Hey everyone,
Gean Chu is a partner at Honey Island Capital, a Curitiba-based venture firm founded by veterans of EBANX. I caught up with him for a deep dive on quantum computing, which has become one of the main factors weighing on crypto markets in 2026.
Quantum went from a distant theoretical worry to the story driving digital asset headlines this year, largely on the back of a March paper from Google Quantum AI that cut the estimated hardware requirement for breaking Bitcoin’s cryptography by a factor of twenty.
Gean and his team spent months researching the threat in order to make their own allocation decisions, and he walked me through what they found.
I highly recommend reading Honey Island’s full report if you want to learn more about this subject.
Another good resource is the Quantum Bitcoin Countdown podcast create by “Bitcoin Historian” (and former CoinDesk colleague of mine) Pete Rizzo.
We recorded this episode on August 25, and two notable things happened in the 48 hours after.
On August 26, StarkWare mined what it says is the first quantum-safe Bitcoin transaction on mainnet, in block 964,199. The method was designed by StarkWare’s Avihu Levy, who published it back in April, and it uses signature grinding to attach a hash-based lock alongside the normal elliptic curve signature. This did not involve a soft fork or protocol change; rather, it was handed straight to MARA’s Slipstream service because Bitcoin nodes will not relay the non-standard format. This is a real world example of the private mempool workaround Gean describes in the episode. Both Levy and the MARA Foundation say private mempools are not a durable answer, and Levy still wants a protocol-level upgrade.
The next day, Blockstream Research published a BIP for SHRINCS, one of the two signature schemes Gean walks through. His position in this episode is that the community should not lock in a scheme before it has been battle tested. That argument is now live rather than hypothetical.
🙌 You can listen to Bits and Borders on your favorite podcast platform YouTube | Spotify | Apple Podcasts
🔥 Join our English language Telegram group to continue the conversation
Sumsub recently released its fourth annual State of the Crypto Industry report. The new research reveals how digital asset companies are balancing fraud prevention, regulatory pressure, and user experience as they scale in 2026. This report draws on Sumsub’s internal data from 2024–2025 and insights from 300 crypto companies surveyed outside Sumsub’s customer base.
Some of the key global highlights include:
Crypto firms are moving away from ‘growth at all costs’, with 74% now prioritizing verification accuracy over user onboarding speed (39%).
Despite fraud rates remaining flat at 2.2% from 2024 to 2025, crypto firms operate in a structurally riskier environment where targeted, automated and AI-driven attacks are the new normal.
Over half (55%) of surveyed companies confirmed they experienced fraud at least once in 2025, with 15% unsure if it happened or not–emphasizing the lag between detection capabilities and growing fraud sophistication.
Roughly 6.9 million bitcoins have their public keys already visible on chain, worth somewhere near $550 billion at current prices. Gean says most of that sits with exchanges and custodians reusing the same addresses, which is an address hygiene problem they could fix today. The stubborn part is the roughly 2.4 million bitcoins in Satoshi-era wallets.
The March 2026 Google paper put the superconducting requirement at about 500,000 physical qubits to crack a key in nine minutes. Google’s Willow chip currently has 105. Gean argues the more consequential development is Oratomic, a Caltech neutral atom startup that launched this year, which he says could do the same job with 10,000 qubits and already has circuits running around 6,000.
The two competing quantum hardware designs create very different problems:
Superconducting machines are fast but enormous, needing roughly 500,000 physical qubits to crack a key in nine minutes, which is quick enough to steal a transaction while it sits waiting to be confirmed.
Neutral atom machines need far fewer qubits but run much slower. Gean puts a 10,000-qubit neutral atom system at about three years to open a single wallet, and a 26,000-qubit system at about three days.
Three days is a real danger to coins whose keys are already exposed on chain, and nowhere near fast enough to beat Bitcoin's ten-minute block time. That ordering is the reassuring part of his analysis. The version of this attack that would break Bitcoin as a payment network is the version furthest from being buildable, and the version that is closest can only reach coins that are already sitting exposed.
Post-quantum signatures exist and NIST (US National Institute of Standards and Technology) has been working on them since 2016, but they do not fit inside a Bitcoin block. Developers are addressing this by building variants tailored specifically to Bitcoin. Gean’s position is that pushing an immature scheme into consensus would inject more risk than it removes.
The Bitcoin community has been more engaged on this issue than its critics allow. Gean says quantum went from around 5% of traffic on the Bitcoin dev mailing list to roughly half of it, and the proposals on the table range from doing nothing at all to freezing exposed coins to a maritime salvage framework where whoever cracks a wallet becomes a custodian rather than an owner.
Regarding “Q-Day” timing, he defers to the institutions that have moved their own deadlines forward, with Google, Cloudflare and the US government all now targeting 2029 or 2030. Once a workable signature exists, he says 98% of bitcoins could migrate in under a week using a quarter of available block space.
I enjoyed this conversation with Gean and I hope you do as well. You can connect with him on LinkedIn.
Have a great weekend everybody.
-AWS
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.