What is a tokenized bond actually for, once it has been minted?
For most of the assets now sitting on the XRP Ledger, the honest answer is: not much yet. They have been issued. They are recorded. They settle in three to five seconds for a fraction of a cent. And then, in the overwhelming majority of cases, they sit there. That gap, between issuance and use, is the gap Flare is now trying to close with the first live piece of its Flare 2.0 architecture.
Flare announced that Flare Confidential Compute (FCC) is set for deployment on Songbird, its canary network, pending a community governance vote that runs from 6 to 13 July. The architecture was first sketched under the Flare 2.0 banner in March 2025. With this proposal, it stops being a roadmap slide and becomes code on a chain with real economic value attached.
The premise is straightforward to state and hard to engineer. Flare wants to extend its consensus beyond its own blockchain by combining Trusted Execution Environments, hardware enclaves that isolate computation from the machine running it, with its existing data protocols. Inside those enclaves, private keys are generated and never leave. That lets a protocol on Flare hold and operate an account on an external network such as XRPL, while everyone else can still verify, on Flare, that whatever the enclave did was authorised and happened as claimed.
Why the timing is not an accident
The case for FCC rests almost entirely on what is happening one chain over.
The XRP Ledger has become one of the fastest-growing venues for real-world asset issuance. According to Evernorth Research, tokenized RWA value on XRPL reached about $4.18bn, a roughly 28-fold increase on the figure of around $147m a year earlier. RWA.xyz data put the network past $3bn in late April, a 59% jump in a single month, with Ripple’s own executives suggesting the true figure was closer to $3.75bn. Fast settlement, sub-penny fees and native compliance tooling, the ability to freeze, restrict and claw back tokens without bespoke code, have made it attractive to institutions that want on-chain plumbing without on-chain chaos.
There is a catch, and it is the whole reason a compute layer has a market. Most of that value is inert.
RWA.xyz splits tokenized assets into two buckets. “Distributed” assets are built to move between wallets. “Represented” assets are recorded on-chain but are not freely transferable outside the issuer’s participant set; the chain functions as a shared ledger for reconciliation rather than a market. On XRPL, the growth is overwhelmingly in the represented category. At the point the network crossed roughly $2.3bn, about $1.49bn of it was represented, the ledger reported just 22 RWA holders, and 30-day transfer volume had fallen about 91% to around $10m. That is the signature of a few large issuances held under tight control, not a liquid, actively traded market.
This is the wedge. XRPL issues and moves assets. It is not, by design, a smart-contract execution environment. So the question of what those assets can do, lend, collateralise, route through a vault, has to be answered somewhere else. Flare’s answer is: here.
What actually ships
Strip away the architecture diagram and three concrete things go live on Songbird if the vote passes.
First, a set of smart contracts that govern the system: registering TEE machines, accepting user instructions, and verifying results. Participating Songbird data providers relay instructions, package the required on-chain and off-chain data, and collectively authorise execution through weighted consensus. The TEE machines themselves, operated initially by the Flare Foundation on Google Confidential Compute, run approved code inside confidential virtual machines and return signed results.
Second, Protocol Managed Wallets (PMW). These let a protocol on Flare create and operate a wallet on an external chain according to rules written in code, rather than handing custody to a person or a centralised service. At launch, PMW supports the XRP Ledger, with keys generated and held inside the TEEs. This is the mechanism that, in principle, lets an XRP-denominated asset be put to work in a Flare application without the user manually shuttling funds across a bridge.
Third, Flare Data Connector V2 (FDC V2). The existing connector verifies external events in batches; V2 moves to faster, individually processed verification and serves as the proof layer underneath PMW, so an application can independently confirm that an action on XRPL actually occurred. FDC V2 is also a general upgrade available to anything building on Flare, not only to the wallet system.
Future releases are meant to open the same relay, verification and consensus machinery to outside developers, who would deploy their own TEE extensions. For now, the surface area is deliberately narrow.
The chain doing the shipping
Flare is deploying this from a position that is credible on engineering and humbling on markets.
The network has compounded liquidity off a genuinely low base. Total value locked climbed from under $10m in mid-2024, around the FIP.09 infrastructure push, past $120m within two weeks of the USD₮0 stablecoin launch in 2025, to somewhere in the $160m to $200m range today depending on the source and date. More than 90m FXRP has been minted, with roughly 80% of it deployed into DeFi protocols, and the network reports over 880,000 active addresses. Flare has, in effect, already built the largest EVM DeFi venue for XRP. FCC is the attempt to widen that franchise from a single wrapped asset into a general execution layer.
Songbird is the right place to find out whether it holds. It is not a testnet. It is a live, EVM-compatible chain with real value, real governance and the same enshrined protocols as Flare. When a potential vulnerability in FAsset bridging surfaced in February 2026, the fix was deployed and verified on Songbird before mainnet resumed, inside 24 hours, with no funds lost. That is the canary network doing exactly what it exists to do, and it is why a confidential-compute system with custody implications should be proven there first.
The wave Flare wants to ride
Zoom out and FCC is a bet on a category, not just on XRP.
Confidential computing, the practice of protecting data while it is being processed rather than only at rest or in transit, has become one of the faster-growing corners of enterprise security. One widely cited forecast from Grand View Research puts the market at about $18.1bn in 2026, rising to $153.8bn by 2030, a compound annual growth rate above 60%. The demand drivers are real: sovereign-cloud mandates, AI workloads touching sensitive data, and a steady fall in the performance overhead of running encrypted computation. NVIDIA’s confidential computing on H100 GPUs, for instance, now runs private-dataset workloads at under 7% overhead.
For Flare, the relevance is that the same TEE primitives underpinning “confidential AI” in the enterprise are what make verifiable cross-chain custody plausible in DeFi. Riding an established hardware and cloud trend is more defensible than inventing one.
Where Flare actually sits
On the competitive map, XRPL is a high-momentum challenger rather than the incumbent, and Flare’s fortunes are tied to it.
XRPL ranks around fifth globally by total tokenized RWA value, well behind Ethereum, which still hosts more than half of all tokenized assets, but it has repeatedly placed near the front on 30-day growth. That combination, large enough to matter, growing fast enough to notice, is precisely the profile that makes an execution layer commercially interesting. If the represented-asset pile on XRPL ever converts into usable, transferable, composable positions, the protocol that sits at the conversion point captures the flow. FCC is Flare’s claim to that position.
Governance and the money question
None of this is settled until token holders vote. The notice period runs 29 June to 5 July, the vote 6 to 13 July, and the proposal can be read on Flare’s voting portal.
The economics during the bootstrap phase are explicitly temporary. Participating data providers are compensated from a Flare Foundation-funded participation pool plus attestation fees while the network gathers operational data, ahead of a permanent incentive model. That permanent model is already taking shape through FIP.16, the tokenomics overhaul that passed in April, which routes protocol revenues, including FCC charges, into an entity mandated to buy back and burn FLR. In other words, FCC is not only a feature. It is meant, eventually, to be a revenue line.
Honest risks and open questions
A clear-eyed reader should weigh at least six things.
1. Centralisation at launch. The TEEs are operated by the Flare Foundation on Google infrastructure. The architecture is designed to decentralise over time, but at genesis the security of any externally controlled wallet rests on a single operator and a single cloud provider. “Verifiable” is not the same as “trustless,” and the distinction matters most precisely where custody is involved.
2. The TEE trust model itself. Confidential computing moves trust into hardware enclaves. That is a real improvement over plaintext processing, but TEEs have a documented history of side-channel and microarchitectural attacks. A system that lets enclaves hold private keys controlling real assets inherits that threat surface. The Songbird-first, caps-on approach is the right instinct; it is not a guarantee.
3. The demand is not proven. The bull case leans on XRPL’s RWA growth, yet most of that value is represented assets held by roughly 22 wallets with collapsing transfer volume. FCC is a bet that programmability unlocks latent demand. It is equally possible that institutions tokenising for internal record-keeping never wanted on-chain composability in the first place.
4. The token backdrop is weak. FLR trades around $0.008, roughly 94% below its all-time high, with a market capitalisation near $700m. The FIP.16 buyback-and-burn machinery only works if usage, and therefore fees, actually materialise. Architecture does not move a token; throughput does.
5. The total addressable market is genuinely uncertain. The confidential computing “wave” is real, but its size is contested to an almost comic degree.
End-period forecasts from eight research firms span from about $9bn to more than $1.4tn, with implied growth rates from roughly 21% to 64%. Any single number quoted with confidence, including the one in Figure 3, should be read as one analyst’s guess inside a very wide cloud.
6. The vote is not a formality, and Songbird is not mainnet. Governance approval is required, and even if it passes, this is a canary deployment with caps, not a Flare mainnet launch. The distance between “live on Songbird” and “load-bearing on Flare” is exactly where past deployments have spent months.
The fair summary: Flare is shipping credible, carefully staged architecture into a real and growing demand environment, while carrying real centralisation, security and demand-proof risks, and a token market that has so far refused to reward the roadmap. The next signal worth watching is not the vote. It is whether, three months after launch, any represented asset on XRPL has actually started to move.