TLDR:
- Horowitz says AI has compute but lacks a network layer that crypto can supply with identity and value structures.
- Crypto creates provenance for content, defending AI from deepfakes and trust collapse.
- AI needs a decentralized registry of truth; crypto provides means to verify origin and authenticity.
- According to a16z, crypto is the connective layer that AI lacks, enabling secure identity and value flows.
AI holds raw power. But it’s incomplete. In a recent tweet, a16z shared Ben Horowitz’s view: computing always rests on two pillars. Machines are there. But networks are missing.
In Horowitz’s framing, crypto becomes that missing layer. He believes it can provide AI with money, identity, provenance, and a decentralized registry of truth.
Now the discussion shifts. Cryptocurrency is no longer niche infrastructure. For some at a16z, it’s central to taming AI’s downstream risks.
Ben Horowitz: Computing has always needed two pillars, machines and networks. AI has the machines but not the network.
Crypto is the missing layer, giving AI money, identity, provenance against deepfakes, and a decentralized registry of truth.
Source: @bhorowitz at Columbia… pic.twitter.com/T8CNp6jCEI
— a16z (@a16z) September 23, 2025
Crypto as the Network Layer AI Lacks
Horowitz frames computing innovation as built on machines and networks. He sees AI as strong on machines, computing, models, and data. But weak on the network. That is where crypto enters. He states crypto offers identity and value structures that AI cannot self-generate.
He also describes a role for crypto in provenance. In an age of deepfakes and synthetic content, verifying origin is vital. Cryptocurrency can anchor content to real identity and timestamping. That way, AI’s output can carry a traceable lineage.
Horowitz warns that without this layer, AI would drift into ambiguity. Who claims authorship? Who ensures truth? He argues crypto is necessary to ground AI in real networks of trust.
This perspective recasts crypto’s mission. It shifts from simply payment rails or tokenization to a foundational web infrastructure for AI systems.
Crypto, Price Signals, and Truth in AI
Embedding crypto into AI systems also enables price signals. Horowitz notes that cryptocurrency is money. It can incentivize correct behavior or penalize misbehavior in AI networks. That gives algorithmic systems economic feedback loops.
He also ties this to the registry of truth. In decentralized networks, truth is not imposed by one actor. Rather, it emerges through consensus and cryptographic proofs. AI’s assertions can then be anchored to that registry. That model defends against manipulation.
In a world where AI can generate content at scale, provenance and price layers become defense lines. Cryptocurrency becomes the guardrail rather than an optional add-on.
Still, implementing this vision is nontrivial. It demands integration across identity systems, cryptographic proofs, consensus, and AI models. It demands design. How to scale these networks is open.
Yet Horowitz’s framing shifts perception. Cryptocurrency is no longer peripheral. In his view, it becomes the network layer AI urgently needs for trust, value, and identity.
