Research · AI investment themes
AI investment themes
AI investment themes span infrastructure, applications, automation, data, developer tools and public-market platform companies. The research challenge is separating durable market structure from hype: where budgets move, where value accrues and which categories become crowded or absorbed by platforms.
Compute and data centres
Compute, data centres, accelerators, power and networking form the most capital-intensive AI theme. These layers have clear demand but require scale, financing and supply-chain access.
Research should distinguish between owners of scarce infrastructure, users of scarce infrastructure and software companies whose margins depend on infrastructure prices falling.
Model tooling
Model tooling includes deployment, routing, evaluation, observability, guardrails, prompt management, testing and cost control. The buyer is often an engineering or platform team trying to make model use reliable.
This theme can support durable companies when the product becomes a control plane. It is weaker when it is a temporary wrapper around one model provider.
Context and retrieval
Context engineering and retrieval connect models to private data, permissions, files, code, customer histories and operational state. This layer matters because generic models become useful when grounded in the right evidence.
Important questions include data freshness, access control, ranking quality, auditability and whether the retrieval layer becomes a product moat or a background capability.
Vertical AI applications
Vertical AI applications target specific workflows in legal, finance, healthcare, industrial operations, sales, support, engineering and back office work. They can look more like labour systems than traditional SaaS.
The investment question is whether the product owns an outcome, replaces a task, expands a workflow or remains a thin assistant. Durable vertical AI usually needs domain context, integration depth and measurable work quality.
Automation
Automation themes include agentic workflows, RPA replacement, industrial automation, AI-enabled services and operating systems for repeat work. The economic promise is labour leverage, but execution risk is high.
Research should test reliability, exception handling, human review, integration burden and accountability. A demo can look complete while the production workflow still fails at the edge cases.
Hype risk
AI markets can overheat quickly. Durable research distinguishes usage from revenue, pilots from deployment, gross revenue from margin, and platform dependency from defensibility.
This overview is general research only. It is not a recommendation to invest in any AI company, theme or security.
How to use this research
Use this AI investment themes research as a map of the relevant market structure, not as a prediction engine. The point is to clarify categories, buyers, public comparables, private-company signals and unanswered diligence questions before drawing stronger conclusions.
For searches around AI investment themes, the useful output is a working view of the category: what belongs in the market, what should be excluded, which company types are public or private, and which signals deserve repeat monitoring.
Signals to monitor
Useful signals include new company formation, funding rounds, hiring patterns, customer evidence, product launches, public-company commentary, partnership activity, secondary-market indicators and changes in valuation tone across comparable categories.
Signals should be dated and sourced. A funding announcement, website claim or public-market multiple can be useful, but it should be treated as one piece of evidence rather than a complete view of quality, durability or risk.
Public and private evidence
Public-market evidence helps frame margins, growth expectations, valuation cycles, platform power and investor appetite. Private-market evidence helps identify category formation, founder activity, product direction and emerging customer demand before it appears in listed-company results.
The research task is to hold both evidence types together. A private company can look compelling until public comparables show weak economics; a public company can look mature until private-company formation reveals a new competitive edge.
What to avoid
Avoid treating AI investing themes, AI infrastructure investing, AI companies market map as a slogan. Durable research should define the market, separate adjacent categories, record assumptions and keep uncertainty visible. It should also distinguish company marketing from independent evidence.
This page is designed for general research context. It does not claim that Meridian manages capital, advises on securities, operates a fund, has a portfolio or offers personalised financial advice.
Next research questions
The next useful step for AI investment themes is usually a tighter company universe: named categories, inclusion rules, source links, confidence levels and a dated view of public comparables. That turns a theme into a working research asset rather than a broad narrative.
The second step is continuous monitoring. Categories change when budgets shift, platforms absorb features, financing markets reopen, regulation changes or customer workflows mature. A serious market map should be revisited as those signals move.
FAQs
What are AI investment themes?
They are market categories and patterns around AI infrastructure, applications, automation, data and platform shifts.
Why is AI infrastructure a major theme?
AI infrastructure controls compute, deployment, data flow, monitoring and cost structure.
What is the biggest AI investing risk?
Hype, crowding, weak defensibility, platform dependence and unclear unit economics are common risks.
Disclaimer
Research content is for general information only and should not be treated as investment, legal, tax or financial advice.