Succinct

Today, we’re introducing data confidentiality to OP Succinct. Institutional adoption is accelerating, but data exposure is a non-starter. Our latest OP Succinct upgrade lets institutions run chains while keeping customer data confidential. When combined with access control configurations over the chain itself, this enables: ↳ Customer data confidentiality ↳ Self-hosted infrastructure ↳ Global liquidity ↳ Ethereum-anchored security We’ve partnered with Polygon Labs to power their CDK privacy configuration with OP Succinct, letting institutions keep transaction data confidential within their own environment.

Coinbase is partnering with Succinct to prove $7.4 billion in deposits on Base, using zero-knowledge proofs. Succinct's SP1 has made proving transactions fast and cheap, enabling every major rollup to begin migrating from optimistic proving to zero-knowledge proving. SP1 has been battle-tested in production environments, securing billions in value for over 35 customers across rollups, bridges, and DA layers. With Base adopting ZK proofs, SP1 will secure nearly $10B in digital assets across major rollups like Optimism Foundation, Mantle, Katana, and more. The Ethereum rollup ecosystem is consolidating around ZK — and SP1 is at the center.

Today, we're launching ZCAM, an iPhone camera app to Prove What’s Real. ZCAM cryptographically signs photos and videos at the moment of capture. Anyone can independently verify the content came from a real device and hasn't been altered or AI-generated. AI-generated content is indistinguishable from reality. AI detectors can't keep up. We benchmarked 7 tools, and simple edits dropped detection rates to as low as 4%. So we asked ourselves, "Instead of detecting fakes, can we prove what's real?" We built ZCAM, a practical way to cryptographically fingerprint media using existing hardware in modern devices. Authenticity shouldn't depend on trusting the platform or the creator. All content on the internet needs to carry its verified chain of custody with it. Download ZCAM on the iOS App Store to capture authentic photos. Developers can build with the ZCAM SDK to integrate provably real proofs into their own app. https://zcam.succinct.xyz/

The quantum threat to crypto was a 2040s problem. A Google Quantum AI paper just cut that timeline by a decade. Here are 4 takeaways the crypto industry needs to reassess: https://lnkd.in/eeXwsJrh

Google just published a landmark paper on quantum vulnerabilities in crypto. Their findings are so sensitive that Google chose to publish a zero-knowledge proof instead of the actual circuits. That ZK proof was built on Succinct's SP1 zkVM. This is one of the most consequential applications of ZK proofs ever. Responsible disclosure of a novel vulnerability protecting trillions of dollars in digital assets. The paper makes the quantum threat concrete. Every blockchain needs a migration plan for post-quantum security. Succinct is working with the Ethereum Foundation to power post-quantum signature aggregation with SP1. Ethereum's PQ roadmap requires a zkVM at the consensus layer, and we're building it. If your chain doesn't have a post-quantum plan, get in touch: https://lnkd.in/g9xeiZDv

The New York Times just tested AI image detectors. They found major problems. We found worse. But we also have a solution. Detection is a losing game. Defenders must anticipate every attack. Attackers only need one. The solution? Don't detect what's fake. Prove what's real. AI has broken our ability to trust what we see. In a few days, Succinct will unveil the first complete platform that restores authenticity to the Internet. Prove What's Real. March 10. Check out our findings: https://lnkd.in/eZR8n6Vs

AI-generated images are fooling everyone. We built the largest benchmark to test whether detectors can keep up. Spoiler: They can't. We built AdversIm, a benchmark of 15,000+ images generated by 5 state-of-the-art AI models, and tested them against 7 AI image detectors. Phase 1: Can detectors catch unmodified AI images? The best detectors caught 90%+ of fakes. Others were barely better than a coin flip. If this were the whole story, the outlook would be cautiously optimistic. But it's not. Phase 2: What happens after basic edits? We applied the kind of edits that happen every time you screenshot or reshare an image — blur, noise, compression. Edits barely visible to the human eye. The top 3 detectors went from detection rates of 98%, 90%, 90% to 36%, 11%, 13%. AI image detection is an arms race that can't be won. Defenders must anticipate every attack across every model and every edit. Attackers only need one that works. The only way forward is to Prove What's Real. Full paper + dataset: https://lnkd.in/enGG2pTZ

SP1 Hypercube is live on mainnet! In May, Succinct became the first team to prove Ethereum in real time. Six months later, we did it with just 16 GPUs, making SP1 Hypercube the most advanced, secure, and trusted zkVM in production with over $3B in assets secured across 35+ production partners. The Ethereum Foundation set ambitious criteria for bringing zero-knowledge proofs to L1 Ethereum. SP1 Hypercube delivers on the toughest benchmarks. We're working directly with the EF cryptography team to scale Ethereum to 1 gigagas per second, unlocking a new generation of real-world applications. Real-time proving just went from research milestone to production infrastructure. https://lnkd.in/eDXivh9z