WuBlockchain

A Reuters investigation found that unlicensed Dubai-based crypto exchange Shelbit processed at least $4 billion since May 2024 through a network linked to Iranian online gambling, the country’s central bank and sanctioned entities, including wallets associated by Israel with the IRGC. Dubai regulator VARA fined Shelbit on July 24 for continuing unlicensed activities after a January 2025 cease-and-desist order and said the case involved potential AML and terrorism-financing violations.

SEC Signals It Will Issue Crypto Market Rules if the CLARITY Act Stalls, Morgan Stanley Launches ETH and Solana Spot ETPs, BNY Moves Fund Recordkeeping On-Chain, Strategy Reports an $8.22B Q2 Loss While Holding Nearly 844,000 BTC, and Digital Asset Treasury Firms Pivot to AI Data Centers. For the complete and weekly curated reports, subscribe to our Substack:

1. If CLARITY Act Stalls, SEC Will Issue Its Own Crypto Market Rules link
Paul Atkins, Chair of the US SEC, stated that if Congress fails to pass the CLARITY Act, the SEC stands ready to craft its own rules governing crypto market structure. However, legislation would offer greater durability and prevent the regulatory framework from shifting with changes in administration. The bill cleared the Senate Banking Committee in May by a 15–9 vote and has yet to advance to a full Senate vote. The SEC has placed rules for crypto asset issuance, custody and trading on its 2026 regulatory agenda.
2. Crypto Ownership in Canada Rises to 25%, Marking Mainstream Financial Adoption link
According to the latest survey report released by the Ontario Securities Commission (OSC), the number of cryptocurrency holders in Canada has more than doubled over the past few years. Currently, one quarter (25%) of Canadians own digital assets or crypto investment funds, up from 10% in 2023 and 13% in 2022. The survey also shows that cryptocurrency use is moving beyond pure speculation. Seventy-four percent of existing crypto owners report having actually used digital assets, while 89% of stablecoin holders have utilized their positions, with 20% employing them for international transfers. Additionally, investor awareness of due diligence has improved: 50% of crypto holders verify whether trading platforms are legally registered before investing, compared with 38% in prior surveys.
3. Morgan Stanley Investment Management Launches ETH and Solana Spot ETPs With Staking Plans link
Morgan Stanley Investment Management, the asset management arm of Morgan Stanley, has launched the Ethereum spot ETP MSSE and Solana spot ETP MSOL, listed on NYSE Arca, both with an expense ratio of 0.14%. The two products plan to stake portions of ETH and SOL, and Morgan Stanley Investment Management will not retain staking rewards. Together with the previously launched Bitcoin product MSBT, its crypto ETP lineup now covers BTC, ETH and SOL.
4. BNY Mellon to Launch Digital Transfer Agency Service, Recording Fund Holders on Blockchain link
BNY, with over $59 trillion in assets under custody and administration, will launch a digital transfer agency service that migrates fund transaction processing and shareholder recordkeeping to blockchain while retaining legacy transfer agency systems. BNY’s transfer agency business currently serves approximately $8.6 trillion in assets and 7.6 million accounts. Baillie Gifford will be the first to deploy the system for the UK’s first fully native regulated tokenized fund, and BlackRock plus BNY’s Dreyfus are expected to adopt it in upcoming funds.
5. Strategy Releases Q2 Earnings: 843,800 BTC Held, $8.22 Billion Loss From Fair Value Changes link
Strategy, formerly MicroStrategy, released its Q2 2026 earnings report. As of July 26, the company held approximately 843,775 Bitcoin, representing a 25% increase year-to-date, and raised around $17.06 billion cumulatively via its ATM offering program. Q2 revenue stood at $122.4 million, up 6.9% year-on-year. Driven by fair-value movements of Bitcoin holdings, the firm posted a net loss of $8.22 billion, compared with a net profit of $10.02 billion in the prior-year quarter. At the end of June, it held $1.71 billion in cash and cash equivalents and $736.1 million in short-term investments.
Strategy stated it has lifted its dollar reserves to $3.75 billion, enough to cover preferred dividends and debt interest for roughly 2.1 years. Year-to-date, the company raised about $218.4 million by selling Bitcoin to fund part of preferred dividends and repurchased STRC preferred stock with a notional value of $28.9 million for approximately $25 million. It also noted convertible debt fell from $8.21 billion to $6.71 billion at quarter-end.
On the Q2 earnings call, Strategy said future capital raises will no longer be fully deployed into Bitcoin; instead, it will dynamically allocate between BTC and dollar reserves based on market conditions to strengthen the stability of its Digital Credit framework. The company reiterated it may sell small amounts of Bitcoin when conditions are favorable to boost dollar reserves, cover preferred dividends and interest, and support share repurchases. It also clarified no current plans to pursue Bitcoin-backed loans, citing counterparty risk, margin risk and vulnerability to short selling. Management aims to gradually reduce convertible debt and maintain dollar reserves at a level covering roughly two to three years of dividends and interest.
6. Robinhood CEO: Firm on Track to Become First Trillion-Dollar Financial Enterprise link
Robinhood CEO Vlad Tenev said on the Q2 earnings call that the company aims to become the first financial firm to reach a $1 trillion valuation, and will continue expanding around Agentic Finance, Robinhood Chain, global asset tokenization and private markets. Robinhood CFO Shiv Verma stated the firm believes its business can grow tenfold over the next decade.
The company has rolled out perpetual futures overseas and plans to make the product available to more customers. On a U.S. launch timeline, it offered no specific date but emphasized that “the absence of a timeline should not be taken as a lack of active pursuit.” Tenev also disclosed that the team is actively building AI Agentic Trading capabilities and intends to open more AI tools to traders in suitable scenarios.
Robinhood’s Q2 2026 net revenue rose 32% year-on-year to $1.308 billion, a record high. Net income grew 48% to $573 million, with diluted EPS of $0.62. Transaction-based revenue climbed 44% to $776 million: prediction markets revenue surged more than tenfold to $156 million, options revenue rose 29%, equities revenue jumped 95%, while crypto revenue fell 38% to $100 million. Quarterly net deposits reached $21.7 billion, platform assets grew 32% year-on-year to $369 billion, and Robinhood Gold subscribers rose to 4.8 million. The company also lowered its 2026 guidance for adjusted operating expenses and share-based compensation to a range of $2.675 billion to $2.775 billion.
7. Among Top 50 Stablecoins by Market Cap, Only USDC, USDG and EURC Meet MiCA Requirements link
Patrick Hansen, Senior Director of EU Strategy and Policy at Circle, wrote that 21 issuers in the EU have launched roughly 35 regulated e-money tokens. Yet among the world’s top 50 stablecoins by market capitalization, only USDC, USDG and EURC comply with MiCA rules. He argued that subsequent MiCA reviews should boost the regime’s competitiveness and strengthen global regulatory coordination. The framework should support EU-based e-money tokens in expanding cross-border payments and tokenized trade, while establishing an acknowledgment mechanism for non-EU regulated stablecoins to bring more global stablecoin activity under MiCA oversight.
8. Multiple DAT Firms Announce Shift to AI Data Center Businesses in Recent Months link
Amid the prolonged crypto slump, at least a dozen Digital Asset Treasury (DAT) firms have pivoted toward AI data centers and similar businesses in recent months to win back investors, with limited results so far. K Wave Media’s stock has fallen roughly 71% since shifting to data center development in May; Lixte Biotechnology dropped around 33% after merging with a battery firm; AlphaTON Capital, rebranded Alpha Compute, also declined about 33%. Multiple law firms note the DAT hype has cooled markedly, and more companies are pursuing new lines including AI, data centers, aerospace and small modular nuclear reactors.
9. July Average Monthly Bitcoin Spot Volume May Hit Lowest Level Since November 2023 link
K33 Research stated that Bitcoin markets remained sluggish in July and are on track for the lowest average monthly spot trading volume since November 2023. The report noted BTC fell around 3% over the past week, oscillating within the $60,000–$66,000 range. CME Bitcoin futures open interest sits at multi-year lows, while perpetual open interest stands at approximately 300,000 BTC. Average daily spot turnover in July was roughly $2.2 billion, reflecting persistently weak market participation.
10. Crypto Market Sees Broad On-Chain Contraction Instead of Sector Rotation in H1 2026 link
Binance Research’s H1 2026 on-chain market report states that the crypto space saw broad-based on-chain contraction rather than sector rotation in the first half of the year. Global DeFi TVL fell by $43.4 billion, a 38% drop, while the combined market capitalization of six major L1s shrank by $246.5 billion, down 42%. Key highlights include: marginal Ethereum holders shifted from ETFs to corporate balance sheets; spot ETF holdings declined to 5.2 million ETH, whereas digital asset treasury firms increased their holdings to 7.7 million ETH. Activity on general-purpose L2s deteriorated sharply, with user operations falling roughly 77% from January to June. Solana network revenue dropped 64.5%. BNB Chain emerged as a leading hub for tokenized equities and stood as the only deflationary major L1 with an annualized burn rate of 5.05%. The industry recorded 207 security incidents in H1 with losses totaling $972 million. Driven by the World Cup and non-sports events, monthly notional volume in prediction markets surged 86% to $51.6 billion; Kalshi and Polymarket accounted for 92% of total June trading volume.
Fundraising
Ethereum Institutional closed its first ecosystem funding round led by BitMine, SharpLink and others. link
AI security firm V12 secured a $10 million seed round to build an automated vulnerability discovery system. link
Robotics data platform Axis Robotics completed a $12 million seed financing. link
Distributed AI infrastructure project ALPHEA announced $5 million in funding. link
Learn more, check out crypto-fundraising.info.
Follow us
Twitter: https://twitter.com/WuBlockchain
Telegram: https://t.me/wublockchainenglish

Glassnode said Bitcoin’s three-month futures basis yield has stayed below the U.S. two-year Treasury yield since February, marking only the second prolonged inversion on record and pushing institutions toward cash and government bonds. Bitcoin spot trading volume has fallen to its lowest level since 2019, while exchange flows and spot ETF demand remain subdued. BTC is trading within a major cost-basis cluster between $62,000 and $68,000, with $69,000 as key resistance. Glassnode said the current bear market is the shallowest on record by drawdown, but has not lasted long enough to confirm a bottom, while its model remains mildly “Risk Off.”

HIVE Executive Chairman: GPUs for AI Compute Earn 10x as Much per Hour as Mining Rigs
Frank Holmes, Executive Chairman of HIVE Digital Technologies, one of the world’s leading publicly listed Bitcoin mining companies, said during TheStreet Roundtable on July 29 that the company is accelerating the expansion of its AI computing business.
HIVE expects AI’s share of revenue to rise from around 10% to more than 50% this year. Holmes said ASIC mining machines generate approximately $0.12 to $0.14 per hour, while an H100 GPU can generate close to $2 per hour. Even older GPUs can earn about $1.40 per hour, roughly 10 times as much as mining machines.

SEC Chair: Will issue crypto market rules if the Clarity Act fails to pass
On July 27, 2026, SEC Chairman Paul Atkins discussed the Clarity Act and the regulation of the U.S. crypto market structure during a CNBC interview. He stated that he is optimistic about Congress passing the bill, with the SEC actively assisting by answering questions and providing technical support.
If Congress ultimately fails to act, the SEC is ready, willing, and able to issue rules to address the issues involved in the Clarity Act and improve regulation in other areas of the crypto market. However, Atkins believed that SEC rulemaking cannot replace congressional legislation. The crypto market ultimately still needs a statute to provide more lasting certainty and establish clear safeguards for future regulatory direction and market development.

Paris-based stablecoin card issuer Kulipa suddenly shut down approximately four months after completing a $6.2 million seed round, rendering U cards issued through its services by around 20 wallets and crypto companies, including Ready and Solflare, unusable. Kulipa announced the funding in early April, co-led by Flourish Ventures and 1kx, with participation from White Star Capital and Fabric Ventures. Because Kulipa used a self-custody model, user funds were only withdrawn when cards were used, and no customer balances were held by Kulipa, leaving users’ assets unaffected.

U.S. Senate Minority Leader Chuck Schumer introduced a bill to create an agency overseeing executive-branch corruption, citing President Donald Trump’s disclosure of over $2B in 2025 investment income, including $1.4B tied to crypto. The bill also cited his family’s ownership of a crypto fund worth over $1B linked to foreign governments. Senators Andy Kim, Alex Padilla and Jeff Merkley cosponsored the bill.

SPONSORED BY SPOONOS
During the 2026 World Cup, SpoonOS put eight large language models through the same experiment: give them the same prompts, information sources, starting capital, and agent framework, then compare how they predicted matches and managed money.
By the end of the 104-match tournament, the models had produced broadly similar prediction scores—but very different trading results.
SpoonOS is an agentic operating system built for Web3 developers and powered by Neo. Neo describes SpoonOS as an operating system designed to support AI agents and AI-native applications in decentralized environments.
SpoonOS Arena was launched during the World Cup as an experimental product. Its format was broadly similar to Alpha Arena, which compares AI models trading in live markets with identical prompts and input data. SpoonOS applied the concept to sports prediction markets.
How the eight models were selected
SpoonOS initially evaluated 11 models before choosing Claude, Gemini, GPT, Grok, Qwen, GLM, MiMo, and DeepSeek for the experiment.
According to the project’s launch announcement, the selection was based on how effectively each model operated within its Agent Group framework. The evaluation considered reasoning quality, response speed and decision-making performance. Models were excluded for reasons including excessive inference latency or weaker results in internal sports forecasting and capital-allocation simulations.
The final list should therefore be viewed as a practical selection for this experiment, rather than a comprehensive ranking of the eight best models available.
Each model operated under the same conditions:
The same prompts and information sources
The same Agent Group orchestration framework
The same decision-making workflow
The same starting capital of $500
A Master Agent coordinated the process, synthesizing the available information and applying the Kelly Criterion to determine whether to place a bet, which outcome to select and how much capital to commit.
The models could also decide not to bet. Their wallet addresses and trading activity were made publicly viewable on-chain.
Similar predictions, different returns
Across the tournament, the eight models made 632 predictions with verifiable results. Of those, 454 were correct, producing an overall accuracy rate of 71.8%.
They also placed 616 bets, but only 263—or 42.7%—were profitable. Even so, the eight wallets ended the session with a combined net profit of $1,826.62.
Prediction accuracy varied relatively little. DeepSeek, Claude, and GLM each recorded 73.4%, while Gemini, MiMo, and GPT finished at 72.2%. Only five percentage points separated the highest and lowest results.
Trading performance produced a much wider range.
DeepSeek finished first with a net profit of $2,591.04, followed by Gemini at $401.71 and GLM at $192.35. The other five models ended the session with losses.
DeepSeek made money on exactly half of its 82 bets. Its advantage came from the payoff profile: the average winning bet returned $175, compared with an average loss of $111.81.
Claude recorded the highest profitable-bet rate at 63%, but still lost $47.77 because its average loss was larger than its average gain. GPT achieved a prediction accuracy of 72.2%, but only 24.2% of its 95 bets were profitable. It finished last with a loss of $468.79.
Individual matches also caused sharp changes in the rankings. DeepSeek earned $1,090.82 from the draw between Spain and Cabo Verde Islands, the largest single-model profit of the session. Qwen and Gemini later backed the draw between Ecuador and Curaçao, briefly moving Gemini into first place.
There were collective failures as well. In Belgium versus Iran, all eight models placed bets and all eight lost money.
SpoonOS said the World Cup session attracted about 29,000 users and generated roughly 38,000 page views.
What comes next
Following the World Cup, SpoonOS plans to continue Arena across football and other competitive sports, including basketball, tennis and esports, according to its session recap.
The project also plans to introduce longer-term rankings and broader model profiles covering prediction accuracy, betting frequency, market preferences and maximum drawdown. More detailed match pages are intended to show why a model placed a bet—or chose to stay out.
The World Cup results should still be read with some limits. Giving models the same prompts, information and execution framework makes the comparison more consistent, but trading profit is not a pure measure of forecasting ability. It also reflects price judgment, position sizing, risk appetite, execution and short-term variance.
In fact, the models did not produce radically different prediction accuracy. The larger difference appeared after the prediction was made: whether to trade, how much to risk and how gains and losses accumulated.
This content is provided for general informational purposes only and does not constitute financial, investment, trading, betting, gambling or prediction-market advice. Prediction markets involve the risk of loss and may be restricted in certain jurisdictions. Results from a single session and past performance do not indicate future outcomes.
Data note: Prediction accuracy is based on predictions with verifiable match results. Trading returns are calculated from final wallet balances at the end of the World Cup session.

Strategy said during its Q2 earnings call that it will continue selling Bitcoin when advantageous to replenish its USD reserve, fund preferred dividends and interest payments, and support share buybacks. Management also said future capital raises will no longer be allocated entirely to Bitcoin purchases, with proceeds dynamically split between BTC and USD reserves based on market conditions. The company added that Bitcoin-backed borrowing is not currently under consideration due to counterparty and margin risks.

Coinkite warned that wallets created on Coldcard Mk3 devices running firmware versions 4.0.1 through 5.0.3 may be at risk due to a seed generation flaw. Users are advised to migrate funds immediately. The warning comes as investigators continue probing the recent theft of 594 BTC from hundreds of single-signature wallets, though no evidence has confirmed a link between the two incidents.

Coinbase reported Q2 2026 total revenue of $1.22 billion, down 19% year over year and 14% quarter over quarter, while transaction revenue fell 21% to $599 million. The company recorded a net loss of $359 million and adjusted EBITDA of $208 million. Crypto spot trading volume declined 24% to $146.4 billion, but Coinbase’s share of total crypto trading volume rose from 9.1% to a record 10.3%.
Subscription and services revenue fell 5% to $555 million, representing 48% of net revenue, while revenue outside BTC spot trading accounted for 88%. Prediction markets revenue increased 106% quarter over quarter and exceeded $100 million on an annualized basis. Average USDC held in Coinbase products reached $20 billion.

Strategy (Nasdaq: MSTR), formerly MicroStrategy, reported a net loss of $8.22 billion for the second quarter of 2026, compared with net income of $10.02 billion a year earlier, driven by fair value changes in its bitcoin holdings. As of July 26, the company held approximately 843,775 BTC, up 25% year to date, and raised about $17.06 billion through its at-the-market (ATM) offering programs.
Strategy said it increased its USD Reserve to $3.75 billion, covering more than 2.1 years of preferred stock dividend and debt interest obligations. Its convertible debt balance declined to $6.71 billion at the end of the quarter, from $8.21 billion previously.

Binance Research report found that the crypto market experienced broad onchain contraction rather than sector rotation in the first half of 2026. Total DeFi TVL declined by $43.4 billion (38%), while the combined market capitalization of six major L1s fell by $246.5 billion (42%). Key findings include: Ethereum spot ETF holdings dropped to 5.2 million ETH, while DAT holdings increased to 7.7 million ETH.
L2 user activity declined sharply, with user operations falling about 77% from January to June; Solana network revenue decreased 64.5%; BNB Chain was the only major deflationary L1, with an annualized burn rate of 5.05%; the industry recorded 207 security incidents in H1, resulting in $972 million in losses; driven by the World Cup and non-sports events, prediction market monthly nominal trading volume surged 86% to $51.6 billion, with Kalshi and Polymarket accounting for 92% of June’s total trading volume.

According to The Block, Bitcoin treasury company Hyperscale Data sold approximately 100 BTC and opened a Bitcoin-backed credit facility to fund its AI data center in Michigan. The company said it is using part of its Bitcoin reserves for AI infrastructure while borrowing against the remaining Bitcoin holdings as collateral. If all AI data center service contracts are extended, the company expects cumulative revenue to exceed $1.2 billion.

Quantum Solutions, Japan’s largest Ethereum treasury company, announced that its subsidiary GPT Pals Studio Limited sold 1,000 ETH for approximately $1.903 million on July 30, with the proceeds to be used for AIDC businesses such as AI data centers. This marks the company’s second ETH sale, bringing total sold ETH to 1,904, while its holdings have declined about 29% from the peak to 4,764.80 ETH. Meanwhile, the company raised the group’s cumulative ETH sale limit from 1,875 ETH to 4,375 ETH, meaning it could sell up to an additional 2,471 ETH, equivalent to about 52% of its current holdings.

U.S. initial jobless claims for the week ending July 25 came in at 197,000, below expectations of 200,000. The previous reading was revised up from 187,000 to 188,000. The U.S. core PCE price index rose 3.3% year over year in June, in line with expectations, compared with 3.4% previously.

The return of term premium, the options market’s bimodal pricing, and what fiscal dominance means for global assets.
Measured by the simplest yardstick — the ten-year Treasury yield minus the federal funds rate — the relative funding cost of the United States, the anchor of the global financial system, now sits in the same band as Germany’s. The shape of the full curve, however, reveals that the point at which the market charges America most dearly is not the ten-year but the twenty-year maturity. This is no ordinary cyclical episode; it is a sovereign credit re-rating written into the curve itself.
On 27 July 2026, the ten-year Treasury yield closed at 4.65 per cent against an effective federal funds rate (EFFR) of 3.63 per cent — a spread of roughly 102 basis points. In isolation, that is barely a third of the 1994 “bond vigilante” peak. But widen the lens from a single point to the whole curve, to the global cross-section, and to the distributions implied by options markets, and a fuller, more cautionary picture emerges: what is being impaired is not any one maturity but an era — the era in which Treasuries, as the world’s risk-free asset, enjoyed the subsidy of a negative term premium.
I. A curve that sits above the policy rate at every point
Subtracting EFFR from every point on the Treasury curve of 27 July yields the first fact: from one-month bills to thirty-year bonds, every tenor trades above the policy rate (Figure 1). One month is 17bp over, one year 51bp, two years 68bp, ten years 102bp — and the twenty-year, at +152bp, is the highest point on the entire curve, above even the thirty-year (+149bp), leaving 20s30s inverted.
Figure 1 The US Treasury curve: September 2024 (before the first cut) versus July 2026. The whole curve has shifted up 240–290bp in 26 months — cyclical steepeners rotate; regime re-pricings shift in parallel.
The distribution of slope across segments is more informative than any single spread. From two to five years the curve travels barely 9bp in three years — almost dead flat: the market prices no path back to the old rate regime. Five to ten years adds 25bp; ten to twenty years jumps 50bp. Nobody prices “the policy rate in year fifteen”; that segment is almost pure term premium and duration-supply premium. The market’s marginal price for American duration peaks at twenty years — precisely where pension-fund demand is thinnest and supply is most purely fiscal.
The front end tells the opposite story. One year at +51bp and two years at +68bp price a hawkish path — no cuts over the coming year, possibly further hikes. That is the policy story of the 2026 Middle East energy shock, not a credit story. Looking at the ten-year point alone conflates the two.
Set against historical cross-sections (Figure 2), three points stand out. First, the wide-spread episodes of 2003, 2010 and 2013 all had front ends belowthe policy rate — cuts were priced, a benign “recovery steepener”. The 1993–94 vigilante episode is of the same family as today: front end above the funds rate, long end at a wide premium. Second, today’s +102bp at ten years is only a third to a half of October 1993 (+219bp) or November 1994 (+330bp); yet the pure duration premium once policy expectations are stripped out (20Y minus 2Y) already stands at 84bp — about two-thirds of the vigilante peak (124bp) — while federal debt, at roughly 120 per cent of GDP, is nearly double the ~64 per cent of 1993. Third, from September 2024 (the whole curve 132–192bp below the funds rate) to today (33–152bp above), the entire curve has shifted up 240–290bp in parallel. When the ten-year touched 5 per cent in October 2023, the curve was deeply inverted — a “tightening shape”; today’s is a “term-premium shape”. The two are wholly different animals.
Figure 2 Yield minus policy rate by tenor (bp): eight historical cross-sections. The darker the red, the more the market charges the Treasury.
KEY TAKEAWAY：
The impairment is structural, not localised: policy expectations explain only the front end (2Y at +68bp over EFFR); the long end is pure term premium, with the twenty-year as the epicentre where the market charges America most. The pure duration premium already equals two-thirds of the 1994 vigilante peak. A parallel 240–290bp upward shift in 26 months is the signature of regime repricing — cyclical steepeners rotate, regimes shift.
II. The global league table: level with Germany, neck-and-neck with India at the front end
Applying the same yardstick to the major economies (Figure 3) yields a counterintuitive result: on “10Y minus policy rate”, the United States (+102bp) sits almost level with Germany (+88bp), and better than Britain (+125bp), France (+167bp), Italy (+169bp) and Japan (+178bp). But 2026 is the year of a global energy shock: the ECB, the Bank of Japan, the RBA, the Bank of Korea and the RBNZ have all turned hawkish, and term premia have expanded across the developed world in unison. America’s mid-table ranking is partly camouflage provided by a crowded ward.
Figure 3 Sovereign yields minus domestic policy rates by tenor (bp). The United States highlighted.
Decomposed by curve segment, three details matter. First, at the front end (2Y minus policy rate), the United States (+68bp) sits almost exactly alongside India (+72bp), Italy (+74bp) and France (+71bp) — a BBB-rated emerging market prices its front end a mere 4bp from the issuer of the global reserve currency. To be fair, this segment chiefly reflects the shared pricing of the 2026 tightening cycle — a policy story, not a credit one; but it also means the Fed’s credibility no longer buys the US Treasury any front-end discount.
Second, on the back-end league table (30Y minus policy rate), America remains on the “core credit” side, but only one position ahead of Britain and India. The full ranking reads: Japan (+288bp) > Italy (+250bp) > France (+244bp) > India (+215bp) > Britain (+192bp) > United States (+149bp) ≈ Canada (+155bp) > Germany (+136bp) > Australia (+118bp) > China (+79bp).
Third, the American impairment takes the form of the whole curve being lifted, not the long end whipping away: the 30Y–10Y slope of the United States (+47bp) is nearly identical to Germany’s (+48bp) and Australia’s (+51bp), and nothing like Japan’s (+110bp) or Italy’s (+81bp).
Adding the debt stock makes the picture actionable. Divide the ten-year spread by debt-to-GDP and you have what the market charges per point of debt (Figure 4): India roughly 2.6, France 1.45, Germany 1.40, Britain and Italy about 1.25 — and the United States just 0.85, the lowest in the table bar Japan (0.77), where the central bank itself is the buyer of last resort. The market is still granting Treasuries a “reserve-currency discount”. Were that discount to mean-revert to the G10 median (~1.25), the ten-year spread would widen to 150–170bp — roughly another 50bp of “normalisation” upside in long yields, requiring no crisis at all, only the market ceasing to price the privilege. Japan illustrates the alternative terminal state: with the central bank buying bonds as a matter of routine, the spread can be suppressed at 0.77 — at the price of the currency and the central bank’s balance sheet.
Figure 4 Debt/GDP versus the ten-year spread: the US sits well below the G10 “price of debt” reference line — the gap is the reserve-currency discount.
Term-premium models say the same. The New York Fed’s ACM ten-year term premium has risen to +0.72 per cent, the San Francisco Fed’s Christensen-Rudebusch model to +1.25 per cent — while the two-year premium is just +0.21 per cent: the impairment is located precisely in the long end. The SF Fed’s decomposition shows that of the ten-year yield, the average expected overnight rate over the next decade is only 3.47 per cent — below today’s EFFR — leaving some 1.25 percentage points of pure term premium: the elevated long end can no longer be explained by “the market thinks the Fed will tighten”; it is a straightforward sovereign-credit and duration surcharge. Between 2016 and 2021 that premium was negative — a global shortage of safe assets subsidised the US Treasury. The refund of that subsidy is the essence of this re-rating.
KEY TAKEAWAY：
In the cross-section America trades level with Germany, in the circle of Canada and Britain, and neck- and-neck with India at the front end — yet as issuer of the global reserve currency it historically belonged systematically below that benchmark. Its “spread per unit of debt” of 0.85 is the lowest in the table bar central-bank-underwritten Japan: the reserve-currency discount survives, but mean reversion to the G10 median of 1.25 implies roughly 50bp more “normalisation” in long yields — requiring no crisis, only the market ceasing to price the privilege.
III. The options market’s bimodal pricing
The cash curve tells us how much has been priced; the options market tells us what is still feared. As of 28 July, four markets are telling four different stories:
The numbers are internally inconsistent: rates options, the SKEW index and gold volatility are pricing fiscal stress, while the 25-delta equity skew and bitcoin volatility price business as usual. In plain terms, the market is pricing a bimodal distribution: a low-volatility centre of inertia, a discontinuous fiscal event in the deep tail, and an empty middle. Historically, such gaps have almost always closed with equity vol catching up to rates vol — October 2022 and October 2023 both followed that script — and the flatness of the 25d skew means equity downside protection is underpriced relative to the risk implied by the rates market.
The cross-asset implications are best read market by market.
US equities: three transmission channels. The discount-rate channel compresses multiples, with long-duration growth first in line; the Kalecki profits channel — a fiscal deficit of ~7 per cent of GDP is, accounting-wise, tantamount to private-sector surplus and nominal corporate earnings — props up nominal profits, producing a grinding, narrow index; and the correlation channel keeps the stock–bond correlation positive, stripping 60/40 and risk parity of their diversifier and forcing vol-target funds to deleverage in tandem whenever rates vol spills over. The winners are pricing-power companies, energy and curve-steepening beneficiaries such as banks; the losers are long-duration tech, bond proxies and small caps reliant on floating-rate funding.
Commodities: gold is the de-dollarisation hedge; oil is the front end’s driver. Gold is the market’s chosen “de-dollarisation hedge” in this episode — after a 27 per cent correction its IV remains pinned at 21–22 with call skew intact, evidence that the structure of central-bank buying underneath and options-market insurance on top is unchanged; oil drives the hawkish front end and is priced as “range-bound with upside skew”.
Crypto: the harshest verdict of 2026. In the first genuine year of sovereign-credit stress, capital chose gold, not bitcoin. Bitcoin traded all year as liquidity beta, suppressed by high real rates; the fulfilment of its debasement-hedge narrative requires the second phase — the central bank forced to monetise the fiscus — not the present first phase of a hawkish front end plus rising term premium.
KEY TAKEAWAY：
The options market is pricing a bimodal distribution: rates options, SKEW and gold volatility already pay for fiscal stress, while the 25-delta equity skew and bitcoin volatility still price business as usual. Historically such gaps close with equity vol catching up to rates vol — and while the centre is calm and the deep tail expensive, downside convexity at 25 delta is the underpriced window.
IV. History offers four endings
American sovereign credibility has been impaired four times before, and the menu of endings is fixed.
1933: creditor terms rewritten. Roosevelt abrogated the gold clauses in Treasury bonds, upheld by the Supreme Court in the Perry cases — America has technically rewritten its creditors’ terms once already.
1942–51: fiscal dominance, literally. The Fed pegged the curve outright for the war effort (bills at 3/8 per cent, bonds at 2.5 per cent); the ending was the inflation tax of 1946–48 (15–20 per cent) plus the 1951 Treasury–Fed Accord that restored independence. That is also the template for “if independence is lost”: yield-curve control.
1971–81: the closest analogue to today. Nixon pressed Burns, central-bank credibility was lost, and through the 1975–77 easing cycle the long end refused to follow — the identical curve shape to today’s. It ended with the 1978 dollar crisis, the Treasury forced to issue Deutschmark- and Swiss-franc-denominated “Carter bonds”, and Volcker taking rates to 20 per cent to rebuild credibility.
1992–94: the good-ending template. The bond vigilantes killed Clinton’s stimulus, forced the 1993 deficit-reduction act, and were rewarded with the surpluses of 1998–2001 and a converging spread.
The rule is singular: the ending is either fiscal consolidation (the 1950s, the 1990s), inflation and monetary subordination (the 1940s, the 1970s), or an external discipline event (Volcker). There has never been a default. And each time, the dollar system emerged more entrenched. So “deep impairment” is not destiny — but the political spectrum of 2026 shows neither a Volcker nor a Clinton, which is precisely why the options market has bid the deep tail so dearly.
KEY TAKEAWAY:
The menu of endings has only ever had three items: fiscal consolidation (the 1950s, the 1990s), inflation and monetary subordination (the 1940s, the 1970s), or an external discipline event (Volcker). There has never been a default — and each repair left the dollar system more entrenched. Impairment is not destiny; but the political spectrum of 2026 shows neither a Volcker nor a Clinton, which is why the deep tail is so expensive.
V. Trump: accelerant, not origin
Attributing the re-rating wholly to the Trump administration fails the timeline: the bear- steepening divergence between the ten-year and the funds rate began in September 2024 — before Trump took office. The full attribution has three layers. The foundation was laid by both parties (2008–21): the crisis response, the 2017 full-employment tax cut, the two rounds of Covid stimulus, and QE suppressing the term premium into negative territory — a subsidy that taught two generations of congressmen that deficits were costless. The trigger was pulled in 2022–24: the inflation breakout activated the “r greater than g” arithmetic, QT removed the marginal buyer of duration, the freezing of Russia’s reserves in February 2022 set off global reserve diversification, and Fitch (August 2023) and Moody’s (May 2025) successively stripped America’s top rating. Trump 2.0 is the accelerant, working through three channels: deficits of ~7 per cent of GDP at full employment, unprecedented in peacetime; public pressure on the Fed and key personnel choices, eroding the “independence premium”; and tariffs plus immigration restrictions prolonging inflation stickiness, pinning the front end hawkish.
In other words, Trump is neither the direct cause nor an irrelevance — he is best understood as a symptom and amplifier of the post-2008 fiscal equilibrium (an electorate that rewards deficits, and two parties in collusion). Even a fiscally conservative successor could slow but not reverse the trajectory: 120 per cent debt with r above g requires primary surpluses, on which no candidate campaigned in 2024.
KEY TAKEAWAY:
“All Trump” fails the timeline — the bear-steepening divergence began in September 2024, before he took office; “nothing to do with Trump” fails on the marginal contributions — ~7 per cent deficits at full employment and a discounted Fed-independence premium are genuinely new. The debt base was built by both parties (2008–21), the trigger was pulled in 2022–24, and Trump 2.0 is the accelerant — and a symptom of the same fiscal equilibrium.
VI. Emerging markets and “neutral” strategies: two shapes of the same storm
For emerging markets, the shock arrives bifurcated — and in the very week of writing, the bifurcation played out in its most extreme form.
(i) The AI-hardware economies: a leveraged mania, liquidated
After peaking near 9,400 in late June — up as much as 116 per cent year-to-date — South Korea’s KOSPI entered a technical bear market on 8 July, crashed 8.95 per cent on “Black Monday” 13 July, fell a further 5.73 per cent on 24 July, and on 28 July plunged 10.84 per cent, its eighth market-wide circuit breaker of the year, to close at 6,023.66: a maximum drawdown of more than a third from the peak. Samsung Electronics and SK hynix fell 13.39 and 14.65 per cent on the day.
Retail leverage was the amplifier. Sixteen single-stock 2x leveraged ETFs, approved on 27 May, attracted nearly 12 trillion won in fifty days — over 90 per cent of it into Samsung and SK hynix alone — and the “fall, forced rebalancing sale, fall further” death spiral left more than 1.2 million leveraged accounts receiving margin calls and hundreds of thousands forcibly liquidated. The exchange has triggered 40 sidecars and eight circuit breakers this year; the KOSPI volatility index reached 97.99 at end-June, near an all-time high.
The transmission chain is traceable. Meta’s latest $12.5bn data-centre bond priced at roughly 5.0 per cent, well above the ~4.2 per cent of its 2025 issuance — the discount-rate repricing of AI capex has begun; TSMC’s June revenue turned negative month-on-month and its capex guidance was raised above $60bn, sparking a global “peak compute, memory glut” chip selloff. Korea is the most concentrated economy on that chain (two stocks exceed half of index market capitalisation), with the most retail leverage, and a central bank still hiking while the won stays weak despite large surpluses. Chinese A-shares moved in sympathy: on 28 July the ChiNext index slumped 7.35 per cent, its worst day in over a year, the Shanghai Composite barely held 3,800, and memory, optical-module and semiconductor names led the decline while banks and liquor stocks rose.
(ii) The failed hedge and the soft dollar: this is no taper tantrum
The most regime-relevant detail is the failure of bonds to hedge: on 28 July, as Asia-Pacific equities crashed, the ten-year Treasury yield rose rather than fell, holding a 4.6–4.7 per cent range — the old “equities crash, Treasuries rally” relationship did not appear. The same day, Nasdaq futures fell 2.29 per cent while Dow futures rose 1.12 per cent, and cash-rich software rallied as chips slumped. This is not recession fear but a repricing of discount rates and cash- flow duration: money has not left the building, it has moved from long-duration assets to cash- generative ones — the standard signature of positive stock–bond correlation and systematic duration repricing under fiscal dominance.
The dollar’s position is equally telling: the dollar index closed at only about 101.6 on 28 July, in the lower half of its multi-year range. Unlike the 2013 taper tantrum, this is not a strong-dollar squeeze — the risk emanates from America’s own fiscal and AI-funding repricing, and the dollar did not strengthen into the shock. The IIF records non-resident portfolio outflows of $26.6bn in May and $17.8bn in June, after a record $98.8bn January inflow — the full stress sequence since the Iran war (Figure 5).
Figure 5 EM non-resident portfolio flows: from a record January inflow to two consecutive months of outflows.
Derivatives pricing matches the cash picture: South Korea’s five-year CDS sits at just 52.5bp, China’s at 31bp — credit is not yet pricing stress; the stress is concentrated in flows and volatility — again the bimodal signature of a calm centre and a moving tail. China’s position is distinctive: a ten-year yield of 1.73 per cent, the lowest curve-to-policy spread in the table, calm CDS, and a Hang Seng index up 9.9 per cent in July — the “anti-trade” pole of the global term- premium storm, exporting disinflation while attracting reserve-diversification demand for renminbi assets. The A-share decline is the resonance of the global AI-chain repricing with domestic liquidity events — ChangXin’s 57.9bn-yuan IPO (the STAR Market’s largest ever, freezing roughly 1.7 trillion yuan of subscription funds), extreme crowding (TMT above 45 per cent of turnover) and a ten-day falling margin balance — a structural clear-out, not a systemic bear market.
(iii) Neutrality is no immunity: three transmission channels
For “strictly neutral” strategies — volatility strategies, dollar-neutral long/short, statistical arbitrage — one illusion must first be dispelled: neutrality hedges direction, not regime. The shock travels through three channels. The funding channel: with EFFR at 3.63 per cent and bills at 3.8–4.0 per cent, the cash hurdle for every neutral strategy is roughly 400bp higher — gross exposure must earn four extra points merely to stand still, while leverage costs (repo, swap financing, borrow) rise in step. The correlation channel: in a positive stock–bond-correlation world, “dollar neutral” is not “duration neutral” — a book long growth and short value carries an implicit short-duration exposure, rate-driven factor rotations trigger crowded unwinds (the 2022 “quant winter” is the template), and pairwise correlations converge to one in the tail, compressing stat-arb Sharpe first. The crowding channel: when term premium becomes the dominant macro variable, every macro-driven quant fund de-risks on the same signal at the same time — neutral strategies rarely die of direction; they die of funding, crowding and correlation spikes, as in the August 2007 “quant quake”, February 2018, and the UK LDI episode of October 2022.
(iv) The other side of the coin: historic soil for volatility strategies
It must be said that the current regime is also historically the richest soil for disciplined volatility strategies. AI-driven concentration offers dispersion opportunities — high single-stock vol against low index vol — in America and Korea alike; the gap between the flat 25d equity skew and the extreme SKEW is a relative-value vol window; and the rates-vol/equity-vol gap offers a convergence trade with negative carry but positive expectation. What should genuinely be avoided is leveraged carry and short gamma: a bimodal distribution means jump risk is rising, and margin and VaR shocks always force deleveraging at the worst moment. For a multi-pod fund built around volatility and derivatives, the message of the current regime compresses into one sentence: the exposure has not disappeared — it has migrated from delta to gamma, funding and crowding. Term premium itself has become a directly tradable risk factor: long 10s20s steepeners and back-end payer vol, long gold 25d call skew, long equity downside convexity while the 25d skew is flat — three legs pricing the same thing, and the difference in how far each has priced is itself the source of alpha.
KEY TAKEAWAY:
Two ends of the same chain: Korea’s leveraged liquidation is where “Treasury term premium → AI funding costs → long-duration repricing” met a uniquely fragile retail microstructure, while the soft dollar, unmoved credit and the failed bond hedge prove this is regime repricing, not a dollar squeeze or recession scare. For neutral strategies, the exposure has not disappeared — it has migrated from delta to gamma, funding and crowding.
VII. In place of a conclusion: a monitoring checklist
This re-rating is neither vindication of the “collapse” thesis nor continuation of “exorbitant privilege” as usual. It resembles a blend of 1993 and 1975: some distance remains in magnitude; the mechanism is already of the same kind. For investors, triggers are more useful than opinions:
· Term premium: whether the New York Fed’s ACM and San Francisco Fed’s CR readings keep rising;
· The equity-skew gap: the direction in which the gap between SPY 25d skew and the SKEW index closes;
· Hedge structures: the resilience of gold’s 25d call skew, and the percentile turn in bitcoin skew;
· Supply digestion: tails at twenty-year Treasury auctions;
· Emerging markets: the IIF’s monthly flows, plus the right-tail frenzy of the AI-hardware economies tracked by VKOSPI and Korea’s leveraged-ETF assets;
· The regime signal: whether Treasury yields still refuse to fall on risk-off days — the failure of the bond hedge is itself the signal.
When the “spread per unit of debt” converges from 0.85 towards the G10 median of 1.25, and when the 30Y–10Y slope migrates from +47bp towards the British and French shapes, “deep impairment” will pass from pricing structure to pricing consensus. History says the window before consensus forms has always been the best entry point for exactly this kind of trade.
This report is based on public data and reasonable inference and does not constitute investment advice.
Follow us
Twitter: https://twitter.com/WuBlockchain
Telegram: https://t.me/wublockchainenglish