Crypto rails won. The crypto narrative didn't.

Two hype cycles, one pattern

Muzamil opens by framing Paul Dubé’s vantage point: a builder who lived the Web3 cycle from inside, now watching AI run a similar playbook at higher speed. Paul does not dodge the comparison. “The parallels are very, very, very overlapping right now,” he says. Most outsiders still think Bitcoin and Ethereum when they hear crypto. Builders know the real story is fifteen years of infrastructure that almost nobody understands.

That gap, between what was built and what was understood, is what Paul keeps returning to. “If you try to get into Web3 today, it’s not easy. You’ve got so many different segments, so much to understand. It would take a year or two just to understand it.” He sees the same fog descending on AI, but faster. “Every morning I wake up, there’s a new agent, a new app, a new tool, a new this, a new that. It’s hard for me to keep up.” The person building the rails is openly saying the surface layer has become unintelligible.

What Sovera is actually doing

Muzamil presses for specifics, and Paul walks through Sovera in plain terms. It is not a bank and does not hold licenses. It is an integrator and payment orchestrator that stitches licensed entities into a single stack. A business onboards funds, either fiat or stablecoins, into virtual accounts with custody and bank-level insurance. From there, the business can swap, hold, or off-ramp through a cards-as-a-service layer that lets it issue its own Mastercards to founders, employees, or affiliates anywhere in the world.

The use case Paul keeps returning to is the distributed company. Payroll on cards. Suppliers paid instantly across borders at a fraction of the wire cost. Liability moved off the parent entity. “The global fin stack today is just very fragmented. It’s a hot mess,” he tells Muzamil. Sovera’s bet is that mid-market companies, gaming firms, and eventually banks themselves will pay for a unified layer that hides the mess.

Built to be bought

The strategic insight underneath the product is sharper than the product itself. Paul built Sovera knowing the incumbents would need to acquire it. He names the buyers directly: “JP Morgan, City Bank, Kraken, Stripe, and the list goes on. I can name 10 other companies that have been acquiring companies like mine.” Companies doing ten to fifteen million in ARR are getting bought for hundreds of millions. The endgame was never to overthrow banks. It was to sell them the rails they were too slow to build themselves.

This reframes the entire Web3 narrative. The revolution didn’t fail. It got absorbed. Paul predicts the US Treasury will eventually issue its own stablecoin backed by bonds and other assets, and that the Clarity Act will crack the system open further. “I would not be surprised if the US Treasury Department ends up being a stable coin.” Less the cypherpunk dream, more a state-sanctioned upgrade to the same plumbing.

Why Visa and Mastercard still exist

Muzamil pushes on the original Web3 promise. If crypto was meant to remove gatekeepers, why is every modern stablecoin product still ultimately routed through Mastercard? Paul’s answer is the quietly damning part of the conversation.

There are two worlds, he says. One has no Mastercard. It runs on non-custodial wallets and a small group of people who know how to move money themselves. “That’s not bringing in the traditional, you know, it’s not bringing in my dad. My dad’s not grabbing a non-custodial wallet and doing this. My sister’s not. Many businesses aren’t.” So the industry built a compliance stack on top, and the off-ramp, the moment crypto becomes spendable, is still owned by Visa and Mastercard. They saw the shift coming, invested early, and are preparing their own tokens. The disruption got absorbed at the edge.

The tax on geography

Muzamil makes the conversation personal. Pakistan, the fifth largest country by population, has been structurally excluded from modern financial rails because the developed world built KYC infrastructure faster than emerging markets could meet the benchmark. He asks Paul directly: has any of this new infrastructure actually made it easier for a freelancer in Karachi or Manila to participate?

Paul’s answer is honest and uncomfortable. Short answer, yes. Long answer, you still have to register a company in the UK or somewhere similar, then onboard that entity into a platform like Sovera, then route earnings through it. “The developers don’t know that world for the most part. So we help educate them.” The freelancer can technically get paid in stablecoins, convert to fiat, and spend it on a Mastercard. But the workaround requires offshore incorporation, which is a fixed cost most freelancers cannot easily absorb.

Muzamil turns the question into a thesis. The current setup taxes geography, not talent. An American worker earning ten to twelve thousand dollars a month does not feel this. A Pakistani freelancer earning around two thousand dollars a month spends a meaningful share of income just to access the system. Paul does not disagree. He mentions that a minister in Pakistan was working with companies like his to solve exactly this. He also notes, dryly, that what governments dislike most is the freedom to move capital without their supervision.

Tokenization is solved. Legislation isn’t.

When Muzamil asks Paul to look past payments, the conversation lands on tokenization. Paul’s framing is the cleanest of the episode: “Tokenization of a real-world asset, kind of like a property, if you will, into digital puzzle pieces.” Slicing the Mona Lisa into tradable shares. Fractional ownership of a hundred Costa Rican beachfront properties, aggregated into a portfolio you can borrow against from inside your own app.

The economic logic is what excites him. Tokenization turns unproductive assets, like a house you live in, into productive ones, like collateral inside a global liquidity pool. The technology already exists. He helped build pieces of it. What’s missing, predictably, is legislation. “The technologists are ahead of the regulations.” Until that closes, retail participants keep getting hurt by the nefarious edges of the market, which poisons the narrative for the legitimate use cases. This is Paul’s through-line. The technology is rarely the bottleneck. Governance is.

Server farms, escrow bonds, and a crisis of thought

The conversation shifts to AI, and Paul offers an example Muzamil does not see coming. US cities and municipalities are courting AI server farm developers with grants and tax incentives in exchange for promised job creation, roughly 200 jobs per facility. The jobs frequently do not materialize, the environmental costs land on residents, and the developer can walk away.

His proposed fix is revealing of how he thinks. Force escrow bonds. Make the developer post collateral. “Let’s ensure that that location that it’s in is being adjudicated properly first for the safety of the people.” It is the same instinct that shaped Sovera, build the rails so the downstream user does not get rug-pulled, applied to physical infrastructure instead of finance.

He is less worried about AI taking jobs than about AI accelerating a thinking crisis that is already underway. “We have a crisis in thought. We don’t really have a crisis in technology and advance. We have a crisis in how we think.” He points to a study on handwriting and brain activation as a small example of what is being lost. Tools that let you sound like a genius from a prompt are also tools that let you stop reasoning.

Survival mode, and what wealth actually means

Muzamil offers a reframe near the end that Paul accepts. The whole jobs debate, he argues, assumes that survival requires employment. If AI and robotics genuinely compound productivity at the rates the absolutists predict, the right question is not whether jobs are created or destroyed. It is whether the new wealth gets distributed in a way that lets people stop living in survival mode at all.

Paul’s response is the most personal moment in the episode. “My ultimate personal hope is that it’s going to automate things to a level that we no longer have to live in survival. We live in creation, and we begin to work on our personal selves.” He is careful not to be utopian. When Muzamil asks him to project twenty-five years out, Paul splits his answer. The world he wants is one where intention becomes the architecture people live by. The world he expects, based on current trajectory, is more bifurcated and more robotic. “I don’t think we can be very optimistic.”

What Paul is actually watching

Muzamil closes by asking Paul for three founders or products that genuinely excite him. He names NoodleSeed, then Sovera itself, then Municipal AI, a product already deployed inside city governments. But the most interesting pick is the last one. Cryptoslam, originally an NFT data aggregator across roughly thirty blockchains, has pivoted into an AI platform called Ethos Swarm. Paul says around 3,000 agents now live on it, built through a tiered progression he describes as “baseline agents that would be like high school kid mentality,” graduated to university-level, then to professional. Animoca Brands’ executive team has built hundreds of agents on it. A token sits underneath, but it is abstracted away from the user.

The reason Paul highlights it tells you what he is actually betting on. Ethos Swarm is Web3 infrastructure that survived the narrative collapse by becoming useful inside the AI stack. Which is, in miniature, the entire argument of the episode. The crypto story died. The crypto rails are everywhere. The AI story will likely follow the same arc, and the builders who recognize the pattern early will own the layer everyone else has to license.