Very few technological developments in recent memory have captivated the post-2010s zeitgeist the way artificial intelligence (AI) has. As the AI talking point reaches a fever pitch, investors continue to debate whether this is a bubble waiting to burst or a structural boom that will remain unstoppable for a long time. To help investors cut through the noise and understand AI’s ever-changing investment implications, we introduce our first-of-its-kind index called the ‘AI bubble meter,’ which measures sector confidence from Worst (<‑3) to Best (>+3).

AI – bubbling with opportunity or fizzling out?

Our AI bubble meter – an innovative metric that weighs five key industry catalysts against five key risks and across seven risk‑reward buckets – has risen from 2 in January to 2.5 in February 2026, signalling a shift from a ‘good’ to a ‘better’ risk‑reward profile. This points to potential returns of 5-10% for the AI theme over the next 3-6 months. Notwithstanding the fragile investor sentiment amid heightened geopolitical volatility this year, our meter suggests that the AI bubble fears are overdone.

Despite the relatively short history of the AI cycle, historical back‑testing indicates that our AI confidence index has solid predictive power for three‑month forward returns, with higher readings generally associated with stronger performance across the AI theme. Using the Nasdaq-100 as a proxy, the correlation between our AI confidence index and subsequent market performance has been consistently around or above 0.5, suggesting a meaningful positive relationship.

The data behind the confidence

We see five main catalysts supporting the AI cycle. These, including capital investment and adoption, continue to show strong momentum:

– AI capital expenditure (capex) has remained a key pillar of the structural AI theme over the past three years. Despite concerns around aggressive buildouts, we expect 40% y/y growth in 2026 and a solid 25% CAGR in 2025-2030. With ongoing Q4 2025 earnings results likely to reaffirm the strong capex outlook, our capex trend score remains unchanged at 4 for February, consistent with January (on a 1-5 scale).

– AI adoption rates, as measured by US Census Bureau data, are likely to continue to trend higher. Latest readings indicate a 15-20% adoption in the US – a significant jump from around 3.8% in late 2023. We maintain February’s AI‑adoption‑rate score at 3.

– Earnings revisions for AI‑related companies remain solid, led by strength across the semiconductor industry. Over the past month, we have seen a further pickup in positive revisions. Consequently, we raise our earnings revision score from 3 to 3.5 for February.

– AI‑related companies’ margins help assess their quality of growth, as margins capture pricing power, operating leverage and overall efficiency. IT companies’ operating margin profiles show margin trends currently remain steady. Accordingly, we maintain February’s quality‑of‑growth score at 2.5.

– Big Tech has been instrumental in driving the global AI boom, supported by its first‑mover advantage and substantial capex. Strong capex intensity (capex‑to‑sales) reinforces its sustained investment momentum. Accordingly, we maintain our big tech commitment score at 3.5 for February.

From DeepSeek to regulatory headwinds, risks remain manageable

While the outlook for AI remains largely positive, lingering concerns remain. Most investors understand that this is not a risk-free environment, but some of their worries can be assuaged:

– Disruption by low-cost China-based large language models (LLMs), such as DeepSeek, remains one of the biggest risks to the market-dominating US-based LLMs. While China-based LLMs have undoubtedly gained traction, we believe the AI market is large enough to accommodate multiple winners, leaving the disruption risk manageable for now.

– Valuation and funding risks have moderated. While AI and broader tech stock valuations spiked in late 2025, the AI theme’s muted performance in Q4 2025 helped normalise valuation levels. Similarly, funding risks remain manageable despite elevated capex weighing on free‑cash‑flow margins across US equities, given credible pathways towards monetisation are emerging. Cash‑flow generation among leading platforms also remains robust.

–The global regulatory and macro environment remains benign and manageable, supported by government policies that favour an enabling versus a restrictive stance on AI.

Strategising for the next phase

With the risk-reward profile improving, we favour an ‘offensively defensive’ approach to AI exposure. On the offensive side, opportunities remain strongest within cyclical semiconductors, including AI accelerators, semiconductor capital equipment and foundries. On the defensive side, we favour US- and China-based internet platforms and early AI adopters in global healthcare. Big tech also offers solid near‑term potential due to strong H1 seasonality. However, we remain cautious on traditional business models facing disruption risk, particularly across hardware, legacy software and other service-related industries.

While the AI bubble debate rages on, substance continues to allay fears about hype. Nevertheless, investors should not be complacent, and instead, should follow our AI bubble meter regularly to monitor AI investment opportunities.