AI will not end the software world, but it will reshape the value distribution across cloud and SaaS

Overview

This report discusses the 5+ year impact of generative AI and subsequent AI technology waves on the global software and cloud computing industries, with the core time anchor set around 2030. It analyzes the software stack by layer, focusing on IaaS/PaaS, databases, developer tools, and SaaS, and asks whether AI will weaken the long-term value of hyperscale cloud providers and enterprise application software. The conclusion is that the software world will not remain static in its current form, but it will also not be eliminated wholesale by AI. Value will shift away from software that is non-innovative, low-complexity, and purely utility-like toward vendors with data, platforms, workflows, and deep domain expertise.

Core views

The report argues that IaaS/PaaS remains relatively early in the cloud migration cycle, and AI will drive more critical workloads to the cloud, expanding cloud vendors' TAM. GPUs, ASICs, and dedicated inference chips will increase infrastructure usage, and as Agentic AI becomes more widespread, demand for CPUs, databases, storage, and networking will also rise. On the SaaS side, AI will bring more personalization, automation, and workflow redesign, and enterprise applications will become the management layer for complex AI agents and business processes. The bearish concerns about development automation and value transfer to model companies do affect some submarkets, but the report believes the view that AI will broadly replace enterprise software is too pessimistic.

Analysis framework

The report uses a framework that combines technology-cycle analysis with a layered software-stack view. It first uses the past 70+ years of technology disruption cycles to show that major technology diffusion typically takes 10 to 20 years, then takes the 2023 ChatGPT breakthrough as the starting point for AI's commercial impact and judges that most of the major effects will become clearer by 2030. It then analyzes IaaS, PaaS, SaaS, and developer tools layer by layer, comparing AI's impact on TAM, margins, customer stickiness, data location, infrastructure complexity, and incumbent moats.

Methodology notes

• Technology cycle analysis10- to 20-year major technology disruption cycle

  Compare the diffusion pace of AI with the commercialization pace of cloud computing, SaaS, and AWS

  The report argues that even if generative AI adoption is very fast, deep restructuring of enterprise software and cloud infrastructure will still take years. 2030 is presented as a reasonable mid-cycle checkpoint for assessing the major effects, with the full impact potentially extending to around 2035.

• Industry stack layeringIaaS/PaaS/SaaS software stack framework

  Break down AI impact by infrastructure, platform, and application layers

  IaaS/PaaS are more likely to benefit from expanding demand for compute, databases, storage, and platform software, while SaaS depends on application complexity, deterministic workflows, domain knowledge, and vendor innovation speed.

• Competitive moat analysisData location and workflow complexity moat

  AI works best where applications and enterprise data already reside

  The report argues that AI is not a standalone technology stack; it is tightly integrated with traditional databases, applications, governance, security, and platform software. As a result, cloud and application vendors with enterprise data, customer relationships, and complex business processes have structural advantages.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Oracle / ORCL

  The report repeatedly cites Oracle as one of the beneficiaries in cloud, database, and SaaS

  Strengths

  Oracle's database migration to the cloud, OCI, Oracle Alloy, sovereign cloud, and PaaS capabilities are seen as supportive of demand capture in AI and deglobalization environments.

  Weaknesses

  Cloud competition still comes from large peers such as AWS, Azure, and GCP, and the economics of AI infrastructure and the pace of customer migration remain uncertain.

  Comparison

  Compared with most regional or smaller IaaS providers, the report views Oracle as having an advantage in white-label OCI, Alloy, and full PaaS capabilities; within SaaS, Oracle is also positioned alongside SAP as a relatively better-positioned vendor.

  Risks

  Margin compression in AI infrastructure, lower customer stickiness, value migration to model or developer-tool layers, and sovereign cloud partnerships not landing as expected.

• Microsoft / Azure

  The report identifies Microsoft as one of the major hyperscale cloud and AI infrastructure beneficiaries

  Strengths

  Its large-scale cloud infrastructure, enterprise customer base, and data and application ecosystem help it absorb AI workloads.

  Weaknesses

  Deglobalization and sovereign cloud demand may require more complex local partnership models and architectural adaptation.

  Comparison

  Along with AWS and GCP, it is one of the world's top three cloud providers; the report says the top three together account for roughly two-thirds of the IaaS market.

  Risks

  Pricing pressure after AI capacity supply and demand normalize, regional sovereign cloud competition, and the commoditization narrative around infrastructure.

• Leading SaaS vendors

  AI will reshape enterprise application software rather than fully replace it

  Strengths

  Complex deterministic workflows, domain knowledge, semantic data relationships, customer relationships, and ongoing maintenance capabilities form a moat.

  Weaknesses

  Vendors that innovate slowly, fail to deeply embed AI into applications, or mainly offer low-complexity functionality are more vulnerable to replacement.

  Comparison

  ERP is the most protected, followed by HCM and full-suite CRM; pure on-premises deployments and low-differentiation software are more fragile.

  Risks

  AI-native applications, automation tools, and new technology waves compress the value of traditional SaaS functionality.

Key data

• Key time frameAround 2030The report uses this as an observation point roughly 7 years after ChatGPT's breakout to assess AI's main impact on the software stack.
• Long-term technology cycle10-20 yearsThe report says major technology disruption cycles typically span 10 to 20 years.
• Cloud penetration referenceSome software submarkets have already reached 70%+ cloud penetrationThe report uses this to explain why growth deceleration in some SaaS and cloud software markets may be driven by the maturity of cloud migration.
• Beneficiary directionIaaS/PaaS, databases, CPU/GPU/ASIC usage, storage, and Agentic AI-related platformsAI workloads combined with enterprise data will create greater demand for infrastructure and platform capabilities.
• Relatively protected SaaS areasERP, HCM, and full-suite CRMThe report believes these areas have moats rooted in complex workflows and domain knowledge, provided vendors continue to innovate and execute.

Impact & implications

The investment implication is to avoid reducing AI's impact to a simple narrative that software as a whole will be replaced. The report instead emphasizes structural divergence: hyperscale cloud providers and database platforms may benefit from AI-driven TAM expansion; SaaS leaders that can innovate quickly and embed AI into core workflows may still remain competitive; while vendors with low complexity, slow innovation, purely on-premises deployment, or weak data and workflow moats face a higher risk of being displaced.

Risks

• Once AI infrastructure supply and demand normalize, compute prices and cloud vendor margins may come under pressure.
• AI development tools may become commoditized, hurting the traditional developer tools market and some PaaS functions.
• Generative AI may later be partially displaced by new technologies such as inference, Agentic AI, First-principles AI, or Quantum AI, causing current investment themes to become obsolete too early.
• If SaaS vendors fail to innovate quickly enough, they may be displaced around 2030 by faster AI-native or AI-augmented competitors.
• Cloud migration in some software submarkets is already approaching maturity, so slower growth may continue to affect valuation expectations.

What to watch

• Whether enterprise AI workloads continue to concentrate on platforms that own both data and applications.
• The real incremental demand for CPUs, GPUs, ASICs, databases, and storage as Agentic AI becomes more widespread.
• The pace of execution for Oracle Alloy, OCI, and sovereign cloud partnerships in Europe, the Middle East, and other regions.
• Whether enterprise applications such as ERP, HCM, and CRM can turn AI into workflow automation and customer retention rather than just an add-on feature.
• Whether the developer tools market stabilizes or continues to rotate rapidly as AI coding and agent-based development platforms replace vendors.
• Margins, customer stickiness, and pricing trends in AI-related cloud infrastructure businesses.