但需求面在短期更复杂。AI 相关超大规模云端业者的资本支出正以爆炸速度增长，主要由未来生产力提升的预期驱动，而且已开始超过营运现金流；若这种由预期带动的需求快于实际生产力改善，短期通膨就可能上升。Lenzu 也提醒，现有货币政策工具可能难以同时处理 AI 带来的通膨与金融稳定风险，尤其在采用摩擦、重组成本或资源错置下，AI 可能造成类似停滞性通膨的过渡期。
欧元收敛交易也重新受到关注。EU 仍有 6 个成员国尚未加入欧元区，其中 Hungary 最受瞩目；新总理 Péter Magyar 已把 2030 年采用欧元列为亲近 EU 的核心承诺。自 3 月底以来，Hungarian 10-year bond yields 已下跌 1.75 个百分点，forint 对 euro 升值逾 6%，而且 Hungarian 殖利率已低于 Poland；但 Hungary 仍面临高达 GDP 的 7% 财政赤字、78% 债务占 GDP，比欧元收敛标准的 3% 与 60% 上限高出许多。
Kevin Warsh has argued that AI will create a productivity boom that lets the Fed cut rates, but most economists doubt it will deliver the broad disinflation he expects. New York Fed researcher Simone Lenzu says the effects run both ways: on the supply side, AI should be disinflationary over the long run because firms can react faster to cyclical shocks, adjust real wages and input prices more quickly, and use better forecasting and inventory management to lower marginal costs.
The demand side is harder. AI hyperscalers’ capex is growing explosively, driven mainly by expectations of future productivity gains and already beginning to exceed operating cash flows. If expectation-led demand rises faster than productivity gains, short-run inflation could increase. Lenzu also warns that existing monetary policy tools may be poorly suited to AI-related inflation and financial-stability risks, especially if adoption frictions, reorganisation costs, or misallocation create a stagflation-like transition.
Euro convergence is also back in focus. Six EU members still have not joined the euro area, and Hungary is the most notable holdout; new prime minister Péter Magyar has made euro adoption by 2030 a central pledge. Since late March, Hungarian 10-year bond yields have fallen 1.75 percentage points and the forint has gained more than 6% against the euro, with yields now below Poland’s. But Hungary still has a fiscal deficit of as much as 7% of GDP and debt of 78% of GDP, well above the euro convergence limits of 3% and 60% respectively.