BANGKOK — The International Monetary Fund has raised its forecast for global growth this year as the AI-driven investment boom helps offset some of the impact from surging energy costs, while warning that economic resilience could be nearing its breaking point.

The IMF expects the global economy to grow at 3.1% in 2026, up from the 3% projected in July, but slower than the 3.5% rate last year. For 2027, it kept the forecast unchanged at 3.4%.

The UN agency expects inflation at 4.1% in 2027, above the 3.9% it projected in July – when it estimated the Strait of Hormuz would reopen by mid-July, with conditions broadly returning to pre-war levels by March 2027. For the current year, it maintained its 4.7% outlook.

“The global economy has been resilient, but that resilience is reaching its limits,” Silvana Tenreyro, economic counsellor and director at the IMF research department, told reporters Tuesday, urging policymakers to restore fiscal space and supply-chain security.

The IMF now estimates that the Strait of Hormuz will reopen in March next year and return to pre-war state by October. Gulf crude exports averaged 88% of pre-conflict levels in September as traffic picked up, the international body said.

The report was released at the IMF and World Bank annual meetings in Bangkok this week, where finance ministers and central bank governors from the fund’s 191 members are meeting to discuss geopolitical tensions, mounting public debt and artificial intelligence.

AI-related investment gave the global economy a “demand impulse” that helped activity in the first half of the year hold up better than expected, despite the Middle East war and trade tensions, the IMF said. Financial markets also stayed largely calm, which the fund said may partly reflect “AI-related exuberance.”

For now, “use of buffers and adaptability have helped economies weather the negative shocks,” the IMF said, citing measures such as diversifying fuel sources, tapping strategic reserves that have contained the energy-market turmoil unleashed by the Middle East war.

But that cushion is “rapidly narrowing” in the near term, it said. Global oil reserves are being drawn down by about 2.7 million barrels a day, nearing levels that could trigger sharp price moves, and energy constraints could impair the AI-led momentum. The growing unease over governments’ fiscal health also adds to the headwinds.

Tenreyro said monetary policy should remain “sufficiently restrictive” to return inflation to target in a timely way, and that central bank independence is critical as governments face mounting financing pressure.

Skewed growth

Overall growth is also becoming increasingly uneven. Thirty-two economies now account for 95% of global growth, down from 54 in April.

Countries that have limited exposure to the war-led supply shock and are catering to the AI-led demand are holding up or strengthening. Taiwan is forecast to grow 10.7% this year and South Korea 3.3%, up from 1.1% in 2025 on semiconductor demand. The U.S. is projected to expand 2.3% in both 2026 and 2027, supported by private consumption, fiscal policy and tech investment.

Energy importers outside that tech supply chain face higher import bills, weaker real incomes and less fiscal room. The euro zone is estimated to eke out just 1.1% growth this year amid higher energy costs and weak consumer confidence. The Middle East and Central Asia is projected to contract 0.3% this year, with Bahrain, Iraq, Kuwait and Qatar facing double-digit declines before double-digit rebounds in 2027.

China is forecast to grow 4.5% this year and 4.0% in 2027, with higher oil prices and structural headwinds weighing on activity. Second-quarter growth of 4.3% was weaker than expected as sluggish consumption and property strains outweighed robust exports.

India remains the fastest-growing major economy, with the IMF forecasting growth of 7.0% in both 2026 and 2027, upgrades of 0.6 percentage point and 0.3 percentage point from July, respectively, driven by robust domestic demand, services activity and AI-related investment.

“The global economy is not simply slowing or accelerating; rather, growth is being reallocated across countries according to their exposure to the war-induced supply shock and AI-related demand strength,” the IMF said.

AI-related risks

While AI has underpinned resilience, the investment boom is also adding to inflationary pressures and financial fragility, the IMF said.

“This is a large, front-loaded demand shock” that sees massive investment pouring into sectors such as semiconductors, data centers, and power generation, before any broad productivity gains have materialized, it said in the report.

A potential reassessment of AI-related profitability and productivity prospects could see excessive investment pull back and force a sharp correction in equities. The IMF warned that in a scenario where AI productivity gains fail to materialize, U.S. equities could fall 20% and global growth could end up 1.1% below its baseline estimates by 2028, with Asia 1.7% lower.

Fiscal constraints

Advanced-economy benchmark yields have climbed to multidecade highs as resurgent inflation, the expected path of policy rates and fiscal concerns push up borrowing costs, the IMF said. Technology firms borrowing to finance AI investment are adding to the pressure.

Public debt in advanced economies is projected to reach 115% of GDP by 2031, from 108% in 2025. In emerging and developing economies, it is forecast to climb to 84% from 74%. Primary deficits in the largest economies remain substantially larger than before the pandemic, and governments face higher real interest rates and shorter debt maturities as defense spending pressures mount.

Rebuilding fiscal space is “increasingly urgent,” it said, urging governments to keep support for households facing higher energy and food costs “temporary, targeted, and nondistortionary.”

Its models put the probability that global inflation exceeds 5% in 2027 at about 35%, up from 30% in April. Risks remain “tilted to the downside,” the fund said.

Anthropic said Monday that small AI models are “months away” from developing the capabilities needed to replicate themselves across internet-connected devices in the home, marking the latest warning from the frontier lab about the technology’s potential safety risks.

Anthropic described how AI models could potentially infiltrate internet-connected devices, such as smart ovens or Android TV boxes, and use their computing power to run themselves and spread to other devices, without relying on expensive data centers.

The AI model would be “challenging to shut down,” Anthropic said in a blog post, “since there’s no centralized point to shut off.”

The company acknowledged that there is not yet a “fully rogue replicator.”

Agents are, however, “becoming increasingly capable of completing the individual steps in the replication chain (and passing on their adaptations to future copies of themselves),” Anthropic said.

Anthropic’s latest research adds to growing safety concerns around AI, such as the ability for models to carry out cyberattacks and the technology going rogue. In September, improvements in frontier models sparked industry-wide warnings of existential risks to humanity.

Executives and researchers at Anthropic and rival OpenAI repeatedly warned about the potential risks of increasingly powerful AI, with some calling for tighter guardrails and measures to control the pace of the technology’s development.

Anthropic said it carried out various tests to see if AI could self-replicate. One of those tests involved running a small Google model, Gemma4-26B-A4B, across a network of 15 cheap off-brand Android TV boxes.

The company noted that these types of devices often have weak security, which could allow a model to infiltrate them. A model that is too large to fit on one device could “split into pieces, or shards” and run across multiple devices, Anthropic said.

“In practice, an AI worm wouldn’t necessarily consist of just one model or strategy. A swarm of different models deployed on different types of hardware using a variety of replication schemes might represent a more realistic ‘worst case scenario’ attack,” Anthropic said.

It’s still “uncertain exactly what form autonomous AI replication will take,” the AI lab said, warning that it could be a “slow, distributed” system sitting on insecure devices or a “powerful cyber model that breaches datacenters at a high pace, or a model that scams people for compute credits.”

“As a prediction that goes beyond our experiments, we expect a future awash with agents running on all types of hardware, including some that aren’t under human control,” Anthropic said.

“The rise of slow replicators would also be a bellwether for more aggressive, faster ones that can breach more hardened machines or be effective API key thieves,” the company added.

BANGKOK – The International Monetary Fund warned Tuesday that government bond yields could rise in a disorderly fashion, as rising inflation, mounting debt and AI-related borrowings pressure sovereign bond markets.

Government bond yields in several advanced economies have climbed to multidecade highs, driven by higher energy costs due to the Middle East conflict and expectations for tighter monetary conditions as central banks look to curb resurgent inflation.

In its latest Global Financial Stability Report, the fund warned that sovereign bond risk could deteriorate, warning of “a sharp rise in sovereign bond yields.” The report was released Tuesday as part of the IMF-World Bank annual meetings in Bangkok, Thailand, this week, where more than 18,000 delegates gathered to discuss the state of the world economy and the financial system.

High debt load and rollover risks make the bond market “susceptible to more frequent fiscal and inflation shocks,” the report said. The IMF had warned last week that global public debt had reached near its highest level since World War II and was on track to soon exceed 100% of GDP.

“Impact could be severe given the foundational importance of sovereign bond markets,” the IMF said Tuesday. A sharp rise in yields would mark down asset values through higher discount rates, including for AI-related assets, and erode the value of government bonds posted as collateral.

IMF also noted the exponentially rising borrowing by AI firms — that compete with sovereign bonds for investor demand — as an additional source of strain on bond markets. Hyperscalers’ share of global investment-grade bond issuance has risen to 10% from less than 1% since 2024, the report said.

The global AI buildout has seen tech giants, such as Alphabet Inc’s Google, Meta Platforms, hyperscalers and AI labs raising hundreds of billions of dollars in debt this year.

“AI-linked issuance – much of it longer-dated – has grown large enough to compete with government bonds for investor capital, even if it does not appear to be the dominant driver of yields,” according to Marc Seidner, CIO of non-traditional strategies at Pacific Investment Management.

Hedge fund worries

Hedge funds’ growing holdings of Treasuries leave the market “susceptible to a disorderly unwind,” the IMF said. Leveraged basis trades have partly unwound without notable pressure in the short-term loan market but the leverage that remains is a concern.

Hedge funds’ share of U.S. Treasuries rose to about 9% in 2025 from about 4% in 2022, mostly through highly arbitraged strategies, the IMF said, citing data from the Federal Reserve.

Caio Ferreira, IMF deputy division chief for financial supervision and regulation, said the worry stems from hedge funds’ rising leverage and their increasingly overlapping positions, meaning multiple funds are holding the same trades, potentially amplifying impact from a collective unwinding.

On the bright side, liquidity has held up in the U.S., Europe and the U.K., the IMF said. Japan is the exception: liquidity in its government bond market has “continued to deteriorate” as yield volatility has risen since the end of yield curve control.

The fund urged debt managers to lengthen maturities, warning that rollover risk could be exacerbated by strategies that replace long-term debt with short-term borrowing on a large scale.

‘Circular’ AI financing

The IMF sounded the alarm over a web of “circular” financing deals among AI firms.

Since 2023, 14 publicly disclosed deals channeled about $300 billion of equity, led by hyperscalers, into financially weaker AI firms, alongside $650 billion of debt. In exchange, the investors secured $1.4 trillion of “future revenue commitments” through services or chip orders, the IMF said.

The revenue promises are what set the deals apart, Jason Wu, assistant director of the IMF monetary and capital markets department, said at a press briefing on Tuesday. “What makes this a little bit unusual compared to previous deals is the future revenue commitments, hence the circularity.”

If AI commercialization falls short, the losses could arrive together. “If there are questions about the monetization in AI, the equity investments will deplete in value, the debt may be difficult to service, and the revenue commitments in the future may not materialize,” Wu said, describing what he called a “triple whammy.”

Such heavy financing and revenue promises cluster in a handful of AI firms, the report said. Due to that concentration, “there are certain companies and nodes that are particularly exposed,” Wu said. “If there are some kind of issues with those entities, then the circularity would pose risk to the market.”