When Software Starts Spending on Its Own: A Pantera Partner Explains How AI Agents Are Reshaping SaaS Business Models and Crypto Payments

Pantera Capital managing partner Paul Veradittakit argues that AI agents are poised to become the primary buyers of enterprise software, upending a SaaS model built for decades around human purchasers. In his thesis, autonomous agents completing tasks—such as buying financial data, renting GPU compute, or calling third-party data-cleaning APIs—require fast, low-latency machine interfaces rather than human-facing UIs, rendering traditional sales funnels and credit-card subscription payments obsolete. He predicts pricing will shift from flat monthly per-seat subscriptions to granular, usage-based micropayments billed per API call, compute load, or business outcome, with crypto-native settlement and decentralized identity/verifiable on-chain wallets serving as the core payment rails for machine commerce. He cites the Fission protocol as an example, where agents on a decentralized network autonomously post needs, deliver services, and settle payments instantly around the clock without banking hours or FX friction. Veradittakit advises startups to rebuild pricing around on-demand billing, make products machine-readable and verifiable for automated trust systems, and track cryptographic payment infrastructure. Watch for adoption of agent-native payment protocols, tokenization momentum among traditional financial institutions, and regulatory clarity across major jurisdictions as signals of how quickly machine-to-machine commerce scales.