Mini-trends in mega-tech year

Amid rising costs, tightening constraints, geopolitical tensions, and tariff negotiations, here are a few trends that emerged in 2025.

Artificial Intelligence: AI did not “arrive” in 2025 but it became an indispensable part. No launch of an electronic product missed out on it. Yet this shift came with a reckoning. Running AI systems at scale proved expensive. Training and inference costs rose sharply as the demand for high-performance chips surged. Firms moved away from the large, general-purpose models toward smaller, task-specific systems. The defining change was not adoption, but discipline. India’s largest IT services firm, TCS, disclosed an annualised AI revenue run-rate of about $1.5 billion. This was the first time it quantified AI-led revenue. The disclosure pointed to rising investor expectations for transparency around AI returns. This was mirrored globally. One of the clearest signals came when Alphabet announced a $4.75 billion acquisition of Intersect Power. The deal was explicitly linked to securing reliable electricity supplies for data centres. In effect, energy and compute emerged as strategic inputs. Spending on AI infrastructure moved decisively out of discretionary innovation budgets, and into core operating expenditure. CFOs treated AI the way they did with manufacturing or logistics. The combination of publicly-disclosed AI revenues, and multi-billion-dollar investments captured the defining AI trend of 2025. The central question was how efficiently firms can finance it, scale it, and justify its impact.

Compute: AI’s rise exposed a vulnerability that few firms had planned for: compute scarcity. In 2025, access to advanced chips, data-centre capacity, and cooling infrastructure emerged as competitive advantages. Cloud providers tightened capacity, long-term contracts became more common, and businesses delayed or re-prioritised workloads to manage costs. Export controls and licensing became part of the story. There was a churn around Nvidia’s H200 shipments to China, with policy shifts and demand for transparency.

Tech Energy: As compute demands rose, so did the demand for electricity. Data centres, manufacturing, and AI infrastructure pushed requirements to levels that utilities struggled to meet consistently. Global data-centre consumption may double to ~945 TWh by 2030, growing ~15 per cent a year from 2024 to 2030. Data centres consumed ~415 TWh in 2024 (about 1.5 per cent of global generation), a baseline that made “AI power” a mainstream debate through 2025. In several markets, power availability influenced where firms expanded. Energy efficiency became a strategic concern rather than a sustainability footnote. Investments flowed into more efficient hardware, alternative cooling systems, and longer equipment life cycles. This resulted in another sub-trend in which tech firms partnered with energy ones. The most telling move was the power land-grab by Alphabet. AI demand forced tech firms closer to the energy supply chains. If energy becomes a constraint, owning or locking in energy becomes a defensible advantage.

 

Hardware: For years, software promised to abstract away physical limits. In 2025, the latter limits returned. Semiconductors, fabrication plants, logistics networks, and cooling systems became bottlenecks. Capital expenditure increased as firms invested in physical infrastructure rather than only digital tools. Margins came under pressure as hardware costs shot up. The digital economy was reminded that its future still rests on steel, silicon, and power.

Supply Chains: They did not revert to pre-2020 configuration. 2025 reinforced a cautious move toward regionalisation. Electronics, auto, and industrial tech firms diversified suppliers, adopted “China plus one, US plus one, and India plus one” strategies, and accepted higher costs in exchange for resilience. Governments remained deeply involved, shaping flows through incentives, export controls, and industrial policy. Predictability, not efficiency, was crucial.

 

Consumer Technology: Innovation slowed noticeably. Smartphones, personal devices, and home electronics delivered incremental upgrades. The growth was largely driven by replacement cycles, and premium products, rather than innovation. According to NielsenIQ, the global consumer technology and durables market grew 4.6 per cent in the first half, generating $403 billion in revenue. The projection for the year is near two per cent, underscoring weak volumes amid inflation and trade pressures. The data pointed to modest smartphone sales growth of 2.6 per cent in Q3-2025, signaling a fragile recovery with demand concentrated at the higher-end, AI-capable devices rather than broad upgrades. Premium segments outperformed, with global premium smartphone sales rising about eight per cent year-on-year, and commanding over 60 per cent of industry revenue in H1-2025. Consumers held older mid-range devices longer while still upgrading to flagship experiences. This cautious buying reflected a shift away from frequent hardware churn toward value and longevity. Enterprise technology saw consolidation as businesses sought stability and integration over fragmented point solutions. Strategic acquisitions and platform play illustrate this shift. Technology distributor, TD SYNNEX, expanded digital services and cloud commerce footprint through multiple acquisitions to bolster multi-cloud orchestration capabilities, and international reseller reach, consolidating vendor offerings under broader enterprise platforms. Qualcomm pursued strategic buys to expand into data-centres, and AI-infrastructure, positioning itself as a supplier across both edge and cloud enterprise segments. The centre of gravity therefore shifted from novelty to reliability.

 

Tech Hiring: Tech labour markets corrected in 2025. Global layoffs that began in late 2022 continued to cast a long shadow. According to Layoffs.fyi, the sector saw more than 2,60,000 job cuts across 1,100 firms between 2023 and 2024. While the pace slowed in 2025, hiring did not rebound. Demand rose for engineers who optimise systems, manage infrastructure, ensure security, and control costs. Generalist roles and speculative teams declined. The emphasis shifted from building new capabilities to sustaining complex ones. The labour market became more disciplined and less forgiving. Speed of hiring was no longer a competitive advantage. Stability, institutional knowledge, and system stewardship carried more weight. This reset signaled a broader maturity phase in tech adoption.

 

KPI Shift: The most consequential shift was cultural. AI was judged by what it cost. The KPI framework changed. Capability metrics such as model size, accuracy, or feature breadth gave way to economic ones such as cost per inference, infra utilisation, and impact on margins. This was driven by scale and expense. Industry estimates show global spending on AI-related infra crossing $300-400 billion in 2025. The effect was visible in how projects were approved and measured. Firms narrowed AI deployments to areas with clearer payback, such as customer support automation, internal productivity tools, fraud detection, and supply chain optimisation. In 2025, AI stopped being a technological race, and became a financial one.

 

Taken together, these trends suggest that 2025 marked a turning point. Technology did not retreat, but matured. Growth slowed, constraints tightened, and ambition was tempered. The defining story was not innovation, but accountability.