QuantRWA

Inspiration

The art market alone is worth over $65 billion, yet a museum-quality painting sitting in a vault generates zero returns for its owner. We asked ourselves what would happen if you could buy a fractional share of a Rolex the same way you buy a fraction of a Bitcoin. The real hurdle was establishing trust. We had to figure out how to prove a photo of a watch uploaded to a website is authentic, unique, and actually worth the claimed amount. That is exactly the problem QuantRWA was built to solve.

What it does

QuantRWA lets users tokenize high-value physical assets like watches, art, or electronics simply by uploading a photo. Gemini Vision appraises the asset in real time to assign an estimated value and a fraud risk score. If the asset passes our security threshold, it gets minted as an SPL token on Solana and fractionalized into 10,000 tradable shares. An AI trading layer then manages a live portfolio of RWA tokens alongside native crypto like SOL and USDC. Every action is safely logged to a transparent MongoDB Atlas audit trail.

How we built it

We built QuantRWA on a four-layer stack.

• The AI layer uses Gemini Vision for multimodal asset appraisal and agentic trading logic.
• The security layer utilizes an AWS Lambda enclave backed by AWS Secrets Manager. This ensures our Solana vault key stays off the frontend, and every mint requires a cryptographic HMAC signature proving AWS approved it.
• The blockchain layer relies on Solana SPL tokens and Metaplex Core for high-frequency RWA trading.
• The frontend was built using Next.js with Tailwind CSS and Framer Motion, deployed on Vercel. MongoDB Atlas stores our time-series tick data and audit logs.

Challenges we ran into

The hardest part was the Oracle Problem. We had to figure out how to trust that a photo represents a real, unique, un-tokenized asset. We solved this using a two-layer verification system. Gemini Vision handles the AI appraisal and deepfake detection, while our AWS Lambda enclave acts as a final security firewall. It rejects any asset with a risk score above 30 before the private key is ever touched. Getting these two systems to communicate reliably with proper CORS configuration and IAM permissions took quite a bit of iteration.

Accomplishments that we're proud of

We are incredibly proud that the core security architecture works flawlessly end to end. A photo goes in, Gemini appraises it, AWS signs off on it, and a mint record with a cryptographic signature comes out in just a few seconds. We are also thrilled with the AI trading dashboard. It makes the platform feel alive and clearly demonstrates the HFT use case to anyone watching the demo.

What we learned

We learned that preparing and verifying assets prior to minting presents the biggest challenge when building on-chain. Verifying real-world assets is a complex problem at the intersection of computer vision, cryptography, and trust, and we have only just scratched the surface. We also discovered the importance of being strategic about what to make real versus reliable in a demo environment. Additionally, we found that technical judges find a rock-solid security architecture much more impressive than just a flashy frontend.

What's next for QuantRWA

The immediate next step is physical verification. We plan to partner with certified appraisers who can co-sign the AWS signature to add a secondary layer of human trust to the AI appraisal. After that, we want to build a secondary marketplace where RWA shares can trade peer-to-peer. We also plan to integrate DeFi primitives so holders can use their tokenized assets as collateral for loans. Long term, we envision QuantRWA becoming the infrastructure layer for any physical asset that deserves liquidity.

Built With

• amazon-web-services
• gemini-api
• mongodb
• next.js
• node.js
• react
• solana
• typescript