Tokenized RWAs surge 589% as AI agent autonomy questioned
A new survey from IC3 researchers challenges the crypto “AI utility” narrative, arguing that giving AI agents a crypto wallet enables automation but does not make models inherently more trustworthy or harder to shut down. The researchers also say blockchains are strong for timestamping artifacts, but weak for reliably verifying authorship at scale.
On the tech side, Xiaomi set an AI inference record: MiMo-V2.5-Pro-UltraSpeed reached over 1,000 tokens per second and peaked near 1,200 using an eight-GPU commodity setup. The company credited FP4 quantization on expert layers and speculative decoding. A limited API trial is scheduled for June 9–23.
Market momentum remains anchored in tokenized RWAs. Tokenized real-world assets grew 589% from early 2025 to June 2026, with dollar gains led by tokenized bonds and money market funds (+$6.5B). Tokenized stocks also rose strongly (+422%), while tokenized precious metals added $1.5B, supported earlier by safe-haven demand amid geopolitical uncertainty.
In parallel, major platforms pushed AI-agent trading and custody rails. ConsenSys launched a non-custodial wallet for AI agents, and Robinhood said customers will soon deploy AI agents to trade crypto on their behalf (initially via an equities beta).
Crypto trading infrastructure and institutional tokenization continued to expand: Kraken added tokenized SpaceX shares via xStocks (over $25B cumulative trading volume in ~8 months), Ondo Global Markets crossed $1B in total value locked, and banks increasingly explore tokenized deposit networks to compete with stablecoins.
Overall, tokenized RWAs growth looks durable, but IC3’s critique is a reminder that on-chain speed and tokenization alone do not solve the trust, oversight, and “real utility” questions around AI and autonomous agents.
Neutral
The news is mixed for traders. Tokenized RWAs data (up 589% YoY-style window) signals sustained demand for on-chain real-economy exposure, which can support risk appetite in RWA-related tokens and improve liquidity expectations. At the same time, the IC3 survey undermines a key narrative: “wallets make AI agents autonomous” is not treated as a trust or safety breakthrough. That can dampen enthusiasm for AI-agent-on-crypto themes, especially if markets had been pricing autonomy benefits.
Historically, crypto rallies driven by an “adoption/infrastructure” story (e.g., earlier waves around stablecoins and tokenized funds) often remain constructive, but can cool when claims about transformative capability outpace verifiable outcomes. Short-term, traders may react positively to the measurable RWA growth stats and big platform announcements (Kraken/xStocks, Ondo TVL, bank settlement experiments). Long-term, the broader impact likely hinges on whether tokenized RWAs deliver clearer settlement, compliance, and auditability benefits beyond token issuance—otherwise the AI-agent hype could face repeated skepticism.