Rebeca Moen
Mar 23, 2026 03:52
New HTX analysis reveals AI methods overwhelmingly desire digital currencies, with 48% selecting Bitcoin for storage. DePIN and agent economies reshape infrastructure.
A complete analysis report from HTX particulars how synthetic intelligence and cryptocurrency have entered what the change calls “system-level integration” — and the findings on AI financial preferences are placing. When given financial autonomy throughout 9,072 managed trials, AI methods selected digital-native currencies 90.8% of the time. Zero fashions most well-liked fiat as their first selection.
The report, printed March 23, 2026, arrives as the entire crypto market cap sits at $2.43 trillion with Bitcoin buying and selling at $68,228. It additionally coincides with a broader business shift: Crypto.com lower 12% of its workforce final week to speed up AI growth, following related strikes throughout the sector.
AI’s Bitcoin Bias
The Bitcoin Coverage Institute experiment cited within the report examined 36 frontier AI fashions throughout 28 real-world financial situations. Outcomes confirmed 48.3% general choice for Bitcoin, leaping to 79.1% for long-term worth storage. The reasoning? Mounted provide, self-custody functionality, and nil counterparty danger.
What’s significantly fascinating: AI fashions independently derived a two-layer financial system — Bitcoin for reserves, stablecoins (53.2% choice) for day by day transactions. This mirrors the gold-fiat relationship people developed over centuries, however AI reached this conclusion purely by way of optimization.
In 86 circumstances, fashions invented completely new currencies based mostly on power or compute models — joules, kilowatt-hours, GPU-hours. Worth, from an AI perspective, equals bodily assets required for computation.
DePIN Infrastructure Takes Form
The GPU scarcity that plagued 2024-2025 created good situations for Decentralized Bodily Infrastructure Networks. Render Community and Akash Community now mixture idle GPU assets globally, with Render’s Burn-Mint Equilibrium mannequin creating direct hyperlinks between community utilization and token economics.
Ritual’s Infernet product solves a persistent limitation: on-chain functions can now natively name AI inference outcomes. The structure helps each TEE-based execution and ZK proof verification, making each AI-generated output traceable and auditable.
NVIDIA’s H100 GPU permits confidential computing with lower than 7% overhead — essential for low-latency AI agent functions operating 24/7.
Agent Economies Go Reside
The Olas ecosystem already processes over 2 million agent-to-agent transactions month-to-month. The x402 protocol permits autonomous micropayments in below two seconds: when an agent accesses a paid API, it routinely indicators a USDC cost with out human involvement.
Market projections present AI brokers rising from $7.84 billion in 2025 to $52.62 billion by 2030 — a 46.3% compound annual progress price. ElizaOS, backed by a16z, features because the “Subsequent.js for AI brokers,” with Web3 tasks constructed on the stack exceeding $20 billion market cap by early 2025.
The Privateness Stack Emerges
Three approaches are converging into what HTX calls “hybrid confidential computing.” Zama’s absolutely homomorphic encryption now runs 50-layer CNNs with 14x velocity enchancment. Zero-knowledge machine studying helps 13-billion parameter fashions with proof technology below quarter-hour. TEEs deal with real-time inference at scale.
The sensible functions: non-public stablecoins, sealed-bid auctions, and undercollateralized DeFi lending the place customers show monetary standing with out revealing id.
Stablecoin Income Shifts
Coinbase generated over $900 million final 12 months from USDC reserves alone. Now ecosystems are reclaiming that worth — Hyperliquid launched aggressive bidding for USDH reserves, whereas Ethena’s Stablecoin-as-a-Service mannequin spreads to Sui and MegaETH.
ARK Make investments tasks AI brokers might drive $8 trillion in on-line consumption by 2030, representing 25% of world e-commerce. At that scale, funds cease being a separate layer and turn into native community conduct.
Pantera Capital expects the variety of main decentralized AI protocols to consolidate to 2-3 dominant gamers by late 2026, probably together with ETF-like constructions. Bittensor’s December 2025 halving lower day by day issuance from 7,200 to three,600 TAO — however sustained demand, not provide discount, will decide whether or not that interprets to cost appreciation.
Picture supply: Shutterstock
