AI Capex Contribution to US GDP Growth Declining: BCA Research Analysis

This analysis is based on Marko Papic’s appearance on CNBC’s “The Exchange” on January 7, 2026, where the Chief Strategist of BCA Research disclosed that AI capital expenditure’s contribution to US GDP growth is declining, currently adding approximately 1 percentage point to economic expansion [1]. Papic’s broader thesis警告 that the AI infrastructure boom will ultimately “end in tears,” similar to historical technology infrastructure cycles such as 19th-century railways and late-1990s fiber-optic builds [2]. The analysis suggests international equities will outperform US stocks in 2026, with the US equity market bearing disproportionate risk when the AI capex cycle normalizes [2]. Current market data shows the Technology sector closing marginally up +0.85% on the day, suggesting limited immediate reaction to this specific commentary [0].

BCA Research’s analysis reveals that AI infrastructure investment currently constitutes approximately 1.0 percentage point of US GDP growth during the first half of 2025, representing roughly 50% of total economic expansion at a time when overall GDP growth stands at 1.6% [3]. AI capex as a percentage of US GDP reached 1.1% in 2025, equivalent to $344 billion from hyperscalers, with projections indicating $404 billion in capital spending for 2026—though this represents a deceleration in growth momentum rather than absolute decline [3].

Goldman Sachs Research corroborates this trend, noting that AI capex currently equates to just 0.8% of GDP, significantly below peak levels of 1.5% or greater observed during previous technology investment booms over the past 150 years [4]. This quantitative context is essential for understanding the declining contribution thesis: while AI spending remains substantial, its relative impact on economic growth is diminishing as base effects and normalization take hold.

The declining contribution trajectory mirrors patterns from historical technology infrastructure cycles. BCA Research draws direct parallels to 19th-century railway construction and late-1990s fiber-optic cable deployment, both of which experienced periods of intense capital spending followed by sharp corrections [2][3]. These historical precedents suggest that the current AI infrastructure buildout may be approaching a inflection point where investment growth slows relative to economic absorption.

Market data from the event date reveals nuanced sector performance that partially validates the BCA Research thesis. The Technology sector closed marginally up at +0.85%, while the NASDAQ Composite gained +0.49%, and the Dow Jones Industrial Average declined -0.69% [0]. This differential suggests sector rotation away from value and into growth positions, though the limited magnitude indicates markets may be processing the nuanced message that while AI capex growth is slowing, it hasn’t collapsed.

Communication Services (-0.45%) slightly underperformed, potentially reflecting concerns about AI-dependent platforms and their ability to monetize infrastructure investments [0]. More significantly, Real Estate (-1.40%) and Utilities (-3.22%) experienced substantial underperformance, consistent with risk-off sentiment around capital-intensive sectors and interest-rate-sensitive industries that face pressure from elevated financing costs associated with infrastructure buildout [0].

BCA Research’s top 2026 investment call centers on international stocks continuing to outperform US equities, a positioning that directly contradicts prevailing market consensus favoring American exceptionalism [1][2]. Their poll of market participants found 32.7% of respondents expecting ex-US stocks to be the best-performing asset class in 2026, reflecting growing institutional skepticism about US equity valuations [6].

The international outperformance thesis rests on several pillars. First, US equity market concentration risk has increased dramatically, with the “Mag-7” technology companies representing an unprecedented percentage of market capitalization [2]. When the AI capex cycle turns, this concentration means US markets will bear disproportionate impact. Second, dollar debasement is projected at approximately 10% against a basket of currencies, driven by policy dynamics and capital reallocation toward international markets [2]. Third, international valuations remain substantially more attractive than US equivalents, providing greater margin of safety even as AI-driven productivity gains begin materializing globally.

A critical insight from the BCA Research analysis is the degree to which the US economy has become dependent on AI infrastructure spending. Peter Berezin of BCA Research explicitly states that without the current AI boom, the US economy “would already be in a recession” [3]. This observation highlights the fragile nature of current economic expansion and the significant tail risks associated with any deceleration in AI capital spending.

The dependence manifests across multiple dimensions. Credit market instability represents a material concern, with companies like Oracle accumulating over $100 billion in debt specifically for AI infrastructure development [3]. These leveraged positions face substantial risk if revenue generation fails to keep pace with borrowing costs, potentially triggering credit events that could cascade through financial markets. Labor market exposure is equally significant, as data center construction supports between 100 and 5,000 jobs per project depending on scale—a slowdown would directly impact employment figures in construction and related sectors [3].

The wealth effect from rising AI stock prices has added an estimated $180 billion in consumer spending over the past year [3]. A reversal in AI-related equity valuations would therefore suppress consumption through the wealth effect channel, potentially triggering a feedback loop where reduced consumer spending leads to weaker corporate earnings, which in turn pressures equity valuations further.

BCA Research’s comparison of current AI infrastructure spending to historical technology booms provides valuable context for understanding potential future trajectories. The 19th-century railway boom and late-1990s fiber-optic buildout both experienced similar patterns: initial enthusiasm driving massive capital deployment, followed by periods of consolidation as excess capacity absorbed and weaker players exited [2][3].

However, a significant limitation in the BCA Research framework is the inability to pinpoint timing for the AI capex cycle inflection. Papic acknowledges “I can’t say when it’s going to happen” regarding the bubble bursting, creating uncertainty for tactical portfolio positioning [2]. This timing ambiguity means investors must balance the theoretical appeal of reducing AI exposure against the risk of missing further upside if the cycle extends longer than anticipated.

Notably, BCA Research has demonstrated accurate policy timing in recent years, correctly predicting the Trump administration’s tariff rollback timing [2]. This track record lends credibility to their broader macroeconomic framework even as specific timing remains uncertain.

The AI capex deceleration presents differential risks across sectors and market segments. NVIDIA (NVDA), trading at $189.47 above its 200-day moving average of $161.55, faces high exposure to AI capex deceleration despite Citi analysts identifying a $500 billion opportunity in the Vera-Rubin platform [5]. The stock’s premium valuation leaves limited margin for error if AI spending growth slows more rapidly than anticipated.

Mag-7 technology companies face dual pressure: capital intensity is increasing as infrastructure requirements escalate, while productivity gains are not being fully captured in market valuations [2]. This compression between rising costs and uncertain returns creates structural headwinds for the sector even as AI rhetoric remains constructive.

Credit-sensitive sectors present another risk vector. Companies that have borrowed heavily to finance AI infrastructure face potential distress if operating cash flows fail to justify capital expenditures. The Oracle example—over $100 billion in debt for AI infrastructure—illustrates the scale of potential leverage-related risks [3].

BCA Research and supporting analysis identify several downside scenarios with quantified economic impacts. A pause in AI investment would reduce GDP growth by approximately 0.5 percentage points, representing a moderate correction that would likely trigger meaningful but manageable market volatility [3]. A full AI capex collapse would subtract approximately 1.0 percentage point from GDP growth, introducing significant recession risk given current economic dependence on AI spending [3].

A 20-30% correction in AI-related technology stocks would result in GDP contraction of 1.0 to 1.5 percentage points over a one-year horizon, representing substantial economic headwind [3]. This scenario accounts for the wealth effect channel, where declining equity valuations reduce consumer spending, and the labor market channel, where construction and technology sector employment contracts.

Despite the cautionary thesis, several opportunity windows emerge from the analysis. International equity outperformance represents the clearest tactical opportunity, with BCA Research’s top 2026 call positioning for non-US stocks to continue beating US equivalents [1][2][6]. Geographic diversification away from AI-concentrated US mega-cap stocks provides both valuation support and reduced single-theme risk.

Dollar debasement creates currency positioning opportunities. A projected 10% decline in the US dollar index would benefit assets denominated in other currencies and potentially trigger capital flows into international equities [2]. This currency thesis dovetails with the international equity outperformance call, creating a potentially self-reinforcing dynamic.

Defensive positioning against credit and liquidity events in AI-leveraged companies offers asymmetric return potential. While many market participants remain committed to AI themes, positioning for stress events in overleveraged AI infrastructure companies could capture alpha during periods of volatility.

The declining AI capex contribution is not an immediate signal but rather an evolving trend that investors should monitor continuously. Key indicators to watch include hyperscaler capital expenditure announcements for signals on AI investment trajectory, with significant deceleration below the projected $404 billion for 2026 serving as a confirmation signal [3]. US GDP growth revisions will confirm or contradict the declining AI contribution thesis, while international versus US equity relative performance will validate BCA’s 2026 call.

Technology sector leadership rotation—particularly a shift from mega-cap tech toward broader industrials and consumer discretionary—would indicate markets pricing in the AI capex normalization thesis [0]. The absence of such rotation despite explicit commentary from influential research firms may suggest either market disbelief in the thesis or simply that the trend remains in early stages.

The analysis reveals that AI capex currently contributes approximately 1.0 percentage point to US GDP growth, representing about 50% of total economic expansion [3]. This contribution is declining relative to historical peak levels of 1.5% observed during previous technology booms over the past 150 years [4]. Hyperscaler capital spending reached $344 billion in 2025 with projections of $404 billion for 2026, indicating growth deceleration [3].

BCA Research’s thesis positions international stocks to outperform US equities in 2026, supported by US market concentration risk, dollar debasement expectations, and attractive international valuations [1][2]. The US equity market faces disproportionate risk when the AI capex cycle turns due to concentration in AI-dependent mega-cap technology companies [2].

Risk factors include economic dependence on AI spending (without which the US would be in recession), credit market instability at AI-leveraged companies, labor market exposure from data center construction, and wealth effect reversal from AI stock valuations [3]. The timing of any AI capex cycle inflection remains uncertain despite the theoretical clarity of the thesis [2].

Several important questions remain unresolved by the current analysis. Timeline specificity for the AI capex cycle turn cannot be determined, creating uncertainty for tactical portfolio construction [2]. Geographic diversification analysis should identify specific international markets and companies best positioned to benefit from productivity gains as AI adoption broadens beyond US hyperscalers.

The alternative investment thesis requires development—if AI infrastructure spending slows, what sectors and asset classes will drive global growth? This question becomes increasingly relevant as the AI capex contribution to GDP normalizes. Policy response mechanisms remain unclear, particularly how the Trump administration would respond to AI-driven economic slowing given competing tariff and policy priorities [3].

The BCA Research analysis presents a significant contrarian perspective in an AI-dominated market narrative. While AI capex remains the primary growth driver for the US economy, its declining relative contribution signals normalization rather than collapse. The investment implications center on reduced AI-driven upside expectations for US mega-cap technology, international equity overweight as the preferred 2026 positioning, defensive positioning against potential credit and liquidity events in AI-leveraged companies, and currency positioning to capitalize on dollar debasement.

The timing uncertainty surrounding the AI capex cycle inflection creates a challenging environment for tactical positioning but supports a more defensive, internationally diversified approach to equity allocation. Investors should monitor hyperscaler capex announcements, GDP growth revisions, and sector rotation patterns as confirmation signals for the BCA Research thesis.