Electronics equipment: AI infrastructure surges while consumer devices contract
Most segments of the electronics equipment market are expected to deteriorate in 2026, with global shipments of smartphones and computers falling by 14% and 11%, respectively, while TV shipments are broadly stable at -0.6%. Servers are the main exception, with shipments expected to grow by 13%, reflecting continued investment in AI infrastructure and data centres. The contrast illustrates the increasingly two-speed nature of the hardware market: AI-related infrastructure continues to expand rapidly, while consumer electronics is being hit by memory shortages, higher component costs and the delayed pass-through of these costs to retail prices.
Despite the cyclical upswing in device shipments, the electronics devices segment remains a highly mature market, with global TV, computer, and smartphone sales peaking in 2011, 2013 and 2016, respectively. Sales are predominantly driven by replacement purchases for existing equipment, and to a lesser extent, by first-time purchases in emerging economies. Limited volume growth has brought about fierce competition for market share over the past decade, with Chinese challengers often displacing former Japanese, European and US leaders.
To escape price-based competition and charge higher prices, companies continue to innovate by introducing new devices featuring improved hardware, including connectivity, screen resolution, processing power and energy efficiency, greater incorporation of additional services such as content subscriptions, as well as software and user-interface improvements. AI-powered devices have so far had no material impact on hardware replacement cycles, given continued consumer scepticism regarding the additional functionalities offered by the technology. In the short term, AI is even hurting parts of the device market indirectly, as demand for high-bandwidth memory used in data centres crowds out conventional DRAM and NAND used in smartphones and PCs, pushing up costs and increasing the risk of production disruptions.
Semiconductors: an AI-driven supercycle with rising instability risks
Global semiconductor sales are expected to rise from around USD 800bn in 2025 to USD 1,500bn in 2026, an increase of almost 90%, before reaching around USD 1,900bn in 2027, according to WSTS, the industry’s global trade association. Sales are being primarily driven by higher prices in critical segments, notably memory chips, reflecting a tight supply-demand balance and an improved product mix, with more chips using the latest technologies hitting the market. The current cycle is unquestionably being driven by soaring demand for expensive chips powering the AI infrastructure boom, especially graphics processors, high-bandwidth memory and advanced logic chips. Because they are more reliant on demand for high-volume, lower-price chips typical of final markets such as industry, automotive and conventional electronics, less AI-exposed segments remain far weaker. A symptom of this two-speed pace in the industry is that many semiconductor companies emerged from the 2022-2023 semiconductor recession only recently, while memory and AI chip suppliers are already operating in an exceptionally tight market. An emerging trend in legacy semiconductor manufacturing nodes is the growing competitiveness of China-based companies.
This AI-driven semiconductor cycle, however, brings its own share of vulnerabilities: capacity expansions are being made on assumptions about future AI demand that remain highly uncertain. Any mismatch, whether excess supply if AI monetisation disappoints or renewed bottlenecks if demand keeps accelerating, could quickly destabilise global supply chains and the broader economy. The memory segment already illustrates this risk: strong demand for high-performance memory chips used in AI data centres is crowding out conventional memory chips used in smartphones and PCs, contributing to contract price increases of 50% to 100% depending on memory type and specifications.
Geopolitical risk is another factor that has been gaining traction in recent months, with US–China rivalry progressively reshaping global trade flows. The imposition of export controls, blacklists and licensing requirements has created new chokepoints in critical areas such as AI-grade memory and manufacturing equipment, which is increasing the likelihood of supply disruptions. At the same time, US semiconductor companies face shrinking market opportunities, with some deriving up to half of their revenues from China. The growing pressure to comply with geopolitical regulations also exposes smaller firms to significant financial and operational risks. Moreover, the divergence of standards and trade policies between the US and China – and even within the US-aligned bloc – adds further complexity and cost, potentially leading to fragmentation and strategic realignment among key players such as Taiwan, Japan, South Korea and Europe.
IT services and software: AI supports demand but threatens the traditional SaaS model
The IT services and software sector comprises the sub-segments of consulting, programming, data processing, managed services and software, collectively generating worldwide sales of USD 2,900 billion. With average annual growth of almost 9% over the past decade, the sector’s robustness stems from the use of information and communication technologies by companies and public authorities to improve their efficiency. Growth is expected to remain positive in 2026, at 5.3%, helped by the deployment of AI, cloud computing, cybersecurity and data-related services.
Essentially made up of national and regional markets, like most service activities, the sector is nevertheless experiencing increasing internationalisation in the software, managed services, programming and data processing segments. For example, over the past decade, IT service exports from India and the US have more than doubled, and now account for over USD 150 billion a year. After big data, cloud computing and cybersecurity, companies in the sector are now focusing on the deployment of artificial intelligence technologies to further accelerate their growth. AI is also transforming the sector’s own operations with coding assistants, automated testing, agentic workflows, knowledge management tools and AI-enabled customer support raising productivity, shortening delivery times and reducing the labour intensity of some tasks.
Meeting growing demand is the main challenge facing companies in this sector, where labour is both the main cost item, between 50% and 75% of sales depending on the segment, and the main source of competitiveness. In recent years, a shortage of qualified profiles has weighed on companies' ability to meet demand, thereby driving up wage costs. AI could partly ease this constraint by improving developer productivity and automating lower-value tasks, but it also raises fears of a “SaaSpocalypse”: the risk that AI agents, low-code tools and internally built applications reduce the need for some traditional software subscriptions. The sector must also contend with the growing demands of regulatory authorities, particularly in terms of data collection, hosting and security under the General Data Protection Regulation in the European Union, and the control of illegal or misleading content under the Digital Services Regulation. The rise of AI adds a further layer of complexity, as companies must invest in new skills, infrastructure, governance and compliance capabilities before the profitability of many AI use cases has been firmly established.
Telecommunications: usage explodes while sales stagnate
The telecommunications sector generates annual worldwide sales of around USD 1,350 billion. Telecoms markets are national, oligopolistic and most often dominated by a former state monopoly. International groups are the exception and most of them have significantly reduced their activities outside their home countries and regions in recent years due to the lack of economies of scale and sufficient synergies. Telecom services are expected to remain relatively resilient in 2026, with sales growth of 4.4%.
Driven by the rapid development of fixed and mobile networks in developed economies between 1995 and 2015, the sector is now largely mature and profitability depends on the size and degree of concentration of domestic markets. In this respect, US operators benefit from a market that is both vast (330 million inhabitants) and concentrated (Verizon, AT&T and T-Mobile hold over 90% of the market), thus enabling high margins. In contrast, European markets remain national and more fragmented, and are therefore comparatively less profitable, although a current shift in EU antitrust doctrine and the Digital Networks Act could open the door to greater consolidation both at a national and European level.
In emerging countries, the deployment of mobile telecoms continues to sustain low-reward growth due to comparatively low average revenue per user (ARPU). The number of mobile and fixed broadband subscriptions worldwide is expected to grow by 1% and 3% per year, respectively, over the next five years. After a decade of almost zero growth, the sector's sales are likely to increase only marginally during the same period.
Telecommunications usage, however, continues to grow at a steady pace thanks to the deployment of faster and more extensive fixed (fibre optic) and mobile (5G) networks. Data volumes exchanged on mobile and fixed networks worldwide are expected to grow by 20% and 12%, respectively, per year over the next five years. The challenge for telecoms operators is to monetise this improvement in service quality with consumers who are opportunistic and price-conscious (in developed countries) or budget-constrained (in emerging countries).
In the absence of sustained business growth, telecom operators will be relying on cost-cutting to boost profits. In Europe, many operators have been gradually withdrawing from their mobile network management activities through the creation of tower companies, whose capital they have often opened up. Outsourcing these activities meets the dual need to reduce the sector's very high capital intensity – capex is equivalent to 15-20% of sales – and to raise funds to reduce high debt levels. A swathe of mergers and acquisitions and share buyback programmes have taken place as a result of major investments in network infrastructures.
ICT
