Citi's Global Chief Economist on Global Growth: What It Means for AI's Next Chapter

Every AI conversation eventually becomes an economics conversation. Chips cost money. Data centers cost more. Electricity, land, water, talent, and time all have a price, and that price is set by the wider economy. So when Citi's Global Chief Economist turns to the subject of global growth, the AI industry should listen closely, because the growth outlook is the tide that lifts or grounds every model, product, and platform built on top of it.

The timing matters. As of October 2026, the AI boom has moved well past the novelty phase. It is now a capital-intensive industrial build-out, with enormous sums committed to compute, power, and infrastructure. That build-out only makes sense if the world economy keeps generating demand for what AI produces. Growth is the assumption baked into every one of those bets, and it is exactly the assumption that a global growth outlook puts under the microscope.

Why a Bank Economist's Growth View Matters to the AI Industry

Banks like Citi sit at the intersection of every major economic flow: corporate lending, cross-border payments, trade finance, capital markets, and treasury operations for the world's largest companies. Their chief economists do not just forecast numbers. They translate the movement of money into a story about what businesses and consumers will do next.

That story reaches the AI world through several channels at once:

• The cost of capital. AI infrastructure is financed. When growth expectations shift, so do interest rates and the price of borrowing, and a single percentage point on a multi-billion-dollar data center program changes the math dramatically.
• Enterprise budgets. Software and AI spending is discretionary at the margin. In a confident growth environment, companies invest ahead of demand. In a cautious one, they demand proof of return before signing.
• Consumer demand. Subscription AI products ride on household spending. Weak growth means fewer upgrades, more churn, and harder upsells.
• Trade and supply chains. Advanced chips, networking gear, and rare materials move across borders. Trade flows follow growth, and growth follows trade.

Put simply: a chief economist's growth outlook is not background noise for AI leaders. It is the operating environment.

The Central Tension: AI Capex vs. a Slower-Growing World

Here is the question that defines this moment. The AI industry is spending as if demand will compound for a decade. The global economy, meanwhile, has been growing at a pace that most economists describe as moderate at best, weighed down by aging populations, heavy public debt, fragmented trade, and sluggish productivity in large parts of the world.

Those two facts can coexist, but only for so long. Either AI delivers the productivity gains that lift global growth, or the spending has to be trimmed to match a slower reality. There is no third path.

This is why the global growth debate is really an AI debate in disguise. Every time an economist revises a growth forecast, they are implicitly revising the size of the market that AI companies are building for.

Three Scenarios and What Each Means for AI

Scenario one: steady, modest growth. The most comfortable outcome. Economies expand at a moderate pace, rates drift lower over time, and enterprises keep investing in AI because the returns are visible and the financing is affordable. In this world, AI adoption spreads quietly through existing budgets rather than through giant new line items. Winners are the vendors who make AI cheaper and more reliable per task.

Scenario two: sticky inflation and higher-for-longer rates. Growth holds up, but the cost of money stays elevated. Capital-intensive AI projects get scrutinized harder. Companies with strong cash flow keep building; companies that depend on debt financing slow down. In this world, efficiency becomes the headline feature, smaller models, better utilization, cheaper inference. The AI industry starts competing on cost per useful output rather than raw capability.

Scenario three: a growth slowdown. Demand softens, budgets tighten, and the weakest AI businesses consolidate. But this scenario has a twist. Historically, downturns accelerate automation, because labor costs stay high while technology costs fall. Companies under margin pressure reach for tools that cut headcount costs or boost output per worker. So a slowdown can hurt AI vendors in the short run while strengthening the case for AI in the medium run.

The takeaway for business leaders is that all three scenarios point toward the same strategic response: build for efficiency and prove value fast. The difference between scenarios is the timeline, not the direction.

Productivity Is the Bridge Between AI and Growth

Economists have a long-standing puzzle: measured productivity growth has been weak in most advanced economies for years, even as digital technology spread everywhere. AI is the biggest test of whether that pattern can break.

If AI meaningfully raises output per worker, the global growth picture improves, and the case for further AI investment strengthens in a virtuous loop. If it does not, the industry faces a credibility problem: enormous spending with modest measurable returns.

This is why the most important AI metric of the next few years will not be benchmark scores. It will be measurable productivity gain per dollar spent. Companies that can document cost savings, faster cycle times, or higher output per employee will win budget. Companies that only demonstrate impressive demos will struggle.

For business leaders, this means building measurement into adoption from day one. Pick a small number of workflows, baseline them, deploy AI, and track the delta. The organizations that can tell a credible productivity story will be the ones that keep getting funded when the economy turns cautious.

Practical Implications for Businesses

1. Stress-Test Your AI Business Case Against Two Economies

Do not model your AI investment on a best-case growth outlook alone. Run the same numbers assuming higher financing costs and slower revenue growth. If the project still pays back, it is robust. If it only works in a boom, it is a bet on the economy, not a bet on AI.

2. Prioritize Cost-Reduction Use Cases Now

Revenue-boosting AI is exciting but slower to prove and more exposed to consumer confidence. Cost-reduction AI, automating back-office work, speeding up document processing, improving forecasting, cutting waste in logistics, pays back regardless of the growth environment. It is also the easiest to defend to a finance committee.

3. Treat Compute as a Financial Exposure

Long-term compute contracts are, in effect, fixed-rate bets on future demand. Finance and technology teams should review them together. Build in flexibility where possible, and model what happens if utilization comes in below plan.

4. Watch the Money, Not Just the Models

For anyone tracking the AI industry, the leading indicators are increasingly macroeconomic: credit conditions, corporate capital spending plans, and the language used on earnings calls about payback periods. When bank economists flag slower growth, it shows up in AI order books months later.

5. Design for a Cheaper Unit of Intelligence

Whatever the growth path, the direction of travel in AI is toward lower cost per task. Products and internal platforms should be built so that the underlying model can be swapped for a cheaper one as pricing falls. Locking architectures to a single expensive approach is a macro risk, not just a technical one.

What This Means for Society

Global growth is not an abstraction. It determines whether governments can fund the transition support that workers displaced by automation will need. Slower growth makes retraining budgets tighter just as the need for them rises.

There is also the question of diffusion. In a high-growth, high-rate environment, AI adoption concentrates in large firms with capital to spare. In a slower-growth world with cheaper money, smaller businesses may find adoption more accessible, but only if tools get cheaper and simpler. The policy takeaway is clear: if the goal is broad-based productivity gains, the priority is making AI affordable and usable for the long tail of small and mid-sized companies, not just for the Fortune 500.

The Bottom Line

The global growth outlook and the future of AI are now the same conversation. Capital-intensive AI build-outs depend on a world economy confident enough to absorb what they produce. And the world economy increasingly depends on AI to deliver the productivity gains that break a long stretch of sluggish growth.

That mutual dependence is the defining economic story of this decade. It will not be resolved in a single forecast, a single earnings season, or a single product launch. It will be resolved slowly, through measurable productivity, disciplined capital allocation, and the quiet accumulation of thousands of business decisions about whether AI is worth the money.

For businesses, the playbook is not complicated. Build the case, measure the results, keep the architecture flexible, and plan for more than one economic weather pattern. The companies that do this will not just survive whatever growth environment arrives, they will be the ones whose AI investments still look smart when the tide changes.