Programme, the Swedish Red Cross, UNICEF and a dog shelter
Three finalists, three kinds of verification
Proof that comes from the person receiving the aid rather than the organisation reporting upward
Where AI needs the blockchain, and where the blockchain needs AI
Tokenisation as three separate domains: money, investing, and everything outside finance
What Blockchain Sweden's ten questions revealed about the political parties
Why blockchain on the main stage matters, and why it is not enough
Over five weeks this summer, three million people opened an account with Kalshi, an American exchange where you can trade on the outcome of almost anything. By the end of the World Cup, prediction markets accounted for more than a quarter of all legal sports betting in the United States — up from nine percent at the start of the year.
The same summer, traders on Polymarket put the odds at eighty-four percent that SpaceX would close its first day above 1.8 trillion dollars. They had priced one of the largest listings in history before the stock exchange had a price at all.
Football and a space company. So what are these markets actually becoming — and is this just gambling and insider trading?
In this episode:
What a prediction market is, and what a perpetual is — one ends when the event happens, the other when your collateral runs out
Why perpetuals now trade many times the volume of the assets they track
Kalshi and Polymarket: one licensed by the federal futures regulator, one running onchain
The University of Iowa market from 1988, which beat the polls three times out of four
The Golden Globes, where twenty-seven bets were placed and twenty-six won
The US soldier charged over Venezuela — caught because every trade sat on a public ledger
Why nobody knows yet whether any of it is illegal
The buyers who never place a bet: banks, funds and newsrooms reading the price as an indicator
Two appeals courts ruling opposite ways, and Europe reaching for a second MiCA
Last year, close to fifteen Nordic companies announced a bitcoin treasury strategy. Mining companies, rig companies, gaming companies. Everyone wanted to be first. Then bitcoin fell from its October peak, and the premium the whole model depended on disappeared.
Most of them left the scene quietly. Four are still standing, and each of them found a different way to earn from the same asset.
In this episode:
What a bitcoin treasury company actually is, and why growth in bitcoin per share only works while the premium lasts
Christoffer de Geer and Alexander Hagen launching the first two Nordic newcomers in spring 2025
H100 raising a quarter of a billion Swedish kronor in 48 hours — and owning more bitcoin than Aker by August
Torbjørn Bull Jenssen's warning in July 2025 about companies with no operational alpha, and how it reads today
The fall from $125,000, and the buyers who were no longer there
Four models still standing: H100, K33, Bitcoin Treasury Capital and Ace Digital
Tom Lee's ethereum treasury earning a yield on its own holdings, and Lyn Alden's cash-flow model
Why both ideas already have Nordic parallels
In this episode of Kaupr Weekly, host Morten Myrstad looks at what happens when buyers start purchasing computing power they have no intention of using. Oil, electricity and freight all made the same journey. None of them stopped being a cost — every company still books them as expenses. What changed is that each was given a standard unit and a price everyone could see, and from that point somebody who would never take delivery could take a position. That is financialization, and it is happening to compute right now.
Compute already has scarcity and a price that moves. What it does not have is a public price or a unit everyone agrees on. The question Morten asks is who is building those two missing pieces — and what it means that two of the world's largest exchanges, six of the largest financial institutions and a handful of startups all arrived at the answer within weeks of each other.
Inside the Episode:
Nvidia's August announcement: financing platforms with Apollo, KKR, Blackstone, Brookfield, BlackRock and Goldman Sachs to mobilise more than five hundred billion dollars — and why the participants say more about the coming product than the amount does.
Jensen Huang on chips as an investable asset class — and why what Nvidia has launched is chips, not compute: one part of the bottom layer of AI, with two more layers above it.
BlackRock's own technology team splits AI into three investment layers, and explains where the money has gone so far — and why the step that matters from here is packaging, not analysis.
The contracts: CME Group with Silicon Data listing compute futures on the fifth of October pending regulatory review, one contract equal to one month's rent for a single chip, the Intercontinental Exchange doing the same with Ornn, and ProShares and Rex Shares already filing for funds tied to them.
Outside the big exchanges: Compute Exchange, ten people in Palo Alto, pricing contracts in inference tokens rather than chip hours — and tokenized GPU forwards on a blockchain, with Falcon Finance in El Salvador and the AI arm of Near as the anchor buyer.
Bob Gelfond's 2013 argument in the Wall Street Journal, back on CNBC last week: compute as the currency itself, not a claim on compute.
The Nordic question: the capacity is built here, and every exchange, fund application and tokenized asset in this episode comes from somewhere else.
In this episode of Kaupr Weekly, host Morten Myrstad looks at a convergence that got measurable this summer. In one week, Hyperliquid — a platform built for crypto — turned over more in stocks, oil and silver than in crypto. In the same season, the New York Stock Exchange started connecting the engine that runs its markets to blockchain settlement, and the clearing house that keeps track of who owns what across the American market ran its first live trades on a blockchain.
Coinbase has a name for where all of this is heading: the everything exchange. One venue, every asset class, one settlement. The question Morten asks is how many different players are now building toward it — and what it means that crypto exchanges, stock exchanges, brokers, neobanks and token issuers have all arrived at the same idea from different directions.
Inside the Episode:
The crypto side moved first: Coinbase selling shares inside the app, Binance offering seven thousand American stocks, Kraken past twenty-five billion in tokenized equity volume — and why these firms have stopped being crypto-only exchanges.
Underneath the exchange: four hours of live trading on a blockchain at the clearing house holding a hundred and fourteen trillion dollars, with the New York Stock Exchange and Nasdaq in the room, and full service planned for October.
The European build: a hundred and sixty year old German stock exchange merging with a crypto firm — and a settlement platform that claims to cut costs by ninety percent.
The customers: close to a million people holding tokenized shares, most trading when Wall Street is closed, and institutions asking for one portfolio across every asset class.
The challengers and the sellers: Robinhood's own blockchain, Revolut opening private markets from one euro, and the Nordic firms positioning to sell crypto infrastructure to the banks.
Two regulatory routes: permission in the US, a pilot regime in Europe — and how each shows up in what gets built.
In this episode of Kaupr Weekly, host Morten Myrstad starts in a room in Oslo in September 2023, on stage at an event called Unleash the Future of Internet. The premise was convergence — AI, Web3 and AR/VR coming together. Three years on, he's honest about what that amounted to: a list of technologies, with no way to say how they'd fit. And when the AI wave arrived the following year, he stayed where he was, because he couldn't see the connection.
This summer, that changed. Sandy Kaul at Franklin Templeton published a paper calling agentic AI the killer use case for blockchain — not a crypto person arguing for crypto, but someone standing inside both, from inside asset management. Two weeks earlier, the Linux Foundation had made the x402 Foundation operational. Three weeks later, Cloudflare gave AI agents an identity and a wallet.
Along the way, Morten sits with a question: if agentic AI really is the killer use case for blockchain, does the sentence read just as well the other way around?
Inside the Episode:
What agents actually buy — compute, a data query, an API call, one at a time. Why a card can't carry that: a card gives you a promise, a blockchain gives you the money.
Response code 402 — set aside by the architects of the web in the early 1990s, labelled "payment required", left empty for thirty years. This story doesn't start with Bitcoin.
Jamie Burke's Post Web — read, write, own, and a fourth function: delegate. Plus the admission you rarely get from inside the industry.
What shipped this summer — the card networks and the blockchains governing one payment standard, and agent wallets from a company sitting in front of a fifth of the web.
What isn't built yet — Coldcard, Hugging Face, and the part no code solves: who answers for an agent that gets it wrong.
In this episode of Kaupr Weekly, host Morten Myrstad takes stock of the final stretch before Europe's biggest crypto regulatory deadline. On 1 July, MiCA's transition period ends — and across the Nordics and Baltics, the picture is split. Firi and K33 in Norway, Safello in Sweden, and a wave of Latvian firms have crossed the finish line. Bitmynt closed instead of applying. Veli got acquired. Kriptomat, eight years in business and fully solvent, is winding down anyway. And Binance — long the dominant exchange in this region — is caught in a dispute with Greek regulators that the industry reads as a test of how seriously Brussels will enforce MiCA against its biggest name.
Along the way, Morten asks a question he has been sitting with for a while: is the licence really the prize everyone thinks it is? Or is the real competition only just getting started?
Inside the Episode:
What MiCA actually changes: from a patchwork of national rules to one licence covering 450 million people — and why more than 2,500 firms are racing to convert before the deadline.
A fly on the wall: the frustration and lost-opportunity sadness Morten has heard firsthand in closed rooms across Stockholm, Oslo, Copenhagen and Helsinki — and why he doesn't believe it comes from bad intentions.
The Nordic and Baltic map: Finland's chaotic head start, Sweden's small steps, Denmark's surprise, Norway's late rush — and the Baltic story of Estonia's paradox versus Latvia's seven licences.
Who didn't make it: Bitmynt closes, Veli gets acquired, Kriptomat winds down solvent — three different exits, one preview of what's coming across Europe.
The Binance test: why a dispute with Greek regulators is being read as a far bigger test of MiCA's teeth.
The next fight: why the licence race was never the real prize — and the two new battles, for retail customers and for wealth managers, that are just getting started.
From the publisher's feed