Verification demand and security risks both rise as AI improves, pushing blockchains toward provenance while expanding attack and surveillance pressure
Software stocks have slid this year as investors reassess which business models hold up as AI tools spread across the economy. Crypto has largely traded in the same risk-off wave, falling alongside the software drawdown.
Grayscale says that correlation is masking a longer-term link. Its head of research, Zach Pandl, argues AI and blockchains are complementary, not competing bets.
The recent selloff has hit companies that sell software and data products, with investors focusing on whether AI agents can replace parts of the application layer and compress pricing power.
The drawdown has been large enough to pull in broader concerns about credit exposure and corporate funding costs, not just equity multiples. That wider spillover has reinforced the risk-off tone in tech.
Grayscale’s thesis starts with a simple premise: AI becomes more economically useful once it can pay and get paid, not just generate outputs.
Pandl argues that if AI agents are equipped with wallets, blockchains offer a direct path to settle transactions globally, around the clock, without relying on bank rails designed around human identity checks and operating hours.
In that framing, blockchains are not a rival to AI. They are an always-on settlement layer that AI systems can plug into when automated commerce moves beyond pilot projects.
Pandl points to stablecoins as the likely first place this shows up. The thesis is not that AI agents will suddenly move large sums. It is that they will move lots of small payments frequently as they pay for services, data, compute, or fulfilment.
A pickup in low-value stablecoin transfers would be an early indicator that automated, wallet-based commerce is moving from demos into production.
Grayscale also argues that AI’s trust problems create demand for better verification systems. As deepfakes improve and synthetic content becomes harder to distinguish, proving origin and timestamping records becomes more valuable.
Public blockchains can provide tamper-resistant logs that support verification workflows, especially when paired with tools that tie credentials, signatures, or attestations to an address.
The pitch is not that blockchains “solve” misinformation. It is that they can provide durable records that reduce reliance on centralised gatekeepers to decide what is authentic.
Grayscale’s argument comes with a warning. Better AI can make it easier to monitor transactions at scale, which can weaken privacy for users who rely on pseudonymous behaviour.
Security risk is also rising. AI agents are getting better at finding software flaws, including in smart contracts. OpenAI’s EVMbench, built with Paradigm, is one recent sign that the industry is trying to measure how capable AI agents already are at detecting, patching, and exploiting high-severity vulnerabilities.
Markets can sell software and crypto together for months. The harder question is whether usage trends start to separate them. If AI agents really do become economic actors, the story will show up in transaction data and payment flows, not in a correlation line on a trading desk.
