AI in IT: Boon or Curse? Winners, Losers & Investment Dilemma

The debate Kepler Cheuvreux set out in its note—“Is AI truly a boon for the entire IT industry, or could it also be a curse?”—captures a growing fault line across technology markets: enormous, record-breaking infrastructure and R&D spending is colliding with an unsettled monetization pathway and rising investor skepticism. The research house argues that AI has already triggered a sweeping re‑prioritisation of capital across the IT value chain, but that the financial payoff remains uneven and far from guaranteed; its analysis warns of winners and losers depending on how quickly firms adapt and whether AI demand sustains the current scale of investment.

Background: why this question matters now​

The AI era’s headline numbers are hard to ignore. Microsoft publicly announced plans to spend roughly $80 billion in fiscal 2025 to build AI‑capable data centers, a move covered across major outlets and confirming hyperscalers’ arms‑race posture on compute and cloud infrastructure. This single commitment, repeated publicly by Microsoft executives, underscores the capital intensity of modern AI and helps explain why companies from chipmakers to colo providers have re‑rated materially. (cnbc.com)
At the same time, hardware and infrastructure players—notably NVIDIA and recent specialty providers—have been among the clearest beneficiaries of the AI cycle. NVIDIA’s data‑center revenue jumps and market‑cap milestones illustrate how the market is rewarding direct suppliers of compute and acceleration for AI workloads. Those gains have not always filtered evenly to traditional enterprise software vendors or services firms, creating the “winners vs. laggards” story Kepler Cheuvreux highlights. (markets.com)
Kepler Cheuvreux’s central framing—AI as both a massive opportunity and a potential source of structural stress for incumbents—reflects three simultaneous realities:

• Record capex and private funding into AI infrastructure and platforms.
• A scramble by incumbents to embed AI into product roadmaps and go‑to‑market strategies.
• Persistent market questions about whether and how that spending converts into sustainable revenue growth and margins.

---

Where Kepler Cheuvreux is right — the immediate strengths of the AI push​

1. Capital deepening and infrastructure scale are real and transformative​

The pace and size of investments into AI data centers, GPUs, and specialized server stacks are materially larger than typical enterprise IT cycles. Microsoft’s public $80bn FY25 allocation and the cascade of multi‑billion deals with niche GPU cloud providers show companies are building a new physical layer for AI at hyperscaler scale. That creates a moat for companies that can match the capital and ecosystem scale. (cnbc.com)

2. Hardware and core platform winners are being re‑rated​

Companies that supply the raw compute, interconnects, and software stacks for model training and inference have seen outsized valuations and revenue growth. NVIDIA’s record quarters and market‑cap milestones are the clearest indicator that the market currently rewards those with direct exposure to model compute and optimization tools. (investor.nvidia.com)

3. AI forces strategic product rewrites and faster innovation cycles​

Large technology vendors are rewriting product roadmaps to make AI a central selling point—bundled Copilot integrations, agent frameworks, and new enterprise AI SKUs are examples. That creates immediate upselling opportunities (e.g., Copilot bundles sold through partner channels) and forces competitors to accelerate R&D, which can benefit customers through better features and tighter platform integration. (microsoft.com)

4. Productivity and outcome focus can unlock new monetization​

When firms can demonstrate clear time‑savings, automation of routine tasks, or measurable revenue uplift from AI features, customers will often pay a premium. For platform owners that can instrument outcomes and tie pricing to value delivered, AI can be a durable revenue driver—if the pricing and contractual models are right.

---

Where the risk, “curse,” and ambiguity are concentrated​

1. Monetization remains uneven and evidence thin for many software vendors​

Kepler Cheuvreux notes a simple but potent fact: investors are still looking for compelling proof that AI spend translates into long‑term revenue growth. Several large software names have lagged the market even while infrastructure and chip vendors soared, indicating the market differentiates between AI‑enablers and AI‑claims. In practical terms, adding an “AI” label to a product is insufficient—vendors must show measurable, recurring revenue tied to AI features and defensible economics.

2. Overinvestment and the risk of stranded capital​

The industry is building data centers, buying GPUs, and signing years‑long supplier contracts at a previously unseen pace. If enterprise adoption stalls, if regulatory constraints materially raise the cost of operation, or if a dominant user scenario fails to emerge, there is a real risk of underutilized assets—empty racks and idle GPUs reminiscent of past boom busts. Kepler Cheuvreux explicitly warns about this overbuild risk, and independent commentators have echoed similar concerns.

3. Services margin compression from AI‑driven efficiency​

IT services companies are experiencing a double dynamic: demand for AI transformation work is high, yet AI tools can automate portions of the work that traditionally justified high bill rates. The net effect is upward pressure on demand but downward pressure on pricing per unit of labor. Public comments from large consultancies and coverage by major outlets show these firms are re‑tooling, reorganizing and sometimes facing headwinds as procurement shifts to outcome‑based or agent‑augmented delivery models. (reuters.com)

4. The “SaaS is dead” overreach and narrative risk​

Kepler Cheuvreux rejects dystopian claims that AI ends traditional software models, but it accepts that business models must evolve. The phrase “SaaS is dead”—originating from a provocative description by Microsoft’s CEO about the agent era—has been widely discussed and misinterpreted across the industry. It’s better read as a warning: standalone, monolithic business apps that don’t embed AI, integrate broadly, or prove outcomes may lose relevance. But the wholesale death of SaaS is an exaggeration; SaaS will evolve into more composable, AI‑first experiences. (windowscentral.com)

---

Detailed implications for each part of the IT value chain​

Software vendors (ISVs): urgency + monetization challenge​

• Immediate pressure: Embed meaningful AI capabilities into core workflows, not as cosmetic features.
• Opportunities: New premium SKUs (agents, copilots, automated insights) and outcome‑based pricing.
• Risks: Failure to show measurable ROI will mean muted investor response and customer churn to AI‑native competitors.
• Action checklist for ISVs:
• Re‑architect critical flows to make AI the default path for users.
• Build telemetry to quantify AI impact and enable outcome‑linked pricing.
• Assess compute dependency: partner where capex is prohibitive.

IT services and consultancies: demand growth but margin risk​

• Immediate pressure: Clients want AI but also expect efficiency and lower TCO.
• Opportunities: High‑value advisory and transformation projects, IP and productization of AI accelerators.
• Risks: Classic labor‑arbitrage revenue models will be squeezed by automation and AI agents that can replace predictable tasks.
• Recommended moves:
• Productize AI assets to move from hourly billing to subscription/IP revenue.
• Upskill client delivery teams and create hybrid human+agent pricing.
• Be selective: focus on industry vertical problems where human oversight remains essential.

Infrastructure and hardware vendors: near‑term winners, long‑term moat questions​

• Immediate advantage: Providers of GPUs, specialized AI servers, and optimized software stacks enjoy clear demand and re‑rating.
• Long‑term questions: Do proprietary stacks (CUDA, specialized interconnects) create sustainable switching costs, or will commoditization and geopolitics fragment this market?
• Strategic imperatives:
• Double down on software value (developer tooling, libraries, blueprints).
• Build supply resilience (geographic diversification, alternative foundries).

Resellers and channel partners: limited direct disruption, but new upsell opportunity​

Kepler Cheuvreux suggests resellers will see limited direct impact but can benefit from the sales motion around AI features (for example, Microsoft Copilot bundles and partner programs). Microsoft’s partner playbook—promotions, Copilot seat benefits, and partner enablement—illustrates how channel partners can monetize AI through bundling, services, and go‑to‑market offers. Channel programs and discounts give partners routes to repackage and upsell AI to existing customers rather than rely on hardware contracts alone. (microsoft.com)

---

The investor’s angle: why valuations diverge and what to watch​

Kepler Cheuvreux highlights a market that is rewarding visible, immediate AI cash flows (infrastructure, chips, cloud) while penalising firms that have AI messaging without proven monetization. This creates:

• A valuation premium for hardware and cloud providers with observable AI revenues.
• A valuation discount or at least skepticism for legacy software firms until they can demonstrate sustainable AI earnings lift.

Key indicators for investors:

• Realized revenue from AI SKUs and ARR tied to AI features.
• Capex utilization rates and long‑term data‑center occupancy.
• Customer retention and net dollar retention for products with AI embedded.
• Regulatory/tax impacts (e.g., energy costs, data governance) that can change economics quickly.

Independent academic and market studies warn of valuation misalignment risk—markets can over‑price future AI payoffs that may never materialize. That’s the “bubble” argument that should temper unbridled optimism. (arxiv.org)

---

Practical recommendations for CIOs, CEOs, and partners​

• Prioritize projects with measurable outcomes. Demand proof points (productivity delta, revenue lift, cost reduction) before scaling.
• Revisit contracting models: insist on pilots that include metrics and staged rollouts to avoid overcommitting to long‑term capex.
• Invest in governance and security early. Regulatory and ethical requirements are moving from “nice to have” to procurement must‑haves.
• Develop a hybrid pricing strategy: blend subscription for core software with outcome‑based or agent‑usage pricing where applicable.
• For services firms: productize repeatable AI solutions to offset labor price pressure and preserve margin.

---

Where Kepler Cheuvreux may overstate or where claims need caution​

• The metaphor “SaaS is dead” is provocative, but it is more useful as a call to action than as a literal prediction. Public transcripts and industry analysis show Nadella’s comments were intended to highlight architectural shifts (agents and AI tiers) rather than announce the demise of software as a procurement model. Treat the phrase as strategic urgency—not an extinction event. (windowscentral.com)
• Any single company’s capex forecasts and multi‑year commitments (including Microsoft’s $80bn FY25 spend) are factual and verifiable through company posts and major press coverage, but the ultimate ROI on those expenditures remains an empirical question—dependent on adoption curves, pricing, and regulatory environment. Investors and executives should treat capex announcements as necessary context, not proof of immediate return. (cnbc.com)
• Claims that every player must “act swiftly and decisively” are sensible strategically, but not every company needs to chase every AI frontier. For many established enterprises, measured integration—prioritizing risk, governance, and actual business outcomes—will outperform reckless, capex‑heavy expansion.

---

Conclusion — a nuanced verdict: boon with conditionality​

AI is neither a pure boon nor an inevitable curse for the IT industry; it is a force multiplier that magnifies strategic competence and execution differences across firms. Where AI creates transparent, trackable business value and where companies own the critical technology stack (compute, platform, developer tooling), the economic upside is real and already being rewarded by markets. Conversely, where investment outpaces adoption, where companies fail to translate features into measurable outcomes, or where regulatory and operating costs rise, the same AI wave can leave stranded assets and compressed returns.
Kepler Cheuvreux’s assessment is a practical reminder: the AI era accelerates already existing market selection dynamics. Companies that treat AI as a tool to deliver measurable outcomes, retool their pricing and delivery models, and partner intelligently to manage capital intensity will likely prosper. Those that adopt the language of AI without the discipline of outcome measurement risk becoming casualties of a technology revolution that rewards proof over promises.

---

Bold, timely choices will determine who rides the AI wave and who is washed ashore—AI is an extraordinary opportunity, provided the industry remembers that scale without monetization and governance is the precise definition of risk.

---

Source: Investing.com Is AI truly a boon for the entire IT industry, or could it also be a curse? By Investing.com