On-Chain Liquidity for Reinsurance Capital: Interview With Veritas CEO Amaury Dalleur 

Reinsurance capital is traditionally fully collateralized and largely illiquid — once committed to a cover, it remains locked until expiry. At the same time, major risks such as cyber incidents and large-scale data center outages remain significantly under-insured in global markets.

Diving into parametric insurance, on-chain mechanisms, and the secondary market for traditionally illiquid risks, Olayimika Oyebanji recently sat down with Amaury Dalleur, CEO of Veritas, to discuss the practical challenges and opportunities in building on-chain liquidity for reinsurance capital.

Can you briefly tell us about yourself and your professional background?

Sure. I’m Amaury, founder and CEO of Veritas. I also lead tokenization at Singularity Venture Hub — which basically means taking real-world financial things and putting them on a blockchain so they can move around more easily.

I have spent my career at that intersection of finance and crypto. Started from something that honestly surprised me: reinsurance — which is just insurance for insurance companies — is this enormous pool of money, hundreds of billions of dollars, and a significant part of it can’t move once it’s committed. It gets locked up until the contract ends. That needs to change.

What does “making reinsurance capital liquid” actually mean in practice, and why has this slice of the market historically been so illiquid?

Let me unpack it, because it sounds more complicated than it is. Reinsurance capital is huge — but a big chunk of it gets committed to a contract and then just sits there, frozen, until that contract ends. You can’t easily pull it out, and you can’t sell your position early. This world — investors putting money behind insurance risk to earn the premium — has a name: ILS, insurance-linked securities. That market is about $120 billion in size.

The one corner of it that already trades a bit is “cat bonds” — catastrophe bonds — basically a bond that pays you nice interest, but you lose the money if a big hurricane or earthquake hits; those change hands here and there. The rest — maybe $25 to $35 billion of it — is effectively impossible to trade at all. Your money’s locked until the very end.

So “making it liquid” means building a place where you can sell your slice partway through and let someone else step in. It never really existed before, and a blockchain is finally the right set of rails to make it work — you can trade one contract at a time, with a live price everyone can see.

Parametric insurance relies on objective triggers. How do you see the role of oracles and on-chain data evolving to support reliable triggers for risks like cyber and data centers?

So “parametric” just means the payout is automatic. Instead of filing a claim and waiting months for someone to inspect the damage and argue about it, the contract pays the moment a measurable thing crosses a line — a wind speed, a flood level, a data center being down for a certain number of hours.

The “oracle” is simply the piece of software that reads that real-world data and reports it to the contract. The rule is: never trust a single source. Pulling several independent feeds, cross-checking them, and adding a bit of smoothing prevents one weird data blip triggering a payout by accident. This is the point where working with Chainlink — the biggest oracle network — is crucial.

For cyber and data centers, the whole game is agreeing on what to measure: “the cloud was down this long,” “this facility lost power.” And the tooling is finally good enough that those things are objectively measurable and hard to fake. That’s what makes them insurable in the first place.

What makes cyber and large-scale data center risks particularly challenging for traditional reinsurance markets?

The reason is simple. Traditional insurance runs on history — decades of data to price a risk. Cyber and AI data centers just don’t have that. They are new, they are all interconnected, and one event can hit everybody at the same moment.

So insurers get nervous and don’t fully cover them — and the gap is honestly wild. On cyber: only about half of the companies that could buy cover actually have it, and cyber is less than 1% of all the insurance premium in the world (that’s Munich Re) — while cyber-crime, by the most-cited estimate, is set to cost the world around $10.5 trillion a year according to Cybersecurity Ventures.

The example everyone remembers is the 2024 CrowdStrike outage: it cost the Fortune 500 alone around $5.4 billion, and only about a billion of that was insured. So roughly 80 to 90% of the bill simply wasn’t covered. Data centers are the same story — S&P says there’s over $2 trillion of insurable value out there, and a single big AI campus, worth $20 to $30 billion, is often only a third to half insured; the rest, the company just eats itself.

Parametric flips it: you don’t need decades of history, you need a clear trigger and someone willing to price it. Even S&P now says closing this gap will need money from outside the insurance industry. And honestly, we’ve got natural first demand in our own backyard — the AI ecosystem we’re part of, through SingularityNet, runs compute and data centers that need exactly this kind of cover.

How does bringing insurance contracts on-chain and enabling a secondary market change the traditional collateral lock-up period reinsurers face?

Today, if you back a cover, you’re married to it for the whole term — no exit. And that lock-up costs you: you demand a higher return just to compensate for being stuck, and it caps how much the whole market can cover at any given moment.

Now put that position on a blockchain — “tokenize” it, which basically means turning it into something you can hold and transfer — and open a marketplace where people trade it, priced by how likely the bad event looks right now. Suddenly you can sell out, or free up your money, halfway through. And the important part: the cash protecting the customer never leaves — the cover stays fully funded the entire time. What moves is your capital behind it.

What are the biggest differences between tokenizing Treasury bills (as Ondo Finance has done) and creating liquidity in parametric reinsurance contracts?

Good question! Oftentimes people lump all “tokenization” together. Tokenizing a Treasury bill — what Ondo does — is easy, because a T-bill is already liquid and every unit is identical; what you are just doing is putting an already-tradeable thing on-chain.

An insurance cover is the opposite: every single one is different — different trigger, different length, different odds — and there’s no existing market price for it. So the hard part was never the token. The hard part is creating a price and a market for something that’s one-of-a-kind. That’s the real work — an order book, buyers and sellers, and a fair price everyone can point to. Ondo made an already-liquid thing easier to move around; we have to manufacture the liquidity from scratch.

From a capital provider’s perspective, what are the main risks or concerns versus traditional private placements?

Totally fair to ask and I would rather just be upfront. There are certain concerns that’s worth talking about.

“Will I get paid?” Yes — the money’s 100% set aside in a bankruptcy-remote structure in Bermuda.

“What if the trigger doesn’t match the real loss?” That’s called basis risk — there are triggers to track the actual loss closely, use multiple data sources, and we walk away from risks where that gap is too big.

“Is the code safe?”

An independent security audit is a hard gate before we go live with real money. And yes — a tradeable position can move against you mid-way, but you’re never forced to sell, and the customer’s cover is never affected. Net-net, versus a traditional private deal you gain an exit you never had, without giving up any of the safety.

How far are we from a functional secondary market for these exposures?

Closer than people think . The tech works today; it is running on a test network right now. The starting point at the moment is working directly with lead reinsurance companies to choose the first pilots to bring on-chain — and starting, on purpose, with extreme weather is a deliberate move because everyone already knows how to price it, so can the rails can be proven on a risk the whole industry is comfortable with.

From there things move up the ladder to the harder, higher-value stuff — cyber and data-center outages — where the real spread is. So a first, real market on the familiar risks is a near-term, concrete thing; the exotic stuff follows as the book and the trust grow.

What are the biggest barriers — technical, regulatory, or market-acceptance — to scaling this?

Honestly, the technology’s the easy part now. The two real hurdles are rules and trust. For rules: we have to keep this unambiguously insurance — not gambling, not a derivative — which is why using a proper Bermuda insurance structure is the way to go, the same path catastrophe bonds use, with real custody and investor protection around anything that trades.

And trust from reinsurance actor requires co-designing the triggers together, especially for cyber, which is a genuinely new risk to put into a parametric, automatic-payout form. Once a respected partner has helped shape a trigger and then watches the money actually move mid-contract, and it works, the rest becomes a distribution game.

Can you share your experience building on-chain infrastructure for reinsurance capital?

Our tech team has been building these on-chain vaults — basically smart-contract “safes” that hold the money — for about four years, through the whole rollercoaster of this space. And the big lesson is almost the opposite of what people expect: putting assets on a blockchain isn’t the hard part anymore, and it’s not where you win.

The tooling got so good that spinning up the money-holding side is honestly kind of easy now. The moat isn’t the tech — it’s the market and the distribution: building the place where these things trade, agreeing on the price, and getting real risk and real capital onto the platform.

Any parting words?

Just this: in reinsurance, the money was never the problem — the fact that it’s frozen is the problem. Unfreeze it, and you don’t just make an old market work a bit better; you make it possible to cover the stuff nobody can insure today — data centers, cyber, the things the whole AI economy runs on. That’s where the real work lies, and I think it’s one of the more interesting problems in finance right now.