a16z Crypto explains AI trends for 2026 and highlights their impact on research, finance, and the future of the open web.

by
Alyssa Davidson

Posted: January 12, 2026 4:49 AM Updated: January 12, 2026 4:49 AM

by well

Edit and fact check date: January 12, 2026, 4:49 AM

We sometimes use automatic translation plugins to improve your local language experience. The automatic translation may not be accurate, so please read the original text for accurate information.

a16z crypto, the blockchain and digital asset fund of Silicon Valley venture capital firm Andreessen Horowitz, has released an overview of expected AI trends for 2026.

The developments highlighted suggest that 2026 could be a pivotal year for the integration of AI, digital agents and blockchain technologies. As AI gains complex reasoning and autonomous problem-solving capabilities, and digital agents require verifiable identities to participate in financial and online ecosystems, new systems will be needed to ensure fairness, transparency, and value flow.

From enabling new research workflows to securing agent-based transactions and maintaining an open web, these trends point to a future where technologies, identities, and incentives must align to support both innovation and a stable digital economy.

Scott Kominers, a16z cryptocurrency research team member and Harvard Business School professor, pointed out in the report that AI will increasingly be applied to practical research tasks.

He emphasized that by the end of 2025, AI models will be able to follow abstract instructions and deliver new, correctly executed outputs, a previously difficult development. According to the researchers, AI is now being used more widely in research, especially in reasoning-heavy areas where models support discovery and solve complex problems autonomously, such as the Putnam math test.

The potential impact of supporting this type of AI research is still emerging, but it could spur new forms of polymath research that emphasize forming hypotheses about relationships between ideas and quickly inferring tentative or partially correct answers. Although these results are not always accurate, they can provide useful guidance similar to leveraging model “hallucinations” for creative innovation. This approach requires a new AI workflow described as agent-wrapping-agent, where model layers evaluate previous outputs and incrementally improve the results.

Scott Kominers has reported using this method in writing academic papers, and other researchers are applying it to patent analysis, artistic creations, or identifying vulnerabilities in smart contracts. He added that implementing ensembles of layered inference agents requires improved model interoperability and mechanisms to fairly evaluate and reward each model’s contribution, a challenge that blockchain and cryptography technologies can potentially solve.

As digital agents expand, financial services will transition from KYC to KYA.

Sean Neville, co-founder of Circle, architect of USDC, and CEO of Catena Labs, explained his trend by noting that the focus in financial services is shifting from “Know Your Customer” (KYC) to “Know Your Agent” (KYA).

He noted that the bottleneck in the emerging agent economy is shifting from intelligence to identity verification. According to Sean Neville, non-human identities currently outnumber human employees by 96 to 1, but these digital agents mostly remain unbanked or unverified.

He emphasized that KYA is a critical infrastructure requirement. Similar to how humans rely on credit scores to access loans, digital agents require cryptographically signed credentials to conduct transactions, tying each agent to principal, operating constraints, and responsibilities. Without these mechanisms, merchants will continue to block agent activity on their access points. He also suggested that while the financial industry has developed KYC systems over decades, implementation of KYA may need to be achieved within a few months.

AI agents impose an ‘invisible tax’ on the open web

Meanwhile, Liz Harkavy of the a16z cryptocurrency investment team said the emergence of AI agents is creating an “invisible tax” on the open web and disrupting the economic foundations.

This problem is caused by a mismatch between the context and execution layers of the Internet, where AI agents extract information from ad-supported sites to provide user convenience, while bypassing revenue streams such as advertising and subscriptions that sustain content creation.

Liz Harkavy noted that new technical and economic solutions are needed to maintain the diverse content that supports the open web and AI development. This could include next-generation sponsored content, micro-contribution mechanisms, or other innovative funding models. She added that current AI licensing agreements are insufficient and often compensate content providers for only a portion of revenue lost due to AI-based traffic.

According to her, the web needs a techno-economic model where value flows automatically and transitions from fixed licenses to real-time usage-based rewards. This approach could include testing and scaling systems using blockchain-based nanopayments and advanced attribution standards to ensure that any entity that contributes information to the actions of an AI agent is automatically rewarded.

As MPost’s resident journalist, Alisa specializes in the broad areas of cryptocurrencies, zero-knowledge proofs, investing, and Web3. With a keen eye for new trends and technologies, she provides comprehensive coverage to inform and engage readers about the ever-evolving digital financial landscape.

• 
  
  

more articles