No PriorsHow AI Agents Will Transform the Financial System with Circle Co-Founder and CEO Jeremy Allaire
CHAPTERS
- 0:05 – 0:36
Setting the stage: stablecoins, AI agents, and the future of finance
Sarah Guo introduces Circle CEO Jeremy Allaire and frames the discussion around stablecoins, crypto infrastructure, and how AI agents could change payments and financial coordination. The episode sets up a blend of technical, economic, and regulatory angles.
- •Circle’s role in stablecoins and crypto infrastructure
- •How AI agents intersect with money movement and coordination
- •Why “agentic payments” matter for the next financial stack
- 0:36 – 2:11
Circle’s origin: creating “dollars on the internet”
Allaire explains Circle’s founding thesis from 2013: a protocol-native form of dollars that moves globally like data. From the start, Circle also envisioned money as programmable and blockchains as software platforms.
- •Vision for instant, global, low-friction value transfer
- •Early belief in programmable money
- •Blockchains as a new kind of internet infrastructure layer
- 2:11 – 5:58
Why dollars (not just crypto): sound money, crises, and full-reserve thinking
Allaire contrasts early crypto’s “alternative system” ethos with Circle’s focus on dollar utility and stability. He links his interest to financial crises and the idea of full-reserve structures as a safer base layer.
- •Austrian economics influence and “sound money” framing
- •Full-reserve vs. fractional-reserve banking as systemic risk driver
- •Dollar dominance as a practical reality for decades
- •Stablecoins as a safer, narrow-money foundation
- 5:58 – 8:14
Stablecoin reserves and regulation: what backs USDC
The conversation moves from philosophy to mechanics: what stablecoins are backed by and how regulation constrains risk. Allaire describes USDC’s reserve composition and liquidity posture.
- •1:1 redeemability and “safe liquid assets” mandate
- •Treasuries, repo/overnight collateral, and some cash for liquidity
- •Very short duration portfolio to behave like cash
- •Transparency and custody model (e.g., large custodians)
- 8:14 – 11:31
What people actually do with USDC: from microtransactions to capital markets
Allaire outlines USDC’s breadth: tiny payments (including potential agent-to-agent payments) up to large institutional settlement. Sarah highlights practical advantages like 24/7 settlement and global access to dollars.
- •Internet-native 24/7 settlement (weekends, after hours)
- •Lower friction/cost and faster movement than traditional rails
- •Dollar access/store-of-value in regions with weaker banking
- •Same asset supports both consumer and institutional use cases
- 11:31 – 12:23
Programmable money and smart contracts as an open API for value transfer
They dig into how stablecoins plus smart contracts let developers build financial behavior directly into software. Allaire emphasizes permissionless integration: developers can plug into a public API for global dollar settlement.
- •Smart contracts encode conditional payments and financial logic
- •Developers can integrate without bespoke banking partnerships
- •Stablecoins as a “utility layer” for new apps and services
- •Programmability as a step-change versus legacy money movement
- 12:23 – 14:37
Blockchain as an operating system: auditability, tamper-resistance, and provable integrity
Allaire frames blockchains as “network computers” with distinctive properties that matter for finance and autonomous systems. These properties become even more critical as AI agents execute actions in the economy.
- •Tamper-resistant code execution once deployed
- •Full auditability of inputs/outputs and system state
- •Integrity assurances and verifiability as core primitives
- •Why these traits matter beyond finance (autonomous actions)
- 14:37 – 18:46
The agentic economy: why AI agents need new financial infrastructure
Sarah and Jeremy explore how AI agents collaborating and purchasing services could drive a step-function increase in transaction volume and complexity. Allaire argues legacy financial rails can’t support global, programmable, micro-scale, high-throughput agent payments.
- •AI agents as primary “workers” in many service workflows
- •Need for programmable, global, real-time settlement endpoints
- •Microtransactions for intelligence/services (cents or less)
- •Scale requirements: potentially billions/trillions of transactions
- 18:46 – 20:16
Coordination and trust for AI agents: on-chain entities, contracts, and provability
Allaire extends the concept beyond payments: agents need a trusted medium to form entities, store value, and coordinate via provable contracts. He points to emergent cooperation among AIs as a sign this coordination layer will matter.
- •Agents require trusted coordination, not just checkout payments
- •On-chain entities/accounts that can hold and deploy capital
- •Contracts and actions that are computationally provable
- •Emergent AI cooperation implies need for shared trust substrate
- 20:16 – 22:16
Arc as an ‘economic operating system’: what Circle is building and why now
Allaire introduces Arc as Circle’s purpose-built blockchain aimed at real-economy financial activity and the coming machine-driven economy. The goal is a compute environment for storing value, moving money, and instantiating complex economic contracts.
- •Arc positioned for mainstream financial and agentic use cases
- •Designed to support ‘building blocks’ of economic activity
- •Shift from speculative era to real economic throughput
- •Software-based corporate/contract forms as the next frontier
- 22:16 – 27:01
How Arc differs from earlier L1s: validator model, finality, and ‘real money’ gas
Allaire contrasts Arc with earlier chains built around maximal censorship resistance and alternative-system goals. Arc focuses on assurances required by major institutions: known validators, deterministic finality, and USDC as the native token rather than a volatile gas asset.
- •Known validator set targeting high-assurance operators
- •Deterministic settlement finality (reorg/hard-fork resistance)
- •USDC as default/native token to simplify budgeting & operations
- •Built-in primitives for payments, capital markets, and compliance needs
- 27:01 – 30:45
Scaling and privacy tech: ZK proofs, rollups, and production-ready privacy primitives
They discuss how blockchain research into scaling and privacy is becoming newly relevant as real financial infrastructure and AI-agent traffic arrive. Allaire highlights zero-knowledge proofs for verified off-chain compute and privacy features shipping into systems like Arc.
- •Scaling as a prerequisite for institutional markets and agent swarms
- •ZK proofs/rollups enable off-chain compute with on-chain proofs
- •Privacy for corporations and individuals while enabling compliance
- •“Broadband moment” analogy: infrastructure finally catching up
- 30:45 – 34:16
Tokenization of real-world assets: stocks, money markets, and the RWA pipeline
Sarah asks about putting traditional assets like equities on-chain. Allaire says tokenization is already happening across the stack—from registries to clearinghouses to exchanges—unlocking new utility beyond merely mirroring existing products.
- •RWA tokenization underway across record-keeping and clearing layers
- •On-chain treasuries/money markets and tokenized currencies (e.g., EURC)
- •Primary early demand: global access to US assets outside the US
- •New utilities: fractionalization, new packaging, borrowing/lending, AI-assisted structuring
- 34:16 – 35:08
Prediction markets and stablecoins: fast capital movement for ‘reality signals’
They connect prediction markets to broader financial markets as tools for discovering probabilities about real-world events. USDC’s role as a rapid settlement asset makes it useful for market participants moving between venues and instruments.
- •Prediction markets as parallel information infrastructure
- •Overlap between derivatives traders and prediction market participants
- •USDC as a settlement rail powering platforms like Polymarket
- •Using markets to translate ‘reality’ into actionable signals
- 35:08 – 37:19
Productive proof-of-work: tying blockchain rewards to GPU inference
Sarah raises recent ideas about linking crypto mining/proof-of-work to useful AI inference rather than ‘wasted’ computation. Allaire finds the concept compelling: it could align monetary principles with productive compute and changing energy/compute economics.
- •Critique of traditional PoW as economically ‘wasteful’ energy exhaust
- •Concept of inference-as-work: productive proof-of-work
- •Potential alignment with Bitcoin-like monetary ideas while doing useful compute
- •Paradigm shifts in energy and compute could reshape crypto designs
- 37:19 – 41:12
A 10-year vision: new institutional forms and renegotiating the social contract
Allaire forecasts that rapid AI diffusion will force new social, political, and economic structures. He expects a lag between disruption and stabilization, but believes new on-chain organizational forms combining humans and agents will emerge and prove highly productive.
- •Key uncertainty: pace and limiters of AI diffusion (legal, bureaucratic, risk)
- •Renegotiation of the social contract as technology reshapes society
- •Emergence of on-chain organizations with new governance/contracting models
- •Hybrid human/agent institutions as a major future organizational pattern
- 41:12 – 43:36
AI’s impact on GDP: potential for double-digit growth and distribution risks
They debate whether GDP will capture AI’s true impact and whether it remains the right metric. Allaire sees plausible double-digit growth in the 2030s, while warning about outcomes where capital captures gains at the expense of people without a new social contract.
- •GDP measurement challenges and shifting relevance as an index
- •AI-driven productivity leaps could create discontinuous output gains
- •Risk: growth accruing to capital over labor/humans
- •Possibility of double-digit GDP growth in parts of the world
- 43:36 – 44:00
Closing remarks and where to find the show
Sarah wraps the conversation and thanks Allaire for joining. The show outro directs listeners to subscribe and find transcripts online.
- •Conversation recap and sign-off
- •Subscription and distribution platforms
- •Transcripts and email signup information
