Industry signals
What changed across the profession.
China Moves to Slow the Yuan’s Rise Before It Erodes Export Competitiveness
The People’s Bank of China will raise the reserve requirement on banks’ foreign-currency deposits from 4% to 6% on November 15, its first increase in nearly five years. The yuan appreciated 4.2% against the U.S. dollar in the first nine months of 2026, and the measure is expected to immobilise roughly US$20–25 billion of foreign-exchange liquidity. Beijing is signalling that it will resist a rapid loss of price competitiveness without setting an exchange-rate target. What to do: exporters, investors and promotion agencies should rerun China-versus-alternative-location cost models at several currency levels; this tool can slow appreciation pressure, but it does not remove wage, tariff or market-access risks.
State Council of China: foreign-exchange reserve requirementOECD Finds IMEC Has Ports but Not Yet a Continuous Corridor
The OECD’s first connectivity assessment of the India–Middle East–Europe Corridor finds strong maritime gateways at Mundra, Khalifa and Piraeus but no committed cross-border rail connection between Jordan and Israel—the central missing link in the inland chain. Western Indian ports are nearing capacity, Eastern Mediterranean spare capacity may constrain future volumes, and fragmented customs systems still undermine reliability. The near-term opportunity is therefore operational, not ceremonial: link national single windows, share inspection results and align trusted-trader systems before the railway is complete. What to watch: which projects receive the new IMEC label and whether the steering committee funds port capacity, border interoperability and the missing rail connection as one network.
OECD: Measuring Connectivity across IMECUNCTAD Says Developing Economies Supply 42% of a US$484 Billion Substitutes Market—but Capture Too Little Processing
Global exports of non-plastic substitutes reached US$484 billion in 2024, with developing economies increasing their share from roughly one quarter in 2005 to 42%, according to UNCTAD. Yet processed products represented only 35% of trade, while raw materials accounted for nearly two thirds. The opportunity is not simply to promote bamboo, fibres, paper, glass or minerals; it is to locate converting, testing and manufacturing capacity where the inputs originate. What to do: export and investment agencies should map value chains by processing stage, standards and buyer requirements, then target technology and capital that turn resource exports into certified packaging, textiles and industrial inputs.
UNCTAD: trade in non-plastic substitutesOECD Turns Heat and Drought Into a Location-Risk Metric
The number of above-average-temperature days worldwide has risen 126% since the early 1980s, and more than tripled in Latin America and the Caribbean, Sub-Saharan Africa, and the Middle East and North Africa. OECD populations experienced an average 17 additional heat-stress days a year in 2021–25, while cropland soil moisture across OECD and partner economies fell 2.6% against the 1981–2010 baseline. These are operating-cost variables for food, power, logistics and labour-intensive projects. What to do: location propositions should disclose heat exposure, water availability, grid resilience and adaptation plans alongside taxes and incentives; investors will price the omitted risk themselves.
OECD: Climate Action Monitor 2026OECD Finds AI Adoption Is Becoming a Firm-Size Competitiveness Divide
Infrastructure attraction is no longer enough: large firms are adopting AI at more than twice the economy-wide rate, and the gap is projected to persist without targeted support.
Only 20% of OECD enterprises with at least ten employees used artificial intelligence in 2025, compared with 52% of firms employing 250 people or more. The OECD’s new Digital Economy and Society Outlook covers policy initiatives across 42 countries and finds that, without targeted support, the gap will persist: by 2030, AI use among large firms could be nearly 80% higher than the average across all firms.
That changes the economic-development question. Governments have spent the first phase of the AI investment cycle competing for data centres, chips and research labs. Those assets still matter. Venture-capital investment in AI infrastructure and hosting firms more than doubled from US$54.5 billion in 2024 to US$123.9 billion in 2025. But the report shows that infrastructure attraction and economy-wide adoption are different contests. A jurisdiction can win compute capacity while its domestic suppliers, exporters and smaller employers fail to use it.
The immediate winners are large, knowledge-intensive firms able to pay for compute, specialised staff, data preparation and systems integration. They can spread fixed implementation costs across larger operations and absorb legal or technical uncertainty. Smaller companies face limited finance, skills and infrastructure, as well as uncertainty over data protection and the cost of connecting AI to existing systems. If that gap holds, productivity and procurement advantages will compound in the firms already best placed to invest.
For national investment agencies, the finding should alter both targeting and aftercare. An AI proposition cannot stop at available power, land and tax support for a data-centre operator. Agencies should ask what domestic firms can buy from, sell to and learn from the project; whether local cloud and compute access is affordable; and which supplier cohorts can move from experimentation to deployed systems. Incentives tied only to capital expenditure may create infrastructure without broad productive use.
Trade agencies have a parallel problem. Export readiness increasingly includes the ability to automate compliance, product development, customer support and supply-chain decisions. A small exporter without secure data, appropriate cloud services or implementation skills will compete against larger rivals using all three. Generic AI workshops will not close that gap. Agencies need sector-specific adoption programmes with measurable deployment, cost and export outcomes, supported by vetted technical partners and clear data rules.
The public-sector response is already becoming more specialised. The OECD identified 1,497 digital-policy initiatives across 42 countries in 2025 and found that governments are moving from technology-neutral policy toward technology-specific measures. AI was the most frequently cited technology, while 60% of countries with a national digital strategy assigned primary responsibility to a dedicated digital ministry, up from 24% in 2016. The institutional shift is real; its test is whether programmes change firm behaviour.
Agencies should therefore measure the adoption funnel as rigorously as the investment funnel: firms assessed, pilots started, systems deployed, workers trained, costs reduced and export or productivity gains achieved. They should segment support by firm size and sector rather than subsidising undifferentiated access. They should also publish the conditions investors need—power, cloud, data governance, skills and customer demand—as a system, not a list of assets.
What to watch: whether governments attach adoption and supplier-development conditions to AI-infrastructure incentives, and whether small-firm usage begins closing the 32-point gap with large enterprises. The next competitive divide will not be between countries with and without an AI strategy. It will be between those that turn infrastructure into broad firm-level capability and those that merely host it.
Practice implications
- ◆Measure firms assessed, pilots started, systems deployed and productivity or export gains—not workshop attendance.
- ◆Tie AI-infrastructure attraction to affordable compute access, supplier development and sector-specific adoption.
- ◆Segment support by firm size and sector because the barriers are finance, skills, data and integration—not awareness alone.
Sources
Get the Brief in your inbox
Each issue is free and arrives the day it publishes.