U.S. Treasury yields retreated sharply on Wednesday from multi-year highs after the Treasury Department announced plans to double the size of its government debt buybacks, providing support to longer-dated bonds following a global selloff driven by concerns over rising public debt, inflation and higher borrowing costs.

The 30-year Treasury yield fell nearly 9 basis points to 5.196%, while the benchmark 10-year yield declined about 6 basis points to 4.647%. The 30-year yield had climbed above 5.33% earlier this week, reaching its highest level in nearly two decades.

The reversal was concentrated at the long end of the yield curve, where investors have been demanding higher compensation to hold government debt amid concerns about the U.S. fiscal outlook and the risk that persistent inflation could keep interest rates elevated for longer.

The Treasury said it would increase the size of its buyback operations, a move designed in part to improve liquidity and market functioning in longer-dated securities.

The announcement helped ease some of the pressure on long-term bonds, but it does not reduce the amount of U.S. government debt outstanding.

“I’m assuming this supply will be replaced by more issuance on the shorter end, particularly bills. This is NOT a debt paydown, it is just a rearrangement of the maturity schedule of Treasuries,” said Peter Boockvar, chief investment officer at One Point BFG Wealth Partners.

Treasury buybacks can improve liquidity by purchasing older or less actively traded securities and can support prices in targeted parts of the market. But they do not eliminate the underlying fiscal challenge facing the U.S. government.

The U.S. fiscal deficit reached $432.3 billion in July, the largest monthly shortfall since March 2021, pushing the cumulative deficit for the year to almost $1.8 trillion. The cost of servicing the debt has also become a growing burden. Interest payments on the nearly $40 trillion national debt have reached about $1.2 trillion this year, increasing pressure on the federal budget as borrowing costs remain elevated.

The latest move in Treasury yields also comes amid a broader global bond selloff.

Japan’s 10-year government bond yield reached its highest level in three decades, while Germany’s 30-year Bund yield climbed to its highest since 2011. France’s 30-year borrowing cost reached its highest level since 2008.

The synchronized rise in long-term yields points to a broader reassessment by investors of sovereign debt risks rather than a problem confined to the United States.

Higher oil prices are adding another layer of uncertainty. Rising energy costs could feed into consumer inflation and make it more difficult for central banks to ease monetary policy. That prospect has encouraged investors to demand higher yields on longer-term bonds to compensate for the risk of persistent inflation.

The U.S. Federal Reserve’s policy outlook will receive additional attention later Wednesday with the release of minutes from its latest Federal Open Market Committee meeting.

The minutes are of the essence because the July meeting exposed unusually strong disagreement among policymakers. Three officials voted for a 25-basis-point rate increase, while the majority opted to keep rates unchanged. Investors will look for clues about the arguments behind those dissenting votes and whether concerns about inflation could prevent the Fed from cutting rates in the months ahead.

The combination of elevated inflation risk, heavy government borrowing and uncertainty over monetary policy has created a challenging environment for the Treasury market. Long-term yields matter well beyond government bonds because they influence corporate borrowing costs, mortgage rates and the valuation of equities.

Therefore, analysts see Wednesday’s retreat as marking a reprieve rather than a resolution. Treasury buybacks can provide additional liquidity and help stabilize longer-dated securities, but they do not address the underlying forces that pushed yields higher in the first place.

With the U.S. running large fiscal deficits, debt-servicing costs rising and energy prices threatening to reignite inflation, investors are expected to continue scrutinizing the government’s borrowing needs and the Fed’s willingness to tolerate higher inflation.

Anthropic is preparing to expand its borrowing capacity beyond a targeted $10 billion, adding another layer of financing to the artificial intelligence boom as the Claude maker lays the groundwork for a potential blockbuster initial public offering.

The company is in discussions with banks over a revolving credit facility that could exceed its original $10 billion target, according to people familiar with the matter cited by Bloomberg. The final size has not been determined and could still remain at $10 billion or be reduced, the sources said.

The move would give Anthropic access to a substantial pool of capital that it could draw, repay, and borrow again as needed, providing liquidity as it continues to spend heavily on computing infrastructure, model development, and other costs associated with scaling its AI business.

The financing discussions also carry significance for Wall Street banks. Lenders seeking a place in Anthropic’s expanded credit facility could improve their prospects of securing roles in what could become one of the largest technology IPOs in years. Morgan Stanley, Goldman Sachs and JPMorgan are reportedly working with Anthropic on the planned listing, while the company has recently held discussions with prospective investors.

Anthropic previously secured a $2.5 billion five-year revolving credit facility from a group of lenders including Morgan Stanley, Barclays, Citigroup, Goldman Sachs, JPMorgan, Royal Bank of Canada and Mitsubishi UFJ Financial Group. Expanding that facility would substantially increase the company’s financial flexibility ahead of a potential public offering.

Anthropic is entering the public markets at a point when investors are paying closer attention to how AI companies finance their extraordinary growth. Unlike more mature software companies, leading AI developers require enormous amounts of computing capacity, much of it backed by expensive Nvidia processors and data-center infrastructure.

A larger credit facility could therefore serve as a bridge between Anthropic’s current private-market funding model and the deeper capital markets available after an IPO. It could also provide the company with additional liquidity without requiring it to immediately issue more equity.

There is precedent for the strategy. SpaceX expanded its credit facility with several banks involved in its planned IPO roughly a month before its June public offering. Such arrangements can strengthen relationships between companies and investment banks at a critical stage of the IPO process.

The broader financing environment shows why Anthropic’s borrowing needs are becoming increasingly significant. JPMorgan estimates AI-related debt financing could reach $4.1 trillion through 2030, up from its previous forecast, as hyperscalers, data-center operators and chip buyers seek to finance the infrastructure required to support AI workloads.

AI-related debt issuance has already surpassed $300 billion in 2026, according to the bank, making data-center financing one of the most important sources of new corporate borrowing this year.

JPMorgan also expects AI capital expenditure to reach $5.5 trillion through 2030, up from its previous estimate of $5.1 trillion. The revised forecast is based partly on expectations that global data-center capacity will expand by 138 gigawatts by the end of the decade, compared with an earlier estimate of 122 gigawatts.

That spending boom is forcing companies across the AI ecosystem to look beyond traditional equity financing. Developers are increasingly using structures such as behind-the-meter power agreements, bring-your-own-power arrangements and more efficient computing systems to overcome constraints on electricity and infrastructure.

For Anthropic, the challenge is more than simply securing enough money to fund growth. The company must convince prospective public-market investors that the revenue generated by its AI models will eventually grow faster than the enormous cost of computing, training, and inference.

That makes the proposed credit facility an important signal of the capital intensity behind the AI race. A multibillion-dollar borrowing capacity would give Anthropic greater room to invest before an IPO, but it would also increase the financial obligations attached to a business whose long-term profitability is still being established.

The potential facility therefore indicates that in the AI industry, the race to build powerful models is becoming not only a technology competition but also a massive financing exercise. That is why banks, private credit investors and public-market investors are increasingly being asked to fund the infrastructure required to sustain it.

European government bond yields climbed to multi-year highs on Wednesday as investors extended a global selloff in sovereign debt, with rising government borrowing, persistent inflation and higher oil prices intensifying concerns about the sustainability of public finances.

The pressure was broad-based. German 10-year and 30-year Bund yields reached fresh 15-year highs, with the 10-year yield moving above 3%. France’s 10-year borrowing costs climbed to their highest level in 18 years, highlighting growing investor unease in some of Europe’s largest sovereign debt markets.

The move followed a similar rise in borrowing costs across the United States and Japan. The yield on the U.S. long bond steadied around 5.28% on Wednesday after climbing to nearly 5.34% on Tuesday, its highest level in almost two decades. Japan’s 10-year government bond yield moved toward 3%, reaching a three-decade high.

Bond yields rise as prices fall, meaning the latest moves represent a significant repricing of government debt. The increase is relevant for financial markets because long-term sovereign yields influence borrowing costs across the economy, from corporate debt and mortgages to the valuation of stocks and other risk assets.

“If you combine a sticky inflation environment and excessive government spending, then the natural move for bond yields is higher,” said Jason Da Silva, director of global investment strategy at Arbuthnot Latham.

He expects investors to exert pressure on governments more frequently as concerns about fiscal discipline grow.

“I think this is going to be the norm going forward. There are no aggressive measures by any Western governments to curb spending,” Da Silva said.

The bond selloff comes as governments across major economies face the difficult combination of elevated debt burdens, large fiscal deficits and higher financing costs. Higher yields increase the expense of servicing existing debt as it matures and is refinanced, potentially putting additional pressure on government budgets.

The problem is particularly sensitive at the long end of the yield curve because investors are demanding greater compensation for holding debt over extended periods amid uncertainty about inflation, economic growth and fiscal policy.

Japan’s rising yields add another dimension to the global bond selloff. For years, extremely low Japanese interest rates encouraged Japanese investors to seek higher returns overseas. A sustained increase in domestic yields could make Japanese assets relatively more attractive and potentially reduce some of the capital flowing into foreign bond markets.

“There is a narrative of are we going to have continued higher-for-longer inflation and what does that mean for longer-term interest rates?” said Neil Fisher, investment specialist at St James’s Place.

“Then you have a narrative around how sustainable is some of this long-term government debt in the UK and Europe, in the U.S. as well?”

Oil Adds To Inflation Risk

Higher oil prices are adding to the pressure on bond markets. Oil futures rose for a fourth consecutive session as expectations faded for a deal that could help end the conflict in the Middle East and ease disruption around the region.

The continuing uncertainty surrounding the Strait of Hormuz is weighing significantly on energy markets. Prolonged disruption to oil flows would increase the risk of higher energy costs feeding into consumer and producer prices, making it harder for central banks to bring inflation under control.

That could force investors to reassess expectations for interest rates and keep longer-term bond yields elevated.

The U.S. Federal Reserve was due to release minutes from its July meeting later Wednesday, when policymakers left interest rates unchanged. Investors will scrutinize the minutes for clues about how officials assess persistent inflation and the potential response if energy prices continue to rise.

The U.S. Treasury was also scheduled to sell $16 billion of 20-year bonds, providing another immediate test of investor appetite for long-duration government debt.

“Governments face a real choice between spending discipline and materially higher borrowing costs, and markets will keep testing which one they choose,” said Nigel Green, CEO of financial advisory firm deVere Group.

Stocks Come Under Pressure

The bond market turmoil spilled into equities, with European stocks edging lower and U.S. stock futures pointing to modest declines after a broad selloff in Asian markets.

South Korea was among the hardest hit. The benchmark Kospi fell nearly 6%, its largest one-day decline in three weeks, as investors reassessed the outlook for semiconductor companies. The sector has been one of the biggest beneficiaries of the global AI investment boom, leaving chip stocks exposed to any shift in expectations for AI-related spending.

Concerns were also heightened by reports that Anthropic’s annual revenue run rate had exceeded $65 billion by the end of July but was below some investor expectations. The development added to questions about whether rapidly rising valuations across the AI ecosystem are fully supported by near-term revenue growth.

The contrast between soaring AI investment and rising financing costs is becoming notable for markets. Higher long-term interest rates can reduce the present value investors assign to future earnings, placing particular pressure on high-growth technology companies whose valuations depend heavily on profits expected several years into the future.

China provided a notable exception to the broader risk-off mood. Unitree, described as the world’s largest humanoid-robot maker, surged 460% on its Shanghai debut after its offering was reportedly more than 8,000 times oversubscribed by retail investors. The extraordinary first-day move illustrates the continued appetite for AI and robotics-related investments even as broader markets become more cautious.

Dollar Steadies As Yen Nears Intervention Territory

The risk-off environment provided some support for the U.S. dollar, although the moves remained relatively modest. The dollar index was last down 0.3% at 99.382.

The euro gained about 0.25% to just above $1.16, while the yen traded around 159.15 per dollar, remaining close to the psychologically important 160 level.

A move beyond 160 could increase pressure on Japanese authorities to intervene in currency markets, given the potential inflationary impact of a weaker yen on imported energy and other goods.

The Canadian dollar strengthened slightly after U.S. President Donald Trump paused the imposition of a 50% tariff on Canadian goods for three days, saying the two countries had reached a deal.

U.S. Consumers in Focus

Investors will also be watching results from Lowe’s, Target and TJX on Wednesday for evidence of how U.S. consumers are coping with elevated prices and borrowing costs.

The earnings come after U.S. retail sales unexpectedly declined last week, raising concerns about the strength of household demand. The results could provide a more detailed picture of whether weaker economic data represents a temporary slowdown or a broader deterioration in consumer spending.

In Britain, inflation rose to 2.9% in July, in line with economists’ expectations, driven in part by higher household energy bills. The data provides another indication of how energy costs remain an important source of inflationary pressure.

The combination of higher oil prices, elevated inflation and rising government borrowing costs leaves global markets facing a difficult policy environment. Economists now expect central banks to balance the risk of allowing inflation to persist against the economic consequences of maintaining restrictive monetary policy, while governments face growing pressure to control spending as the cost of financing their debt rises.

Washington’s strategy of restricting China’s access to Nvidia’s most advanced artificial intelligence chips is facing a significant loophole: Chinese companies may be able to use the computing power of those chips without ever taking physical possession of them.

Several Chinese AI companies have reportedly accessed advanced Nvidia processors through data centers and cloud providers in Southeast Asia and other parts of Asia, exposing a gap in U.S. export controls that generally focus on where chips are shipped and who owns them rather than who can remotely use their computing capacity.

The issue is becoming crucial as Washington and Beijing compete for leadership in artificial intelligence. Nvidia’s most powerful processors are among the most critical pieces of infrastructure for training frontier AI models, making control over access to them a central element of U.S. efforts to slow China’s progress.

Yet the emergence of overseas compute markets means that restricting physical shipments to China may not be enough.

Chinese AI Firms Turn To Overseas Compute

The issue came into sharper focus after Moonshot AI released its Kimi K3 model in July.

Less than a week later, White House official Michael Kratsios accused Moonshot of using Nvidia GB300 chips through a facility in Thailand. Kimi K3 is among a new generation of Chinese AI models that have made significant gains in recent months. DeepSeek and Alibaba have also released models that have performed strongly on industry benchmarks.

Industry observers say access to large amounts of advanced computing capacity has been an important factor behind the rapid improvement of Chinese AI systems.

Cassia King, a senior researcher on the Compute Policy team at the Institute for AI Policy and Strategy, said Moonshot’s reported access to computing power in Thailand could be legal under current rules if the company did not actually purchase or own the physical chips.

“The U.S. export control regime controls physical AI chips. It does not cover remote access to those chips,” King said.

That creates a potentially important gap. A company in China may be prohibited from importing an advanced Nvidia processor, while a cloud provider outside China could potentially operate the same processor and sell computing time to Chinese customers.

The customer receives the computational output rather than the chip itself.

Southeast Asia Becomes A Strategic Computing Hub

Reports indicate that Chinese technology companies including ByteDance, Alibaba and Tencent have accessed Nvidia computing capacity remotely through facilities in countries such as Thailand, Malaysia and Japan.

ByteDance was reportedly working with Singapore-headquartered cloud provider Aolani to obtain access to Nvidia-powered computing infrastructure in Malaysia, according to a source familiar with the arrangement. The Wall Street Journal previously reported the arrangement.

Aolani said it serves a “global and diversified customer base spanning customers from North America and Asia.”

“The companies we service do not have ownership, potential future claim or physical access to the chips that power our solutions,” an Aolani spokesperson said, adding that permitted access to its services and infrastructure complies with applicable regulations.

The growth of Southeast Asia’s data-center industry makes the issue harder for Washington to contain.

Companies are investing heavily in data centers across Malaysia, Indonesia and Thailand as demand for AI computing increases. JLL estimates that global data-center capacity could roughly double to 200 gigawatts by 2030.

DC Byte data show 31 planned data centers of at least 100 megawatts across Malaysia, Indonesia and Thailand, compared with only two currently operating facilities of that scale.

That expansion is creating an increasingly important pool of computing infrastructure close to China, potentially giving Chinese AI companies another route to advanced processing capacity even as direct shipments of restricted chips remain tightly controlled.

Why Remote Access Matters

The strategic importance goes beyond individual Chinese companies gaining access to particular Nvidia chips. Training advanced AI models requires enormous amounts of computing power. If Chinese developers can obtain sufficient access to high-end processors through overseas cloud infrastructure, U.S. restrictions on physical chip exports may have a weaker effect on their ability to develop competitive models.

Michelle Nie, a visiting fellow in technology and national security at the Center for a New American Security, described the loophole as a threat to U.S. national security.

“The point of chip export controls is to deny China the ability to train frontier AI using advanced U.S. chips,” Nie said.

The challenge for Washington is therefore shifting from controlling the movement of hardware to controlling access to computing capacity itself.

That is considerably more complicated.

A physical semiconductor can be tracked through export documentation, ownership and shipping records. Computing power delivered remotely through the cloud can potentially be accessed by customers in another country without the underlying hardware crossing a border.

Lawmakers are considering legislation aimed at closing the gap.

The proposed Remote Access Security Act, or RASA, would expand U.S. export controls to cover remote, cloud-based access to certain critical hardware and software. The legislation passed the House of Representatives in January but has yet to clear the Senate.

If enacted, the measure could give U.S. authorities a legal framework for regulating who can remotely access advanced computing infrastructure. But passing the legislation would not immediately resolve the problem.

Nie said the bill would give the government authority to regulate remote access, but the administration would still need to establish specific rules governing which computing resources are restricted and which customers should be prohibited from using them.

King said the Bureau of Industry and Security could potentially move quickly once it had the necessary authority, but designing an effective system would be more difficult.

“The challenge will be in making a rule that’s effective and enforceable,” she said.

Policymakers would need to determine which levels of computing capacity should be covered, which customers should be restricted, and how cloud providers could implement reliable know-your-customer and customer-verification systems.

But the compliance burden could fall on cloud providers.

Under a remote-access regime, cloud providers operating advanced Nvidia hardware could potentially become responsible for screening customers and monitoring how their computing resources are being used. That could increase compliance costs and complicate the business of serving international customers.

Nvidia’s Chips Remain At The Center Of The AI Contest

The dispute reveals that Nvidia has become so important to the geopolitical competition over AI. The company’s advanced accelerators provide much of the computing power required to train and operate the world’s most capable AI models. Washington has therefore treated access to Nvidia’s most advanced processors as a strategic issue rather than simply a commercial trade matter.

But the emergence of overseas data centers shows that controlling the hardware supply chain and controlling access to computing power are two different challenges.

China does not necessarily need to import every restricted chip if companies can obtain access to the same processors through cloud infrastructure elsewhere.