The technology sector continues to be at the forefront of innovation, driven by the rapid adoption of artificial intelligence (AI).
While the sector has delivered robust equity performance over the past three years, concerns about a potential valuation bubble and challenges about monetisation persist. However, current evidence indicates that the AI investment cycle remains in its early-to-middle stages, with returns being fuelled by tangible earnings growth rather than speculative excesses.
Although the journey may be marked by periods of volatility, we are confident that the substantial investments in building out AI infrastructure are both justified and essential to meet the future demands of the global economy. The high visibility of incremental revenue from cloud computing, coupled with the significant cost-saving opportunities that AI offers to global corporations, underscores the rationale behind these investments. With a projected USD 500 billion in AI-related spending targeted for 2026, this ambitious goal appears not only achievable but also necessary to unlock the transformative potential of AI across industries.
Moreover, AI has emerged as a national strategic priority for many countries, as it is increasingly recognised as a critical driver of economic growth, national security, and global competitiveness. Governments around the world are implementing policies and allocating resources to ensure leadership in AI development, understanding that dominance in this field will shape the future of innovation, defence, and economic resilience. This alignment of public and private sector efforts further reinforces the long- term growth trajectory of the AI sector.
Valuations: elevated, but not a bubble
Technology valuations are higher than historical averages, yet they remain far outside bubble territory. The Nasdaq 100 is trading at roughly 26x forward earnings, a premium of around 17% to its long-term average, but dramatically lower than the 50x multiples seen during the dotcom peak. Crucially, recent share-price appreciation has been driven primarily by earnings growth rather than multiples’ expansion. Looking ahead, consensus expectations point to technology sector earnings growth of roughly 26% in 2026, with semiconductors leading at around 50%.
The AI capex supercycle continues
The backbone of the AI boom is capital expenditure by ‘Big Tech’. Hyperscalers such as Microsoft, Alphabet, Amazon, and Meta are engaged in what is effectively an arms race to build AI infrastructure, with no major platform able to afford to fall behind. Capital spending is expected to rise by more than 34% again in 2026, sustaining strong demand for semiconductors, networking equipment, data centre hardware, and cloud infrastructure.
Importantly, these investments are largely funded by free cash flow (with the notable exception of Oracle), rather than by excessive leverage. While depreciation expenses will rise sharply in 2026 as new assets are absorbed into balance sheets, revenue growth is expected to offset most of the margin pressure.
Cloud providers: the clearest visibility
Along the AI value chain, cloud computing providers offer the highest revenue visibility. Company take-up of AI is still in its early stages, and broader AI usage is accelerating cloud migration, with penetration still below 50% globally. Backlogs at major cloud providers are growing faster than revenues, pointing to strong future demand.
Large cloud platforms benefit from scale, recurring revenue, and a broad moat. While competition is intensifying, the economics remain attractive: strong graphics processing unit (GPU) demand allows cloud providers to generate double-digit internal rates of return on new investments in data centres.
Semiconductors: the earliest beneficiaries
Semiconductors remain the earliest and most direct beneficiaries of AI investment. While Nvidia continues to dominate AI computing, growth opportunities extend across chip designers, factories, and networking suppliers. Demand for high-performance chips and data centre networking remains strong as AI workloads are scaled up.
That said, this part of the value chain is more cyclical and more sensitive to changes in capital spending. Investors should expect volatility, particularly as expectations for long-term returns are reassessed. Selectivity matters more than ever, with Broadcom ideally positioned to gain share in the AI chip market.
Concerns over company software overdone
While consumer take-up of AI tools is accelerating, monetisation, especially for large language models such as ChatGPT, remains uncertain as competition rises, models become commoditised, and operating costs stay high. These concerns have driven underperformance in company software, amid fears that generative AI will lower barriers to entry and reduce software seat demand through automation. The sector now trades at around 7.7x forward EV/sales, below its 10-year average of 9.6x. While point- solution software faces disruption risk, large company platforms with complex workflows and proprietary data are better positioned to benefit from AI-driven automation.
The AI investment cycle is far from over. While risks exist, from over-optimistic projections to rising debt loads and energy constraints, the structural drivers remain powerful. For long-term investors, diversified exposures across cloud platforms, semiconductors, and power infrastructure providers may offer the most resilient positioning in an era defined by AI-driven growth.
WHERE ARE WE IN THE AI INVESTMENT CYCLE?
Read our Investment Outlook 2026 for more insights.
The financial instruments and investment strategies portrayed in this document are for informative purposes only. They may differ from those effectively held in an investor’ portfolio. Depending on the jurisdiction and investment profile, one or some of these instruments and strategies – including, where applicable, options – may not be permitted, available or suitable. The opinions expressed herein are correct as at 19 January 2026 and are subject to change without notice. Any forecast, projection or target, where provided, is indicative only and is not guaranteed in any way.
.jpg)
