1. Why This Sector Exists [60 WORDS MAX]
Every business on earth needs to store, move, and process information. IT is the infrastructure of modern commerce — the plumbing and electricity rolled into one. Customers don't cancel servers the way they cancel magazine subscriptions; switching costs are enormous. A portfolio needs it for growth, margin expansion, and a hedge against human labour inflation.
2. What's Happening Right Now [150 WORDS MAX]
What happened: April was a monster month.
The Technology Select Sector SPDR Fund posted its best month since October 2002, recording gains of 20%.
The PHLX Semiconductor Index soared, gaining nearly 40% during the month.
Earnings drove it:
Alphabet's Q1 EPS came in at $5.11 on $109.9B revenue versus Wall Street's expectation of $2.63 EPS on $107.03B.
Intel blew past expectations — EPS of $0.29 vs. $0.01 estimated, revenue $13.58B vs. $12.42B — and shares jumped 20% after hours.
Microsoft beat on earnings and revenue but slipped after reporting $31.9B in capex for its fiscal Q3.
Why it happened:
The blended IT sector earnings growth rate hit 46.3% for Q1.
Geopolitical de-escalation hopes unlocked suppressed valuations.
Software companies mostly struggled on fears that AI will disrupt their businesses, creating a historic divergence between hardware and software.
What it sets up: Hardware re-rating looks extended at these levels; the market now turns its scrutiny to whether AI capex translates into enterprise software revenue over the next 4–8 weeks.
3. How the Money Works [100 WORDS MAX]
Revenue comes from licences, subscriptions, and hardware cycles — subscriptions being the landlord analogy: once a tenant moves in (installs your software), eviction is expensive. The two costs that matter are R&D (to stay ahead of the next disruptor) and cost of goods sold on hardware (wafer costs, packaging). Scale helps dramatically: a cloud platform's marginal cost of adding the millionth customer approaches zero. What separates great from average is switching cost depth. Microsoft is the gold standard — Office, Azure, Teams, and Active Directory are a spider web; pulling one strand tightens the others.
4. The 4 Macro Drivers
Driver 1: Interest Rates & Discount Rate Sensitivity
Mechanism: IT generates a disproportionate share of its cash flows in the distant future. Higher rates raise the discount rate, compressing the present value of those future flows — like repricing a 30-year mortgage upward. Long-duration growth stocks de-rate fastest. Now:
The forward 12-month P/E is 20.9, above both the 5-year average of 19.9 and the 10-year average of 18.9, and above the 19.7 at end of Q1.
Multiples are re-expanding, implying the market is pricing rate relief. 2nd-order effect: Rate cuts don't just re-rate multiples — they trigger a venture capital funding restart, seeding the next generation of disruptors that will eat at incumbents 3–4 years later. Threshold: 10-year Treasury yield sustained below 4.0% unlocks full multiple expansion; a move back above 4.75% reverses it.
Driver 2: AI Capital Expenditure Cycle
Mechanism: Hyperscalers (Microsoft, Amazon, Alphabet, Meta) are building out data centres at historic rates. That spending flows directly into semiconductor revenue, power infrastructure, and cooling hardware — then, with an 18-month lag, into software as enterprises deploy AI tools on top of that infrastructure. Now:
Microsoft reported $31.9B in capex in a single quarter
— annualising above $120B.
Anthropic committed to spending more than $100B on AWS over the next 10 years.
2nd-order effect: Massive capex crowds out free cash flow, pressuring hyperscaler buybacks. That weakens EPS growth even as revenue grows — the market hasn't fully priced this. Threshold: Any guidance cut in hyperscaler capex is a red flag for chip stocks within 48 hours.
Driver 3: Enterprise IT Budget Cycles
Mechanism: CIOs set annual budgets in Q4 for the following year. When macro uncertainty rises (tariffs, recession fears), discretionary IT projects are deferred — the first line item cut, like a restaurant pausing a kitchen renovation. This hits IT services and enterprise software before it hits semiconductors. Now:
Infosys crossed $20B in FY26 revenue, growing 3.1% in constant currency
— solid but not spectacular, signalling caution in discretionary enterprise spend.
Adam Parker of Trivariate Research thinks software companies are "structurally in trouble," seeing better opportunities in semiconductors.
2nd-order effect: Deferred IT projects don't disappear — they stack up, creating a pent-up spending surge when confidence returns. Threshold: ISM Services PMI sustained above 55 signals CIO budget unlocks.
Driver 4: U.S.-China Tech Decoupling & Export Controls
Mechanism: Export restrictions on advanced chips to China reduce total addressable market for semiconductor companies and force supply-chain relocations that increase cost of goods sold. Think of it as a government-mandated shrinkflation of your customer base. Now:
Intel has been championed by the Trump administration, which made the U.S. government its largest shareholder as part of efforts to bring chip manufacturing stateside.
Intel announced it will join Musk's Terafab chip complex in Austin to help design, fabricate, and package ultra-high-performance chips for SpaceX, xAI, and Tesla.
2nd-order effect: Reshoring raises domestic chip production costs short-term but creates a moat for U.S. fabs over 5 years — TSMC's Arizona ramp is the real watch item. Threshold: Any new BIS entity-list additions or chip licensing rule changes are immediate revenue events for NVDA, AMD, and INTC.
5. Sector Map
| Sub-Industry | What It Does | Key Driver | Main Risk |
|---|---|---|---|
| Semiconductors | Designs/manufactures chips | AI capex cycle | Export controls, cycle turns |
| Cloud Infrastructure | Rents compute & storage | Enterprise IT budgets | Capex overcapacity glut |
| Enterprise Software | Sells licences/subscriptions | Rate environment (multiples) | AI disruption of legacy apps |
| IT Services & Consulting | Implements tech for clients | Corporate capex budgets | Automation reducing headcount |
| Hardware & Networking | Servers, switches, PCs | AI build-out demand | Component cost inflation |
6. Company Case Studies
Case Study 1: Intel (INTC) — Fallen giant re-rating on foundry + AI inference revival
Business: Intel earns revenue from PC/server CPUs, data centre chips, and an emerging foundry business that manufactures chips for third parties. The key cost is R&D and fab depreciation — enormously capital-intensive. At scale, the foundry business has high fixed costs but near-zero marginal costs per wafer batch once utilisation is full.
Moat: Manufacturing process technology and advanced packaging are Intel's moat.
Intel is one of only three global companies offering the most advanced type of chip packaging
— a genuine bottleneck. The moat was eroding for a decade; it is now tentatively widening under CEO Lip-Bu Tan.
Macro Linkage: Driver 4 (decoupling) is Intel's biggest tailwind and Driver 2 (AI capex) is its direct revenue source.
The once-sleepy CPU market has taken off as agentic workloads shift compute needs beyond Nvidia's GPUs.
More AI agents running simultaneously means more CPU cycles — Intel owns that market.
Watch: (1) Foundry external revenue — currently small; crossing $2B/quarter signals real third-party traction. (2) Advanced packaging bookings:
CFO Zinsner said he's confident packaging will bring billions per customer, up from prior hundreds of millions estimates
— track whether that guidance firms or slips. Risk:
Some analysts are waiting to see promising yields on Intel's next-generation 14A technology, planned for 2028 or beyond
— a yield disappointment is an existential signal. Early warning: TSMC takes a key foundry customer Intel was chasing.
Valuation: Trades ~28x forward non-GAAP earnings post-rally. Fair to slightly expensive given execution risk on 14A ramp; re-rating only justified if foundry customers multiply.
Case Study 2: Microsoft (MSFT) — The spider-web incumbent monetising AI on top of an unbreakable installed base
Business: Revenue from Office 365, Azure cloud, Dynamics, LinkedIn, and Xbox. Azure is the growth engine; Office is the annuity. Key cost: data centre capex (now $31.9B/quarter) and R&D. At scale, Azure's incremental margin is very high — the 1,000th dollar of cloud revenue costs almost nothing extra to deliver.
Moat: Active Directory and Office are embedded in virtually every enterprise on earth. Replacing them requires retraining every employee — prohibitively expensive. The moat is widening as Copilot AI layers onto the existing subscription, raising revenue per seat without raising churn.
Macro Linkage: Driver 1 (rates) sets the multiple; Driver 2 (AI capex) is both a cost (capex drag on FCF) and a revenue driver (Azure AI workloads). The tension: heavy capex compresses near-term FCF yield, making the stock rate-sensitive even though fundamentals are strong.
Watch: (1) Azure revenue growth rate — current ~35% YoY; any deceleration below 30% triggers multiple compression. (2) Copilot seat additions — the per-seat price-up is the margin expansion engine; flat adoption kills the thesis. Risk: The bear case is simple: $31.9B/quarter in capex that doesn't convert to revenue fast enough crushes FCF and forces a dividend/buyback cut. Early warning: Azure growth misses two consecutive quarters.
Valuation: ~31x forward earnings. Expensive on traditional metrics but fair when priced on FCF-to-growth (PEG ~1.5x) if Azure sustains 30%+ growth. Any rate spike makes this multiple untenable.
Case Study 3: Infosys (INFY) — AI-era IT services adapter: lower multiple, real execution risk
Business: Infosys sells IT consulting, outsourcing, and AI-transformation services to large enterprises. Revenue is primarily time-and-materials plus managed services contracts — sticky but not SaaS-sticky. Key costs are labour (offshore engineers in India) and sales/bid costs. Scale helps margins modestly through utilisation rates.
Full-year FY26 revenue crossed $20B, growing 3.1% in constant currency, with a 21.0% adjusted operating margin.
Moat: Multi-year enterprise contracts and deep domain knowledge in verticals (banking, energy, retail). The moat is eroding as AI reduces the need for large implementation teams — the core risk. Breadth of AI tool partnerships partially offsets this.
Macro Linkage: Driver 3 (enterprise IT budgets) dominates. When CIOs pause discretionary spending, Infosys feels it first — new bookings slow before revenue falls.
Infosys reported large deal wins totalling $14.9B in FY26, with FY27 guidance pointing to modest growth and stable margins.
Modest = risk.
Watch: (1) Large deal total contract value per quarter — below $3B signals demand softness. (2) Headcount net additions — AI-driven productivity should reduce headcount needs; if Infosys is adding aggressively, margins will disappoint. Risk: AI automates the exact work Infosys charges for — application maintenance, testing, and basic development. If agentic AI reaches 20–30% of their service mix, headcount and billable hours shrink structurally. Early warning: revenue-per-employee growth stalls.
Valuation: ~22x forward earnings — reasonable but not cheap given 3% revenue growth. A re-rating requires evidence of AI-led margin expansion above 22%.
7. How to Value These Companies [80 WORDS MAX]
Use EV/FCF and EV/Sales alongside P/E. Why: earnings are manipulated by D&A, stock comp, and capex timing; FCF is harder to fake. Typical ranges: 20–40x FCF for cloud/software; 12–20x for IT services; 15–25x for semiconductors at mid-cycle. The most common junior mistake: using GAAP P/E for high-capex companies, ignoring that $30B in annual data-centre investment is being expensed over decades — massively understating true cash earnings power.
8. KPIs That Actually Matter
| KPI | What It Signals | Why It Beats EPS | Benchmark |
|---|---|---|---|
| Net Revenue Retention (NRR) | Existing customers spending more | EPS misses expansion vs. churn | >120% = excellent |
| RPO (Remaining Performance Obligations) | Contracted future revenue | Forward revenue EPS can't show | Rising QoQ is bullish |
| Azure / AWS growth rate YoY | Cloud demand health | Revenue quality vs. one-time items | >30% = strong |
| Semiconductor book-to-bill ratio | Orders vs. shipments | Leads revenue by 1–2 quarters | >1.0 = demand healthy |
| Free Cash Flow margin | True cash profitability | Strips out non-cash stock comp noise | >25% = world-class |
| Large deal TCV (IT services) | Pipeline 12–18 months out | Lags revenue; leads it in services | Rising trend critical |
9. Risk Map
Risk 1: AI-Driven Software Commoditisation
AI coding tools (GitHub Copilot, Cursor) can replicate in minutes what took developers weeks — compressing the value of software licences and IT services contracts. Transmission: software pricing power falls → NRR drops below 100% → multiple contracts from 30x to 15x revenue. Historical precedent: SaaS de-rated 60–70% in 2022 when growth slowed. Early warning: two consecutive quarters of NRR below 110% at any major SaaS vendor.
Risk 2: Hyperscaler Capex Reversal
If AI ROI doesn't materialise for enterprise customers, hyperscalers cut data-centre build plans. Transmission: chip order cancellations → semiconductor inventory glut → ASP collapse → earnings cuts of 30–50% within two quarters. Precedent: 2022–23 cloud spending pause crashed Marvell, AMD, and Nvidia earnings estimates by 40%+. Early warning:
a Wall Street Journal report flagged that OpenAI may be on track to miss key revenue and user targets
— AI demand disappointment is the trigger to watch.
Risk 3: Export Control Escalation
Additional U.S. chip export restrictions to China remove a material revenue stream overnight — no hedging, no repricing. Transmission: 10–20% TAM reduction → revenue guidance cuts → re-rating. Precedent: Nvidia lost $5B+ in annualised revenue from the A100/H100 China ban in 2023. Early warning: BIS rule-change rumours on Capitol Hill, or Chinese state media escalation language on semiconductors.
Risk 4: Multiple Compression from Rate Resurgence
If inflation re-accelerates (tariff pass-through, commodity spike), the Fed pauses cuts. Discount rates rise; long-duration IT multiples compress fastest. Transmission: 10-year yield +50bps → P/E compression of 3–5 turns → 15–20% price decline even with flat earnings. Precedent: 2022, when the Nasdaq fell 33% as the Fed hiked 425bps. Early warning: 5-year breakeven inflation rate crossing 2.75% — inflation expectations unanchoring.
10. Cycle Playbook
| Phase | Sector Behaviour | Why | What to Own |
|---|---|---|---|
| Early Expansion | Strong re-rating, multiples expand | Rate cuts, risk-on flows | Cloud, high-growth SaaS |
| Mid Cycle | Earnings-driven gains, stable multiples | Revenue beats, buybacks | Semiconductors, mega-cap IT |
| Late Cycle | Divergence: hardware slows, software sticky | Capex budgets tighten | Subscription software, IT services |
| Recession | Severe de-rating; hardware down 40–60% | Capex cancelled, multiples crushed | Cash-generative incumbent software |
| Recovery | Semis lead, then cloud, then services | Inventory re-stocking, capex restart | SOX index plays, foundry names |
Now: We are in late Mid Cycle transitioning to Late Cycle — earnings beats are robust but multiples are already full at 20.9x forward P/E, meaning further gains require earnings growth to do the work rather than multiple expansion. Own cash-generative quality over speculative growth.
11. Structural Themes
Theme 1: Agentic AI Shifting Compute from GPU to CPU+Edge
The first wave of AI was training massive models on GPU clusters — Nvidia's kingdom. The next wave is inference at the edge: running AI agents on laptops, phones, and on-premise servers.
The once-sleepy CPU market has taken off as agentic workloads shift compute needs beyond Nvidia's GPUs.
This accelerates as enterprises prioritise data privacy. Winners: Intel (CPUs, advanced packaging), Qualcomm (mobile inference), edge server vendors. Losers: pure GPU-training plays if inference efficiency improves faster than expected. Position before consensus: buy CPU-exposed names while the market still sees Nvidia as the only AI chip trade.
Theme 2: AI Commoditising IT Services Labour
IT services firms (Infosys, Wipro, Accenture) built empires on human labour arbitrage — cheap offshore coders doing repeatable work. AI agents now write, test, and deploy code faster and cheaper.
Adam Parker of Trivariate Research thinks software companies are "structurally in trouble."
The effect propagates into IT services over 3–5 years as contract renewals reflect lower headcount needs. Winners: platform vendors whose tools do the displacing (Microsoft Copilot, Salesforce Agentforce). Losers: pure-play IT service body shops. Position before consensus: underweight labour-intensive IT services now, while the market still prices them as stable compounders.
12. Portfolio Reference
| Factor | Value |
|---|---|
| S&P 500 weight | ~31–32% (largest sector) |
| Typical dividend yield | 0.7–1.2% |
| Beta vs S&P 500 | ~1.15–1.25 |
| Overweight when | Rate expectations falling; AI capex accelerating |
| Underweight when | Rates rising; recession; valuation >25x forward P/E |
| ETF | Focus | Expense Ratio |
|---|---|---|
| XLK – Technology Select Sector SPDR | Broad large-cap IT | 0.09% |
| SOXX – iShares Semiconductor ETF | Pure semiconductor exposure | 0.35% |
| IGV – iShares Expanded Tech-Software | Software & cloud focus | 0.41% |
13. Three Questions You Should Be Able to Answer
Q1: Why does Microsoft stock sometimes fall on a great earnings report?
A: Because the market is forward-looking and already priced the beat. What it watches instead is the quality signal inside the numbers: Azure growth trajectory, capex guidance, and operating leverage.
Microsoft beat earnings but slipped after reporting $31.9B in capex
— the market read that as future FCF impairment, not just investment confidence. Great EPS today means nothing if $120B/year in capex doesn't yield proportional cloud revenue tomorrow.
Q2: How do interest rates hit IT stocks through a channel most juniors miss?
A: The obvious channel is discount rates compressing long-duration multiples. The missed channel is venture funding: when rates are high for 12–18 months, VC funding dries up, startup formation collapses, and 3–4 years later there are fewer AI startups competing with incumbents. That reduces disruption risk for mega-cap IT. Conversely, rate cuts today are seeding the disruptors of 2029.
Analysts call for earnings growth of 21.3% through 2026
— but the competitive map underneath that consensus is quietly shifting.
Q3: What is the bull and bear case for IT given today's macro?
A: Bull: AI earnings power is real and durable —
IT sector blended earnings growth hit 46.3% in Q1.
Capex builds a moat around hyperscalers. Rates trending down re-rate multiples. Bear:
The forward P/E of 20.9x is already above historical averages
; every good news is priced in.
OpenAI may be on track to miss key revenue and user targets
— the demand thesis cracks if AI monetisation disappoints. The view flips bearish when two consecutive quarters show Azure/AWS growth decelerating below 25%.
Research via live web search | Saturday, May 02, 2026 | GICS Rotation Series