Interactive figure

The same market, three sizes

Juniper Research34

> US$34bn by 2026

Mordor Intelligence21.3

US$21.27bn in 2026

Gartner15.7

US$15.7bn in 2025, US$27.4bn by 2029

Estimate for

Juniper Research: > US$34bn by 2026. Source. These firms are not contradicting each other. They are counting different things — some include the carrier fees passed through a platform, some count only the platform's own revenue. Ask what is inside a number before you plan against it.

Each bar is a published estimate. None is wrong. They are answering different questions.

The published positions#

Juniper Research projects the CPaaS market surpassing US$34 billion by 2026, up from around US$30 billion in 2025.

Gartner forecasts spending at US$15.7 billion in 2025, reaching US$27.4 billion by 2029 — roughly 14% compound annual growth.

Mordor Intelligence sizes 2026 at US$21.27 billion, reaching US$41.05 billion by 2031 at about 14% compound growth.

The 2026 figures span from roughly fifteen to thirty-four billion dollars. That is not a rounding difference; it is more than a factor of two.

Three reputable firms, three figures for the same year, differing by more than a factor of two. None of them is wrong.

CPaaS market size for 2026, as published

34Juniper Research

Over US$34bn by 2026, up from about US$30bn in 2025

21.27Mordor Intelligence

US$21.27bn in 2026, reaching US$41.05bn by 2031

15.7Gartner

US$15.7bn in 2025 spending, growing to US$27.4bn by 2029

A 2.2× spread between the highest and lowest. That gap is definitional, not a quality difference between the firms.

Where the divergence comes from#

Carrier pass-through. The largest single factor. A2P messaging revenue includes substantial fees that platforms collect and remit to operators. Count gross and the market is very large; count net platform revenue and it is much smaller.

Category boundaries. Whether contact-centre software, email delivery, video SDKs and verification services sit inside or outside "CPaaS" changes the total materially, and there is no industry-standard answer.

Spend versus revenue. Buyer-side spending includes things vendor-side revenue does not, and the two are measured from opposite directions.

What each firm had to decide before counting anything

  1. Carrier pass-through

    Whether the fee a platform collects and hands to an operator is market revenue. Include it and the number roughly doubles.

  2. Category boundary

    Is a contact-centre suite CPaaS? Is email delivery? Is a video SDK? Each line drawn is a different market.

  3. Spend or revenue

    Buyer-side spending and vendor-side revenue are different quantities and diverge systematically.

  4. Geography and segment

    Which regions and which customer sizes are inside the count at all.

  5. Forecast base year

    A different starting point compounds into a very different figure five years out.

Five decisions, each with two or three defensible answers. That is enough to produce every figure above without anyone being careless.

The agreement is the signal#

All three firms independently project sustained double-digit compound growth. That agreement, across different definitions and methodologies, is a far stronger finding than any of the individual totals.

When independent methods disagree on magnitude but agree on direction, direction is the robust result.

How to use this in a decision#

Never quote a single figure without its scope. "The CPaaS market is $34 billion" is not a fact; it is a fact with its conditions removed.

Match the figure to your question. Sizing your addressable revenue as a platform? Net revenue is the relevant number. Sizing what enterprises spend on this capability? Gross spending is.

Distrust any deck that shows only the largest number. Not because it is dishonest — the number is published and real — but because the selection tells you what the deck is for.

The methodological lesson, which generalises#

This pattern recurs everywhere in this field. Deliverability rates, containment rates, engagement benchmarks: wide published ranges usually indicate unstable definitions rather than genuinely variable performance.

When you meet a benchmark with a wide range, the first question is not "which end am I at?" It is "what are they counting?" — and the answer is frequently that they are counting different things and calling them the same name.

What to take away#

Three firms, one market, a 2.2× spread, and no error. Read the agreement rather than the level, cite the boundary alongside the number, and treat any figure whose methodology is not available as decoration.

Sources

Every claim worth checking, with somewhere to check it.

  1. Forecast Analysis: Communications Platform as a Service, WorldwideGartner
  2. Market Share Analysis: Communications Platform as a Service, Worldwide, 2024Gartner
  3. CPaaS Market Research Report: Size, Share, Trends 2025-29Juniper Research
  4. Communication Platform-as-a-Service MarketMordor Intelligence