China’s manufacturing sector showed a widening split in February, with large state-owned factories struggling to regain momentum while private exporters reported a surge in overseas orders.

The February report has prompted concern that escalating conflict in the Middle East could sharply undermine export performance in March.

Data from the National Bureau of Statistics of China showed the official manufacturing purchasing managers’ index (PMI) slipping to 49.0 in February from 49.3 in January, a four-month low. The reading remained below the 50-point mark that separates expansion from contraction and came in slightly below the 49.1 median forecast in a Reuters poll. The sub-50 reading signals persistent weakness in large and medium-sized enterprises, many of which are state-owned and more dependent on domestic demand and infrastructure-linked activity.

In contrast, the Caixin China General Manufacturing PMI, compiled by S&P Global, rose to 52.1 from 50.3, beating forecasts of 50.2. The private-sector survey showed new orders increasing for the ninth consecutive month, at the fastest pace since December 2020. Export orders in the Caixin survey expanded at their strongest rate since September 2020, highlighting the continued resilience of smaller, export-focused manufacturers.

The divergence underscores structural fault lines within China’s economy. The official survey skews toward large industrial firms serving domestic markets, including construction, heavy equipment, and property-related supply chains — sectors still grappling with subdued demand and overcapacity. The Caixin survey captures more nimble firms, often embedded in global supply chains and clustered around coastal manufacturing hubs.

Zichun Huang, a Chinese economist at Capital Economics, said averaging the two surveys provides a broader gauge of industrial conditions.

“On this basis, the headline reading picked up from 49.8 to a five-month high of 50.5,” Huang said. That composite view suggests stabilization at the margin, though the underlying composition of growth remains uneven and externally dependent.

Exports have been a critical support for the economy. In 2025, China posted a record $1.2 trillion trade surplus, navigating tariff pressure from U.S. President Donald Trump through supply chain adjustments and strong demand from emerging markets. But that cushion now faces mounting risks from escalating tensions in the Middle East.

The closure of the Strait of Hormuz following U.S.-Israeli military action against Iran has introduced a new layer of uncertainty. The waterway is a vital artery for global energy shipments and broader trade flows. Disruptions there have already led to higher shipping insurance premiums and delays in container traffic, according to analysts, threatening to raise logistics costs for Chinese exporters.

“If the war and closure of the Strait of Hormuz lasts for three or four months, or even longer, then it will be a nightmare for every country, including China,” said Xu Tianchen, senior economist at the Economist Intelligence Unit.

He noted early instances of shipment disruptions and higher freight charges.

While February’s private-sector export surge suggests orders were still flowing before the full impact of the Middle East escalation was felt, economists expect March data to tell a different story. Contracts negotiated earlier in the quarter may have buoyed February readings, but rising freight costs, longer transit times, and greater geopolitical uncertainty are likely to weigh on new orders and shipment volumes in the coming weeks.

The official PMI’s export sub-index already hints at softening demand. New export orders in that survey fell to 45.0 in February from 47.8, a 10-month low, signaling contraction. That drop may reflect early caution among larger manufacturers more exposed to commodity-intensive supply chains and shipping routes vulnerable to disruption.

Zhiwei Zhang, chief economist at Pinpoint Asset Management, warned that the Middle East conflict would likely weigh on the global economy, including China, at least through March.

“If economic activity slows further in the coming months, I expect the government to boost investment moderately to mitigate pressure on the economy,” Zhang said, pointing to signals expected from the National People’s Congress.

The March performance of China’s export sector will be closely watched as a barometer of how quickly geopolitical tensions translate into real economic impact. China’s manufacturing model is deeply integrated into global supply chains, particularly in electronics, machinery, and intermediate goods. Any sustained disruption in maritime trade routes could not only delay outbound shipments but also impede the import of key components and energy supplies, compounding production bottlenecks.

Domestic demand remains a weak link. Household consumption has yet to fully recover from the property downturn and pandemic-era shocks, while investment in real estate remains subdued. Overcapacity in sectors such as steel, chemicals, and solar manufacturing continues to pressure margins. That makes the economy more reliant on external demand at a time when global trade conditions are deteriorating.

Premier Li Qiang is expected to outline new measures to boost domestic demand and address structural imbalances when he presents the government’s work report and the next five-year plan. Economists are looking for concrete fiscal initiatives after targeted interest rate cuts and potential reserve requirement reductions delivered only limited traction.

Lynn Song, chief economist for Greater China at ING, said the mixed PMI readings suggest a continuation of last year’s pattern.

“The mixed bag of manufacturing PMI data suggests a similar trajectory to what we observed in 2025: resilient external demand continuing to drive growth,” Song said.

However, he added that domestic demand has remained soft and is unlikely to rebound meaningfully without stronger policy support.

Economists polled by Reuters in January forecast growth slowing to 4.5% this year and holding at that pace in 2027. If the Middle East conflict persists and trade routes remain disrupted into the second quarter, those projections could come under downward pressure.

The February data present a snapshot of an economy still benefiting from external demand momentum built earlier in the year. But with geopolitical risks intensifying and shipping channels under strain, March is shaping up as a critical test of how China’s export engine can withstand another external shock.

OpenAI is developing an internal code-hosting platform that could eventually compete with GitHub, according to a report by The Information, marking what could become one of the most consequential strategic pivots in the artificial intelligence industry.

The project, still in its early stages, was prompted in part by repeated service disruptions that left GitHub temporarily unavailable in recent months, the report said, citing a person familiar with the matter. Those outages reportedly exposed operational risks for OpenAI engineers who rely heavily on external infrastructure to manage and deploy code at scale.

GitHub is owned by Microsoft, which is also OpenAI’s largest strategic backer. Any move by OpenAI to commercialize its own repository product would therefore place the AI company in direct competition with a key investor and infrastructure partner.

At a technical level, the rationale appears rooted in reliability and control. Code-hosting platforms are central to modern software development workflows, handling version control, collaboration, deployment pipelines, and integrations with testing and security tools. Service interruptions in such systems can delay releases, disrupt model training pipelines, and slow product iteration cycles.

For an organization operating frontier AI systems — including ChatGPT — uptime and development velocity are strategic imperatives. Building an in-house platform would allow OpenAI to reduce dependency on third-party providers, tighten security around proprietary model code, and potentially tailor repository features to AI-native workflows such as large-model versioning, dataset governance, and inference deployment.

The Information reported that employees have considered making the repository available for purchase to OpenAI’s broader customer base. If that path is taken, the initiative would shift from internal resilience to revenue diversification.

The competitive optics come with an enormous impact due to the relationship between the duo. Microsoft not only owns GitHub but also provides OpenAI with cloud infrastructure through Azure and holds a substantial equity stake in the AI firm. A commercial code-hosting platform from OpenAI would introduce a rare point of competitive overlap between the two companies.

So far, the partnership has been defined by deep integration: Microsoft has embedded OpenAI’s models across its enterprise products, while OpenAI has leaned on Microsoft’s cloud capacity to scale training and inference. A direct challenge to GitHub would test the boundaries of that alliance.

The development also comes amid heightened scrutiny over competitive dynamics in AI, as regulators in the U.S. and Europe assess whether close partnerships between dominant cloud providers and AI startups concentrate market power.

Valuation signals and AI capital race

The move surfaces against the backdrop of OpenAI’s latest funding round, which reportedly valued the company at $840 billion. The raise drew participation from major technology firms and investors, including Masayoshi Son and his conglomerate SoftBank, reinforcing the scale of capital flowing into artificial intelligence infrastructure.

The size of the valuation signals that investors continue to price in aggressive expansion beyond core chatbot and API services. A code-hosting platform, particularly one optimized for AI development, could serve as both a defensive tool and a new monetization layer within OpenAI’s ecosystem.

GitHub, with its vast developer base and embedded enterprise presence, remains deeply entrenched in global software workflows. Any credible alternative would need to offer differentiated functionality — potentially AI-native features such as automated code review powered by large language models, model lifecycle tracking, or integrated prompt engineering environments.

Execution risks and timeline

According to the report, the project is unlikely to be completed for months. Building a secure, scalable, enterprise-grade repository platform is a complex undertaking, requiring robust version control systems, authentication layers, compliance tooling, and integration ecosystems.

Moreover, monetizing such a product would require navigating customer perceptions about vendor neutrality, especially among enterprises that use both Microsoft and OpenAI services.

At this stage, OpenAI’s initiative appears exploratory. But if it advances to commercialization, it would represent a broader shift toward vertical integration — consolidating infrastructure, development tooling, and AI services under one corporate umbrella.

That evolution would mark a wider industry trend: AI firms are no longer content to build models alone. They are increasingly positioning themselves as full-stack technology providers, competing not just on algorithms, but on the platforms that power the software economy itself.