Foreign investors have continued pouring money into U.S. corporate bonds even as Treasury yields have surged to their highest levels in years, challenging expectations that higher borrowing costs and currency-hedging expenses would push overseas investors away from American credit markets.

Net foreign purchases of U.S. corporate bonds reached $251 billion through the end of June, putting 2026 on track to approach last year’s record $392 billion, according to an analysis by Goldman Sachs chief credit strategist Amanda Lynam.

The resilience of overseas demand is notable because the benchmark 10-year Treasury yield closed at 5% on Tuesday after reaching its highest level since 2007. The move has intensified concerns about how much further yields could rise and whether expensive U.S. borrowing costs could weaken demand for dollar-denominated assets.

So far, foreign investors have continued buying.

“This is notable as the foreign appetite for US credit has persisted despite a range of headwinds in recent years, including fluctuations in the strength of the dollar and the cost of hedging,” Lynam wrote in a Tuesday note.

Foreign investors own roughly 29% of the U.S. corporate bond market, giving them an important role in determining demand and financing conditions for American companies.

Their continued participation suggests that higher Treasury yields have not yet fundamentally altered the relative attractiveness of U.S. corporate credit. Instead, overseas investors appear willing to absorb higher rates in exchange for the income and liquidity available in the world’s largest corporate bond market.

Higher Yields Have Not Broken Foreign Demand

The resilience of foreign demand comes at a critical point for U.S. fixed-income markets. Investors are assessing whether higher borrowing costs will put additional pressure on corporate bonds and other risk assets as they await the Federal Reserve’s policy decision on Wednesday. The rise in Treasury yields has already increased the benchmark against which corporate borrowing costs are priced.

Normally, a sharp increase in domestic yields can create several obstacles for foreign investors. Higher Treasury rates can make other markets relatively more attractive, while a stronger or more volatile dollar can alter returns for investors whose portfolios are denominated in other currencies. Currency hedging can further reduce the effective yield earned by overseas buyers.

Those forces have generated repeated predictions that foreign investors would eventually reduce their exposure to U.S. assets.

The data have not yet supported that outcome.

Europe has been the biggest source of foreign demand for U.S. corporate bonds since early 2022, accounting for 52% of net foreign purchases, according to Goldman. Asia accounted for 21%.

The figures also put the focus on Japan, where domestic bond yields have risen as policymakers have moved away from the ultra-loose monetary conditions that for years encouraged Japanese institutions to invest abroad.

A sustained increase in Japanese government bond yields could make domestic assets more competitive and encourage insurers, pension funds and other institutions to repatriate capital. That possibility has been closely watched by U.S. bond investors because Japanese institutions have historically been major participants in global fixed-income markets.

Goldman, however, expects any further decline in Japanese holdings of U.S. investment-grade and high-yield corporate bonds to remain manageable relative to the overall market. That assessment rests partly on the sheer scale and liquidity of the U.S. corporate bond market.

Few Alternatives Match the U.S. Market

For international investors, the decision is not simply whether U.S. yields have become more expensive. It is also what assets can replace them.

The U.S. corporate bond market offers a combination of scale, liquidity, credit diversity, and tradability that is difficult to reproduce elsewhere. Large institutional investors can deploy substantial amounts of capital across investment-grade and high-yield securities without necessarily sacrificing the ability to trade.

That market depth comes to the fore when investors are managing large portfolios. A modest improvement in the relative attractiveness of domestic bonds may not be sufficient to justify moving hundreds of billions of dollars out of a market that offers a much broader pool of issuers and securities.

Goldman therefore expects a floor to remain under foreign purchases of U.S.-domiciled credit and considers a broad repatriation of overseas capital unlikely.

“We continue to expect a floor to remain under foreign purchases of US-domiciled credit, and view a broader repatriation of flows as unlikely,” Lynam wrote.

The distinction between foreign demand for Treasuries and demand for corporate credit is also important. Corporate bonds offer investors additional compensation for taking credit risk, meaning that higher Treasury yields can actually provide a larger overall income opportunity when corporate spreads remain contained.

That does not make the market immune to a further rise in rates. If Treasury yields remain elevated, companies refinancing debt will face higher interest expenses, while existing bonds can suffer price declines as newly issued securities offer higher coupons. A deterioration in economic conditions could also widen corporate credit spreads and increase losses for investors.

But the foreign-flow data suggest that those risks have not yet been sufficient to trigger a broad withdrawal.

The bigger test may come if high Treasury yields persist rather than simply rise temporarily. A 5% 10-year yield changes the economics of global portfolios more substantially when investors believe rates will remain elevated for years. It can also alter corporate financing decisions, equity valuations and the relative appeal of other developed-market bonds.

For now, however, foreign investors appear to be treating the higher U.S. yield environment as an opportunity rather than a reason to abandon the market. That leaves U.S. corporate borrowers with an important source of external demand even as domestic financial conditions tighten. It also suggests that the long-running concern over foreign investors turning away from U.S. assets may be more complicated than a simple comparison of Treasury yields.

The dollar, hedging costs, domestic yields, and geopolitical considerations all matter. But so does market structure. Until another market can offer comparable scale and liquidity, Goldman expects foreign capital to keep finding its way into U.S. corporate credit.

South Korea’s SK Hynix is in talks with Intel about a potential deal that could bring the memory-chip maker’s manufacturing operations to U.S. soil for the first time, three people familiar with the discussions told Reuters, in a move that could reshape part of the supply chain for chips critical to artificial intelligence and data centers.

One option under discussion would see SK Hynix lease part of Intel’s long-planned semiconductor manufacturing complex in Ohio, two of the people said. Another possibility is a joint venture involving Intel and major cloud companies that want to secure reliable supplies of memory chips as AI infrastructure investment drives demand and tightens availability.

The discussions remain exploratory, and no agreement has been reached. One source said SK Hynix is considering multiple structures and that the company could ultimately pursue a different arrangement.

The talks nevertheless illustrate the growing pressure on memory-chip manufacturers to establish production capacity closer to their biggest customers. They also offer Intel a potential way to put its underused or delayed Ohio investment to work while sharing some of the enormous cost of building semiconductor manufacturing capacity in the United States.

For the Trump administration, an SK Hynix commitment would provide another high-profile example of a major Asian semiconductor company expanding manufacturing in the United States. Washington has been pushing chipmakers to invest domestically as it seeks to reduce dependence on Asian supply chains and capitalize on the rapid expansion of AI computing and data centers.

The discussions could become more complicated if SK Hynix seeks to manufacture advanced memory technologies in Ohio.

Seoul Faces a Semiconductor Dilemma

SK Hynix produces a broad range of memory products, including DRAM used in servers, PCs, and smartphones and NAND flash used for data storage. It is also one of the world’s leading suppliers of high-bandwidth memory, or HBM, a critical component in AI processors because it allows chips to move large amounts of data rapidly.

The sources said any proposal involving advanced memory such as HBM, or potentially even certain DRAM technologies, could face scrutiny from the South Korean government because the technologies are considered sensitive.

SK Hynix told Reuters that it was “reviewing various measures, including establishing additional production bases, to strengthen the competitiveness of its memory business,” but said that “no matters have been determined at this stage.”

Intel declined to comment on what it called speculation, while saying it continues to invest in Ohio and prepare the site.

South Korea’s trade ministry said decisions about overseas investment would be up to the company, but projects involving a designated “national core technology” would be subject to review under the country’s Industrial Technology Protection Act.

The situation has resulted in a difficult policy balancing act for Seoul. SK Hynix is under pressure to serve rapidly expanding global demand, particularly from U.S. technology companies, but South Korea also wants to preserve semiconductor manufacturing and technological capabilities domestically.

The issue has become sensitive because the government has been encouraging SK Hynix and rival Samsung Electronics to accelerate construction of a semiconductor manufacturing cluster in southwestern South Korea. Moving some advanced production to the United States could therefore satisfy Washington while potentially complicating Seoul’s efforts to deepen its domestic semiconductor base.

There are also significant economic disadvantages to U.S. production. Manufacturing chips in America is considerably more expensive than in South Korea because of higher labor and construction costs, according to two of the sources. Much of the semiconductor supply chain, including equipment, materials, and specialized suppliers, is concentrated in Asia, adding further costs to U.S. operations.

The commercial case for expanding in America therefore depends on more than manufacturing economics. Customer commitments, government incentives, supply security and the political risks associated with remaining heavily concentrated in Asia could all become part of the calculation.

SK Hynix already has a U.S. footprint under construction. The company completed a secondary Nasdaq listing in July and is building a chip-packaging facility in Indiana, but it does not currently fabricate memory chips in the United States.

SK Group Chairman Chey Tae-won said in July that the company was facing “enormous pressure” and lobbying from both customers and governments to increase chip supplies.

“I think we need to build a factory in the United States. If possible, I believe we should build it,” he told reporters.

The comments suggest the question is now about when and how SK Hynix expands U.S. manufacturing rather than whether the company should establish a larger American presence.

An Opening for Intel as AI Reshapes Chip Manufacturing

For Intel, a partnership with SK Hynix could help address one of the biggest challenges surrounding its Ohio project: the sheer scale and cost of the investment.

Intel announced in 2022 that it planned to invest as much as $100 billion to develop what it described as potentially the world’s largest chipmaking complex in Ohio. Production was initially expected to begin in 2025, but construction schedules have since been pushed back, with the site’s two planned plants now expected to be completed in 2030 and 2031.

Bringing in an established memory producer could give Intel a potential tenant, partner, or source of capital for infrastructure that otherwise requires years of investment before generating meaningful returns. It could also diversify the economic purpose of the Ohio site beyond Intel’s own manufacturing ambitions.

The potential timing is significant because AI has transformed memory from a relatively standardized component of computing into a major constraint on the expansion of data-center capacity. HBM has become more relevant because advanced AI accelerators require large amounts of high-speed memory, while only a small number of companies have the technology and manufacturing scale to supply it.

That dynamic is giving memory makers unusual bargaining power, but it is also increasing pressure from customers to expand capacity. A joint venture involving cloud companies would make economic sense in that context: hyperscalers could provide long-term demand commitments or capital in exchange for greater certainty over future memory supplies.

The proposal would also fit a broader shift in semiconductor manufacturing economics. Companies are now being asked to build where customers and governments want capacity, even when those locations are not the lowest-cost manufacturing bases.

Washington is intensifying that pressure. U.S. Commerce Secretary Howard Lutnick has threatened tariffs of as much as 100% on South Korean and Taiwanese companies unless they commit to greater production in the United States.

At the same time, Washington and Seoul are still negotiating the implementation of a major investment commitment South Korea made in the United States last year in exchange for lower U.S. tariffs. Of the proposed $350 billion commitment, about $150 billion has been earmarked for shipbuilding, while the remaining $200 billion has yet to be allocated.

Two of the sources said the unresolved investment discussions have complicated SK Hynix’s position. Seoul wants to use potential U.S. investment by its companies as leverage in broader negotiations with Washington, while the Trump administration is pressing SK Hynix to make a rapid commitment to expand its American manufacturing base.

The result is a semiconductor investment decision caught between three competing forces: the economics of manufacturing, the strategic demands of governments and the increasingly urgent needs of AI customers.

If SK Hynix ultimately manufactures memory in Ohio, the significance would extend beyond the individual deal. Analysts believe it would mark a step toward moving a portion of the world’s most important memory supply chain closer to U.S. data centers and technology companies. But the cost of that localization, and the question of which technologies Seoul is willing to see produced abroad, will determine how far that shift can realistically go.