Prediction markets, zero-knowledge and media… a16z’s three biggest crypto themes for this year

Venture capital (VC) firm Andreessen Horowitz (a16z) identified three key keywords that will define this year’s digital-asset (crypto) market: ▲ smarter prediction markets ▲ broader adoption of zero-knowledge (ZK) proofs ▲ the rise of staked media.

On the 9th (local time), a16z Crypto released its “2026 Industry Outlook,” compiling analysis from its researchers and partners, and made the announcement.

Andy Hall, a professor of political economy at Stanford University and a16z research advisor, said he expects “prediction markets to get bigger and smarter.” He added that “by 2026, not only major events like elections but also complex, granular social phenomena will be trackable through real-time odds.”

A particularly notable point is the combination with AI. Hall said “AI agents will scour global information to gain a short-term trading edge and help improve the accuracy of forecasts for political and social events.” He also predicted that, to resolve disputes that arose when centralized platforms monopolized outcome resolution, decentralized governance and large language model (LLM) oracles will serve as arbiters to determine the “truth.”

Justin Thaler, an associate professor of computer science at Georgetown University, said “zero-knowledge proof (SNARKs) technology will transform the cloud computing market beyond blockchains.” Until now, ZK technology had been impractical outside blockchains because computation verification required massive resources.

Thaler explained that “this year, the efficiency of zkVM (zero-knowledge virtual machine) proving will improve dramatically, reducing overhead by a factor of 10,000 versus existing levels,” adding that “it will be usable even on smartphones.”

He said, “This makes it possible for a single GPU (graphics processing unit) to verify a CPU’s work in real time,” and predicted that “an era of ‘verifiable cloud computing’ will open, allowing low-cost proofs that work performed in the cloud was executed correctly.”

Robert Hackett, an editor at a16z, stressed that amid a flood of limitless content and misinformation produced by generative AI, “staked media” will emerge as a new alternative.

Staked media refers to a format in which a speaker stakes digital assets against a claim or forecast to back credibility. Hackett said, “We’re past the point where simply claiming to be ‘objective’ is enough,” and added that “a model that demonstrates truthfulness by taking on financial risk for one’s statements will draw attention.”

For example, a podcaster could lock up tokens to prove there is no intent to engage in a “pump and dump” (price manipulation), or an analyst could link forecasts to on-chain markets and publish a transparent track record.