Stablecoin profits quantified; on-chain buybacks and AI monetization advance in parallel
Five newly disclosed facts as of October 17, 2026, 10:30 (UTC+8):
1. On October 16, Circle disclosed that of approximately $1.4 billion in revenue in the first half of 2026, $819 million went to distribution partners and transaction costs, while $473 million was spent on licensing, banking, liquidity, compliance, and products. Net profit ultimately came to approximately 7 cents per $1 of revenue. Stage: revenue recognition → profit realization; evidence: A2. Research implication: USDC scale cannot be equated directly with Circle’s profits; customer access and control over balance distribution are important points where value is captured.
2. From October 16, 10:30, to October 17, 10:30, the Hyperliquid Assistance Fund used 1,784,272.13 USDC to buy 22,277.65 HYPE tokens and permanently remove them from circulation. Stage: capital deployment → token value realization; evidence: A1. Research implication: Actual supply reduction is confirmed, but net supply changes still need to be assessed alongside token emissions, contributor releases, and unlocks.
3. On October 16, Cloudflare raised the per-account routing limit from 1,000 to 50,000 and priced additional Tunnels at $4.50 per month; enterprise customers can pay based on committed volumes or actual usage. Stage: purchasing mechanism → product delivery; evidence: A2. Research implication: Edge network demand from multi-tenant and AI applications is beginning to translate into billable usage, though actual purchases and revenue contributions have yet to be disclosed.
4. As of October 17, 10:30, BNC closed at $4.49, while BNB was approximately 720.99 USDT. Calculated using the report’s methodology, mechanical net NAV was approximately $6.847 per economic share, implying a discount of approximately 34.42%. During the period, there were no filings indicating new financing transactions, additional BNB purchases, fee waivers, or net node revenue. Stage: market pricing; capital commitments have not advanced; evidence: A2/B/C. Research implication: A widening discount does not in itself mean an increase in per-share value; the capital flywheel still requires evidence of actual financing and increased BNB per share.
5. U.S. Treasury TIC data released on October 16 showed net dollar portfolio inflows of $199.8 billion in August. Foreign investors made net purchases of $300.7 billion in long-term securities, including $109.0 billion in medium- and long-term U.S. Treasuries and $153.9 billion in U.S. equities. Stage: actual capital inflows; evidence: A1. Research implication: Long-term dollar assets continue to attract strong demand; capital flows cannot be inferred solely from the “de-dollarization” narrative.
Sources: Official materials from Circle, Hyperliquid, Cloudflare, the SEC, and the U.S. Treasury, as well as public market data. See today’s research report for the full links.
Five newly disclosed facts as of October 17, 2026, 10:30 (UTC+8):
1. On October 16, Circle disclosed that of approximately $1.4 billion in revenue in the first half of 2026, $819 million went to distribution partners and transaction costs, while $473 million was spent on licensing, banking, liquidity, compliance, and products. Net profit ultimately came to approximately 7 cents per $1 of revenue. Stage: revenue recognition → profit realization; evidence: A2. Research implication: USDC scale cannot be equated directly with Circle’s profits; customer access and control over balance distribution are important points where value is captured.
2. From October 16, 10:30, to October 17, 10:30, the Hyperliquid Assistance Fund used 1,784,272.13 USDC to buy 22,277.65 HYPE tokens and permanently remove them from circulation. Stage: capital deployment → token value realization; evidence: A1. Research implication: Actual supply reduction is confirmed, but net supply changes still need to be assessed alongside token emissions, contributor releases, and unlocks.
3. On October 16, Cloudflare raised the per-account routing limit from 1,000 to 50,000 and priced additional Tunnels at $4.50 per month; enterprise customers can pay based on committed volumes or actual usage. Stage: purchasing mechanism → product delivery; evidence: A2. Research implication: Edge network demand from multi-tenant and AI applications is beginning to translate into billable usage, though actual purchases and revenue contributions have yet to be disclosed.
4. As of October 17, 10:30, BNC closed at $4.49, while BNB was approximately 720.99 USDT. Calculated using the report’s methodology, mechanical net NAV was approximately $6.847 per economic share, implying a discount of approximately 34.42%. During the period, there were no filings indicating new financing transactions, additional BNB purchases, fee waivers, or net node revenue. Stage: market pricing; capital commitments have not advanced; evidence: A2/B/C. Research implication: A widening discount does not in itself mean an increase in per-share value; the capital flywheel still requires evidence of actual financing and increased BNB per share.
5. U.S. Treasury TIC data released on October 16 showed net dollar portfolio inflows of $199.8 billion in August. Foreign investors made net purchases of $300.7 billion in long-term securities, including $109.0 billion in medium- and long-term U.S. Treasuries and $153.9 billion in U.S. equities. Stage: actual capital inflows; evidence: A1. Research implication: Long-term dollar assets continue to attract strong demand; capital flows cannot be inferred solely from the “de-dollarization” narrative.
Sources: Official materials from Circle, Hyperliquid, Cloudflare, the SEC, and the U.S. Treasury, as well as public market data. See today’s research report for the full links.