AI productivity upside could be 10x the current estimates: BofA

Bank of America says AI could lift global economic growth by up to 1 percentage point annually over the next decade, from about 3.5% to 4.5%, with task-level productivity gains already visible in software development (+55%) and writing (+40%). However, broad economy-wide productivity is still rising only about 0.1% per year because adoption, retraining, and integration remain barriers. The report also argues that faster AI adoption could raise neutral interest rates in leading economies as productivity and investment demand strengthen.

Analysis

The market is still pricing AI as a capex story, but the more important second-order effect is a re-rating of labor-adjusted growth expectations and, eventually, the neutral rate. If productivity diffusion keeps broadening, the winners are not just the model vendors but the firms that can turn workflow automation into sustained margin expansion: platform software, cloud infra, semis, and select consulting/integration names. That argues for relative strength in the AI enablers while traditional labor-intensive service firms face a slower, more uneven margin tailwind.

The key distinction is between task-level gains and enterprise-level monetization. The adoption curve is high enough to support multiple expansion in the near term, but the macro payoff likely arrives with a lag of several quarters to years as retraining, workflow redesign, and regulatory clearance catch up. That creates a window where AI beneficiaries can compound revenue growth before the market fully capitalizes the productivity upside into rate expectations.

For rates, the contrarian implication is that AI may be less deflationary than the consensus expects in the medium term because higher productivity can coexist with stronger investment demand. That is constructive for real yields and bearish for duration-sensitive assets if the market starts to reprice a higher terminal neutral rate in the US versus Europe. Regional divergence matters: faster adopters should see relative growth and rate premia widen, while slower adopters risk chronic discount rates and lower earnings leverage.

The main risk is that adoption statistics remain headline-high but economically shallow, leading to a sentiment air pocket if monetization disappoints. Another reversal trigger is faster-than-expected supply response in AI infrastructure, which could compress returns on the picks-and-shovels trade even as usage grows. In that scenario, the best risk/reward likely shifts from pure beta to cash-generative incumbents with real distribution and pricing power.

Key Decisions for Investors

• Long MSFT versus a broad software basket for 3-6 months: best positioned to convert AI adoption into durable operating leverage and attach higher-value workflow spend; target 1.5-2.0x the upside of a generic SaaS rally with lower drawdown risk.
• Pair trade long SMCI / short a non-AI hardware index ETF for 1-3 months: if AI buildout remains capital-intensive, the levered infrastructure names should outperform, but size modestly because the trade is vulnerable to supply-chain normalization and multiple compression.
• Initiate a duration hedge via short TLT or buy put spreads on long-duration Treasuries over the next 2-4 months: rising neutral-rate expectations from stronger productivity and investment demand create asymmetric downside for bonds if growth data starts confirming the AI narrative.
• Use BAC as a relative beneficiary basket proxy, long BAC against a regional-bank ETF over 6-12 months: a higher neutral-rate regime and stronger nominal growth should favor balance-sheet scale and trading/markets sensitivity over rate-sensitive smaller lenders.
• Buy APP on pullbacks as a higher-beta AI-adjacent monetization expression for 3-6 months, but pair with a stop: upside is attractive if AI spend continues flowing into digital ad/automation tooling, yet the stock will be more vulnerable if the market rotates from narrative to fundamentals.