I’m a structural engineer with 25+ years of experience in construction (157 built structures, CC-1 to CC-3), currently researching the Physical Oracle Problem in RWA tokenization.
Centrifuge has built excellent infrastructure for the financial layer. But for pools backed by physical assets — real estate, infrastructure, equipment — there’s a gap that no oracle currently addresses: cryptographic proof that the physical asset actually exists and matches its on-chain representation.
Chainlink delivers data faithfully. But who verified the source?
MANTRA/OM showed what happens when $4.7B is built on unverified physical claims. Tangible/USDR showed it with UK real estate.
I’m working on a verification framework combining drone photogrammetry + hardware attestation (TEE) for construction assets. Happy to discuss how this layer could complement Centrifuge’s infrastructure for physical asset pools.
New Silver is one of the most professionally run physical asset pools in DeFi — this analysis is not about their performance, but about a structural gap that exists in the entire category.
To make this more concrete, I looked at the New Silver pool documentation — one of Centrifuge’s longest-running physical asset pools.
Here is what the documentation says, and what it does not say:
What New Silver discloses:
Asset type: short-term bridge loans for fix-and-flip residential construction (12-24 months)
Average loan size: $190,000
Collateral: residential real estate under active renovation
Risk factors (their own words): “failure to recoup full value of the underlying asset in foreclosure” and “lack of knowledge in certain geographic areas”
What is missing — from a structural engineer’s perspective:
No draw inspection mechanism. In standard construction lending, funds are released in draws after an independent inspector confirms each stage of work (foundation, framing, rough-in, finish). New Silver’s documentation mentions no such mechanism. Who verifies that renovation is 30%, 60%, 90% complete before each disbursement?
ARV cannot be verified on-chain. After Repair Value is set by an appraiser at origination. But if the contractor cuts corners — cheap materials, poor waterproofing, structural shortcuts — the ARV is never reached. A property appraised at $280,000 ARV may sell for $190,000 after a low-quality renovation. There is no mechanism to flag this gap between projected and actual quality.
A fix-and-flip property mid-renovation has 30-50% of its ARV. If a borrower defaults at month 6 of a 12-month loan — the physical asset may have no roof, no windows, exposed framing. Foreclosure in many US states takes 300-500+ days. During this period, the token continues trading at face value. No on-chain signal reflects the deteriorating physical condition.
New Silver’s own disclosure: “lack of knowledge in certain geographic areas.” This is a direct admission that they lack local technical expertise across all 39 states they operate in. This is exactly the gap a decentralized network of local certified inspectors could fill.
I am not criticizing New Silver — they have an excellent track record. The gap is structural and exists across all physical asset pools, not just theirs.
The question for Centrifuge: as the protocol scales to institutional partners like Apollo and Janus Henderson, how will physical asset pool integrity be verified on an ongoing basis?
Happy to share how InfraVeritas approaches this technically.
Thank you. The problem is structural and affects every physical asset pool. Happy to discuss how IPAS addresses it specifically for Centrifuge infrastructure.
Hi Alex, good to hear from someone working on the same problem. What angle are you approaching it from — technical infrastructure, compliance, or data collection? Happy to connect and compare notes. — Petro
Hi Petro, great to see someone with real structural engineering depth working on this, the Physical Oracle Problem is exactly what I’ve been researching for the past year or two.
My angle is primarily the hardware/data capture layer. I come from aeral imaging and drone payload engineering - building the sensors and systems that actually observe physical assets. I’ve been designing an architecture that combines multi-modal sensing (spectral, photogrammetry, IoT) with cryptographic attestation at the point of data creation, then feeds that into a decentralized verification network before anything touches the chain.
The way I see it, there are three distinct unsolved layers: (1) tamper-evident data capture at the source, (2) decentralized consensus on physical state, and (3) on-chain enforcement with dispute resolution. Chainlink and Chronicle solve the plumbing for financial data feeds, but nobody has built the equivalent for physical reality. Your TEE + photogrammetry approach sounds like it’s attacking layer 1–2 from the construction side, which is exactly where I think the breakthrough needs to happen.
Would be keen to compare notes, particularly on how you’re handling the attestation chain from sensor to smart contract, and what construction asset classes you’re targeting first. I’ve done detailed work on the architecture and threat model side that might complement your domain expertise.
Your three-layer breakdown is solid and aligns with what I’ve been working on from the construction side.
Before we compare architectures in detail, I’d find it useful to understand your project context better — are you building this as an independent system, within a team, or as part of a funded initiative? What’s the current stage?
Happy to connect on LinkedIn and continue from there.