Valuing AI and Technology Companies in Malaysian M&A

The standard EBITDA multiple framework that works well for F&B, manufacturing, and services businesses breaks down when applied to technology companies. This is not a weakness in the framework — it is because technology companies, particularly SaaS and AI-driven businesses, have fundamentally different economic structures. Their value lies in recurring revenue, software gross margins, customer retention dynamics, and proprietary intellectual property. None of these are well-captured by a trailing EBITDA multiple.

As more Malaysian technology businesses reach the scale and maturity that makes acquisition attractive, buyers and sellers need a framework for valuing them accurately.

Why Tech Companies Are Valued Differently

The fundamental difference between a technology company and a traditional service business is the relationship between revenue and cost at scale. A restaurant group with ten outlets cannot serve its tenth outlet without employing comparable labour and paying comparable rent to its first outlet — revenue and cost scale roughly together. A SaaS company that builds its software once can serve its hundredth customer at near-zero marginal cost — gross margins expand as the customer base grows.

This means that a technology company in growth mode may have low or negative EBITDA while being economically valuable — the losses today are an investment in a high-margin revenue base tomorrow. Applying an EBITDA multiple to a growth-stage SaaS business produces a nonsensical result: zero or negative EBITDA generates zero or negative enterprise value, which clearly does not reflect reality.

Conversely, a profitable technology business that is no longer investing in growth may show strong EBITDA while its ARR growth has stalled and customer churn is rising — EBITDA-based valuation will overstate value relative to the business's trajectory.

The Primary Metrics for SaaS and Technology Valuation

Annual Recurring Revenue (ARR). For subscription businesses, ARR is the total annualised value of all active subscription contracts. It is calculated at a point in time — the sum of all monthly or annual contracts annualised. ARR is the primary top-line metric because it represents the predictable, contracted revenue base of the business.

Monthly Recurring Revenue (MRR). ARR expressed monthly. For tracking purposes, MRR movements (new MRR from new customers, expansion MRR from upsells, contraction MRR from downgrades, churned MRR from cancellations) provide the most granular view of revenue health.

Gross Revenue Churn Rate. The percentage of ARR lost each month or year from customer cancellations and downgrades. A gross churn rate of 2% per month means the business loses 24% of its ARR annually from existing customers — which requires significant new customer acquisition just to maintain flat ARR. Best-in-class SaaS businesses run gross churn below 5% annually.

Net Revenue Retention (NRR). NRR captures the total revenue change from existing customers — combining expansions (upsells) and contractions/churn. An NRR above 100% means existing customers are spending more over time, which is the hallmark of a high-quality SaaS business. NRR above 120% indicates a business where growth is embedded in the existing customer base.

Gross Margin. For SaaS businesses, gross margin (revenue less direct cost of delivering the software service — hosting, support, third-party licensing) is typically 65–80%+. This is significantly higher than F&B (50–70%), manufacturing (30–50%), or professional services (40–60%). The high gross margin is what makes SaaS multiples higher than other sectors — more of each revenue dollar reaches EBITDA as the business scales.

If a Malaysian SaaS business has gross margins below 50%, investigate why. It may indicate that the business is being delivered as professional services with software as an output — which is a fundamentally different economic model, valued differently.

ARR Multiples vs EBITDA Multiples

ARR multiples for Malaysian technology businesses in current market conditions generally run 3x–6x ARR, with the specific multiple determined by: growth rate (higher growth = higher multiple); NRR (above 110% commands a premium); gross margin (below 60% attracts a discount); addressable market size; and competition dynamics.

Proprietary Technology vs Off-the-Shelf

One of the most important value differentiators in Malaysian tech acquisitions is the distinction between businesses built on proprietary technology and those built on top of third-party platforms (AWS Marketplace, Salesforce AppExchange, SAP add-ons).

Proprietary technology — where the core software is developed in-house and owned by the company — creates sustainable competitive advantage and attracts higher multiples. The IP is an asset that the buyer acquires and controls.

Platform-dependent technology — where the business's value depends on continued access to a third-party platform under licensing terms that the company does not control — carries platform risk. If the platform operator changes its terms, restricts access, or enters the same market directly, the business's competitive position can be impaired. Buyers discount for this risk.

In due diligence, buyers should obtain: all software development agreements; all third-party licensing agreements with costs and renewal terms; IP assignment agreements from developers and contractors (ensuring IP developed by contractors is assigned to the company, not retained by the individual); and a technical review of the codebase architecture.

Developer Key-Person Risk

Malaysian technology businesses are frequently dependent on two or three senior developers who understand the codebase, manage deployments, and are the knowledge centre for the technology. When these individuals leave — which they may do when an acquisition occurs — the technology operations can stall.

In due diligence, buyers should assess: How well is the codebase documented? Are there comprehensive technical specifications? Is the knowledge embedded in the organisation (in wikis, runbooks, architectural documents) or in individuals' heads? What are the retention plans for key technical staff post-acquisition?

Technical debt — accumulated shortcomings in the codebase from rapid development, shortcuts taken under time pressure, or undocumented legacy systems — also affects value. A technical due diligence (code review by an independent engineering firm) is appropriate for technology acquisitions above RM 5 million in value.

Data Assets and PDPA Compliance

For AI-driven businesses, the training data is often the most valuable asset. Who owns the data? Under what terms was it collected? Does the data include personal data subject to PDPA? Is it properly consented and documented?

AI models trained on customer data may constitute processing of personal data under PDPA 2010. The business should have a legal basis for this processing (consent, legitimate interest, or contractual necessity) documented. A buyer acquiring an AI business that has trained its models on data without proper consent inherits a regulatory liability.

For any Malaysian technology business with data as a core asset, a specific PDPA compliance assessment is warranted in due diligence.

Ongoing Compute Costs

For AI businesses, the cost of running AI models at scale — GPU compute costs, cloud infrastructure, model inference costs — is a significant and growing component of the cost base. Unlike traditional software that runs efficiently on commodity hardware, large language model inference and training can cost RM 50,000–RM 500,000+ per month at meaningful scale.

Buyers should model these costs explicitly: are they currently captured in the cost of goods sold (as they should be), or have they been treated as R&D and expensed below the gross profit line? The answer affects gross margin calculation and therefore the valuation framework applied.