The AI supercycle strengthens the case for tighter policy by some Asian central banks

Overview

This report discusses the impact of the AI boom on the policy cycle of Asian central banks. It notes that Asian central banks were already facing challenges such as Middle East geopolitical tensions, energy and food prices, El Niño, and uncertainty over Federal Reserve policy, and that AI infrastructure buildout has added new growth and inflation variables. In the short term, AI supports growth through hardware exports and data center investment; in the medium to long term, it may raise potential growth through productivity improvements. But if AI data center construction acts as a negative supply shock, it will reinforce the tightening bias of some central banks.

Core views

The core view is that AI’s impact on Asian economies is not simply growth-positive; it will also reshape the inflation path. Taiwan, China is a key node in the global AI hardware ecosystem, South Korea benefits from the memory-chip supercycle, Japan holds an important position in semiconductor equipment, chemicals, and materials, Mainland China is accelerating semiconductor self-sufficiency, and ASEAN and Australia benefit from data center investment. On labor, AI’s second-order effects are currently mild, but Mainland China could become an early sample for observing AI-driven labor substitution in Asia. On inflation, AI cost pressures first appear in upstream producer economies and are then gradually transmitted through midstream assembly and downstream computing-service prices, making South Korea and Taiwan, China the most likely to face tightening pressure earlier.

Analysis framework

The report analyzes AI’s impact on Asian central banks through three macro transmission channels: growth, labor, and inflation. The growth analysis focuses on AI hardware exports, data center capex, and long-term productivity; the labor analysis uses explicit and implicit signals of AI labor impact from corporate earnings call transcripts; the inflation analysis classifies Asian economies by their position in the AI technology stack as upstream producers, midstream assemblers, and downstream adopters, and examines producer prices, export prices, PMI input and output prices, and lagged CPI transmission.

Methodology notes

• Macro transmission frameworkThree-dimensional analysis of growth, labor, and inflation

  Decomposing AI’s macro impact on central bank policy

  The report separately evaluates the growth upside from AI hardware exports and data center investment, the second-order effects of AI adoption on employment and productivity, and the transmission of AI supply-chain costs to inflation and policy rates.

• Industry-chain position frameworkUpstream producers, midstream assemblers, downstream adopters

  Layered AI inflation pass-through

  Upstream economies such as South Korea, Taiwan, China, and Japan bear chip, equipment, and talent cost pressures earlier; Mainland China’s midstream assembly segment absorbs part of the shock through scale, competition, and efficiency; downstream adopters such as India, Indonesia, and parts of ASEAN face rising cloud-computing, hardware, and service costs with a longer lag.

• Text signal analysisAnalysis of AI labor impact in corporate earnings calls

  Explicit and implicit labor-substitution signals

  HSBC’s data science team compares signals such as AI adoption, layoffs, substitution, and hiring slowdowns mentioned in earnings calls by companies in Asian emerging markets and developed markets to assess potential labor-market disruption.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• South Korea

  An upstream beneficiary of AI memory chips and also an economy where inflation pressure is transmitted earlier

  Strengths

  Samsung and SK Hynix hold leading global positions in DRAM and NAND, and HBM demand supports a supercycle.

  Weaknesses

  Material shortages, delivery delays, and capacity shifting toward higher-margin AI products may raise production costs.

  Comparison

  Compared with downstream adopter economies, South Korea benefits earlier from AI demand but also bears upstream costs and PPI pressure earlier.

  Risks

  Cost pass-through to CPI, wage spillovers, and further central bank tightening.

• Taiwan, China

  The core node of the global AI hardware ecosystem

  Strengths

  TSMC holds a central position in advanced process technology and global foundry manufacturing, and AI hardware exports contribute significantly to economic growth.

  Weaknesses

  Supply-chain tightness, inventory stockpiling, and surging semiconductor demand may increase cost pressure.

  Comparison

  Compared with Japan, AI hardware accounts for a larger share of Taiwan, China’s economy, and CPI pass-through may be faster.

  Risks

  Rising producer prices, lagged consumer inflation, and tighter monetary policy.

• Japan

  A supplier of semiconductor equipment, chemicals, and materials

  Strengths

  It holds an important position in chip manufacturing equipment, materials, and investment in critical technologies, and also benefits from data center investment.

  Weaknesses

  AI hardware accounts for a smaller share of the overall economy than in South Korea and Taiwan, China, and companies are more inclined to absorb costs.

  Comparison

  Japan’s transmission from PPI shock to CPI may be slower than in South Korea and Taiwan, China.

  Risks

  Pressure on corporate profit margins, policy investment execution risk, and weakening cost-absorption capacity.

• Mainland China

  A midstream AI assembly and supply-chain localization economy

  Strengths

  Semiconductor self-sufficiency is advancing, data center investment is large-scale, and the assembly segment has strong scale and competitive advantages.

  Weaknesses

  Upstream imported chip inflation combined with export prices constrained by a domestic non-inflationary environment may compress margins.

  Comparison

  Compared with upstream supplier countries, Mainland China acts more like a disinflationary buffer layer that absorbs cost shocks.

  Risks

  Margin compression, AI labor substitution amid youth employment pressure, and technology restrictions.

• India and the Philippines BPO industry

  An early-observation sector for AI labor substitution

  Strengths

  Routine processing jobs can be shifted toward higher value-added activities through retraining.

  Weaknesses

  Repetitive tasks such as data entry are more easily displaced by automation.

  Comparison

  Compared with upstream hardware economies, the BPO industry’s main risk is labor reallocation rather than the direct hardware export cycle.

  Risks

  Job substitution, slower hiring, and insufficient retraining.

Key data

• AI hardware export exposureTaiwan, China and Vietnam are close to half of GDPThe report says AI hardware exports are one of Asia’s most direct growth channels.
• AI capital expenditure by the four largest U.S. hyperscale cloud providersRising from about USD 150 billion in 2022 to an estimated more than USD 700 billion in 2026This capex drives full-stack data center buildout from chips to servers, fiber networks, and liquid-cooling systems.
• TSMC’s position in global foundry manufacturingAbout 70% market shareThe report emphasizes TSMC’s central role in advanced process technology and the AI hardware value chain.
• South Korea’s position in memory chipsSamsung and SK Hynix account for about 70% of the global DRAM market and more than half of NAND FlashSouth Korea is a major beneficiary of the AI-related memory-chip supercycle.
• Mainland China data center investment planAbout RMB 2 trillion over the next five years, equivalent to about USD 295 billionThe goal is for at least 80% of AI chips to come from domestic suppliers.
• Japan strategic investment planAbout JPY 370 trillion over 14 years, equivalent to about USD 2.3 trillionInvestment priorities include AI, semiconductors, and other critical technologies.

Impact & implications

The investment and policy implication is that the impact of the AI theme on Asia must be assessed from both the growth benefit and the inflation cost. Upstream South Korea, Taiwan, China, and parts of Japan’s industry chain may enjoy export and capex strength while also facing earlier pressure in PPI, export prices, and wage costs; midstream Mainland China provides some disinflationary buffer to global pricing; cost pressures in downstream adopter economies such as India, Indonesia, and ASEAN may show up with a lag through cloud services, corporate IT budgets, and service prices. For central banks, if AI lifts productivity without generating broad inflation, it can support stable rates; if it behaves more like a negative supply-chain shock, it will strengthen the case for rate hikes or maintaining high interest rates.

Risks

• AI upstream cost pressures pass through to CPI faster than expected, forcing central banks to become more hawkish.
• Middle East geopolitical tensions, energy prices, refining margins, and El Niño jointly push up food and fuel prices.
• If AI data center construction creates supply-chain bottlenecks, it may resemble a negative supply shock.
• More visible AI labor substitution or hiring slowdowns emerge in economies such as Mainland China, India, and the Philippines.
• Midstream companies absorb costs, causing margin compression and weakening investment and employment.

What to watch

• PPI, export prices, CPI, and wage spillovers in high-tech industries in South Korea and Taiwan, China.
• Supply-demand changes in AI servers, HBM, DRAM, NAND, data center equipment, and liquid-cooling systems.
• Signals of AI adoption, layoffs, substitution, and hiring slowdowns in Mainland Chinese corporate earnings calls.
• Job reallocation, automation, and retraining progress in the BPO industries of India and the Philippines.
• Rate hikes, pauses, or policy-guidance changes by Asian central banks in the second half of 2026.
• Whether cloud computing and corporate IT contract renewal prices are passed through to downstream service prices.