Skip to main content
Information Technology

AI's Impact on the IT Sector: A New Technology Cycle

Pomegra Learn

How Is Artificial Intelligence Reshaping the IT Sector?

Artificial intelligence represents the most significant structural shift in the Information Technology sector since the cloud computing revolution of the 2010s — and arguably since the internet itself. The scale of capital investment flowing into AI infrastructure, the speed of adoption across software categories, and the potential for AI to fundamentally alter the economics of knowledge work are creating both extraordinary opportunities and genuine disruption risks within the IT sector. Understanding which companies are primary beneficiaries, which face competitive threats, and how to assess the durability of AI-driven earnings growth is the central analytical challenge for IT sector investors in the mid-2020s.

Quick definition: AI's impact on the IT sector encompasses the demand shock for AI-specific semiconductors and infrastructure, the integration of AI capabilities into software products to drive pricing and retention, and the potential long-term displacement of certain IT services and software categories by AI-native alternatives.

Key takeaways​

  • Nvidia's data center GPU revenues grew roughly 10x in two years (2022–2024), driven almost entirely by AI training demand
  • The four largest hyperscalers committed to $200+ billion in annual data center capex by 2024, primarily AI-motivated
  • Software companies are embedding AI features to justify price increases of 10–30% for existing products
  • AI coding assistants may reduce software development labor demand, threatening IT services company margins
  • The AI cycle is creating an "arms race" in semiconductor and cloud investment that may ultimately prove cyclical

The semiconductor demand shock​

AI has created an unprecedented and rapid revenue cycle for AI-optimized semiconductor companies. The core driver is the architecture of AI training: large language models require performing trillions of floating-point matrix multiplication operations during training, which can be parallelized across thousands of GPUs far more efficiently than on traditional CPUs. This makes Nvidia's H100 and subsequent GPU architectures the essential hardware of AI infrastructure.

The revenue numbers are extraordinary by any historical comparison. Nvidia's data center segment revenue grew from approximately $4.3 billion in fiscal year 2022 to approximately $47.5 billion in fiscal year 2024 — more than a 10x increase in two years. Operating margins expanded from roughly 25% to more than 60%, generating free cash flow that transformed Nvidia into one of the most valuable companies in the world.

The AI semiconductor opportunity extends beyond Nvidia. High-bandwidth memory (HBM) — the high-speed memory stacked alongside GPU processors — is supplied by SK Hynix and Micron, both experiencing surging revenue from AI demand. Custom AI accelerators (ASICs) designed by hyperscalers and developed with foundry partners like Broadcom are creating an alternative to pure GPU-based compute. Power semiconductors, voltage regulators, and thermal management components for AI servers are all experiencing demand surges.

Cloud infrastructure: the capex super-cycle​

AI infrastructure has driven hyperscaler capital expenditure to levels that represent a structural step-change from the pre-AI period. Amazon, Microsoft, Google, and Meta collectively spent approximately $200+ billion in capex in 2024 — roughly double their 2022 combined spending.

This spending is driven by several imperatives:

  • Training new AI models: Foundation model training requires massive compute clusters running for weeks or months
  • Inference at scale: Serving AI results to hundreds of millions of users requires additional GPU or specialized inference chip capacity
  • Competitive necessity: No major hyperscaler can afford to fall significantly behind in AI capabilities without risking cloud market share

The near-term financial impact on hyperscalers is capex-driven free cash flow compression. Microsoft, for example, saw its capital expenditures rise from roughly $20 billion in fiscal 2022 to approximately $44 billion in fiscal 2024. Amazon's capex approached $75+ billion by 2024. This spending compresses near-term FCF yield and P/E ratios at these companies, but if the investments yield superior cloud revenue and margins in 3–5 years, the long-run value creation could be substantial.

How it flows​

Software monetization through AI​

Software companies across the IT sector are embedding AI features into their products to justify price increases and differentiate from competitive offerings. The primary monetization strategies are:

Copilot premium tiers: Microsoft's Copilot for Microsoft 365 carries a premium of approximately $30 per user per month on top of the base Microsoft 365 subscription. Salesforce's Einstein Copilot, ServiceNow's Now Assist, and similar products carry similar premium price points. At scale, these AI add-ons can drive 15–25% average revenue per user (ARPU) increases.

AI-enhanced product capabilities: Many software companies are using AI to automate workflows, improve search and discovery, and provide intelligent recommendations — capabilities that make the products stickier and improve NRR by reducing churn.

New AI-native product categories: Companies like GitHub (owned by Microsoft) have launched AI coding assistants (GitHub Copilot) that have achieved millions of paid users rapidly, creating entirely new revenue lines.

The sustainability of these AI-driven monetization gains is uncertain. If AI capabilities become commoditized and available from multiple vendors at zero marginal cost, the premium pricing associated with AI features may compress toward zero. The companies with the deepest customer integration — where AI features are embedded in core workflows that would be expensive to replace — are better positioned to maintain AI pricing than those offering easily substitutable point solutions.

AI's threat to IT services​

The very productivity gains that make AI attractive to enterprise software buyers create a potential disruption threat for IT services companies. If AI coding tools increase software developer productivity by 20–40% — as various studies suggest — then IT services companies could deliver the same scope of work with fewer people, or the same number of people could deliver more work at the same price.

In the near term, this dynamic is playing out as a productivity opportunity rather than a purely deflationary threat. IT services companies are using AI tools to improve delivery efficiency, expanding margins rather than reducing prices. But as AI capabilities improve and adoption becomes industry-standard, competitive pressure will require IT services companies to pass at least some productivity gains to clients through reduced prices, compressing revenue growth.

The most sophisticated IT services companies (Accenture, in particular) are positioning AI implementation and governance as a high-value consulting service that requires deep human expertise — attempting to shift their revenue mix toward higher-value services that AI complements rather than replaces.

Real-world examples​

Nvidia's fiscal year 2024 performance is the defining real-world example of AI's IT sector impact. Revenue of approximately $60.9 billion, up from $26.9 billion the prior year, represented the largest single-year absolute revenue increase of any company in history at that point. Gross margins exceeded 74%. Operating income exceeded $32 billion. The stock price rose from roughly $150 at the start of fiscal 2024 to more than $600 at fiscal year-end, adding more than $1 trillion in market capitalization in a single year.

This performance, while extraordinary, also illustrates the concentration risk of AI hardware spending. Nvidia's success depends substantially on continued hyperscaler willingness to spend at current rates on GPU infrastructure. If AI revenue generation disappoints, if alternative chip architectures gain market share, or if training efficiency improvements reduce the compute required per model, Nvidia's revenue could moderate significantly from its peak rates.

Common mistakes​

Treating the AI hardware cycle as permanent rather than cyclical. Technology hardware investment cycles always overshoot — companies over-invest relative to near-term demand, creating inventory corrections. The AI infrastructure cycle has characteristics of both structural demand (AI is a real capability that creates real value) and cyclical momentum (some hyperscaler spending is competitive insurance rather than demand-driven necessity).

Assuming all software companies benefit equally from AI. AI is a genuine tailwind for well-positioned software companies that can embed AI functionality into sticky enterprise workflows. It is a competitive threat to software companies whose products could be replaced by AI-native alternatives. Distinguish between the two before assuming uniform AI-driven upside.

FAQ​

How do I evaluate AI exposure in a technology ETF?​

Review the ETF's top 10 holdings and their exposure to AI infrastructure (GPU, HBM, data center), AI software (Microsoft, Salesforce, ServiceNow), and AI services (cloud hyperscalers). For more targeted AI exposure, consider SOXX for semiconductor focus or look for thematic AI ETFs that focus specifically on AI infrastructure and applications. Consult sec.gov for full ETF holdings disclosures.

Will AI reduce IT sector employment and stock prices?​

AI's labor productivity effects are likely to show in IT services company margins over 5–10 years, not immediately. Short-term impact on stock prices is ambiguous: higher margins from AI efficiency tools could offset revenue growth moderation. The companies at greatest risk from AI labor displacement are those providing routine, codifiable IT services with limited value-added differentiation.

Summary​

Artificial intelligence represents the IT sector's most significant structural investment cycle in decades, creating extraordinary near-term earnings growth for AI-optimized semiconductor companies (led by Nvidia), driving hyperscaler capex to record levels, and enabling software companies to monetize AI features through premium pricing. The cycle also creates disruption risks for IT services companies facing AI-driven productivity pressure and for software companies whose products compete with AI-native alternatives. Whether the AI cycle proves to be a durable multi-decade shift in technology economics or a faster-saturating infrastructure cycle depends on the pace of AI capability improvement, the efficiency gains that reduce compute requirements per model, and ultimately whether enterprises find the value justifying their AI investments. Investors who remain anchored to valuation fundamentals — rather than extrapolating cycle momentum indefinitely — are best positioned to navigate the opportunity and risk this transformation presents.

Next​

→ IT Software Earnings Quality Metrics