CA API Management · Exit

What CA API Management migration economics actually look like in 2026.

The destination quotes for Kong, Apigee, and Mulesoft against CA API Management look very favourable on the surface. The all in economics tell a more complicated story, and the parts of the story buyers underestimate are the parts that drive the closing math.

By The Negotiation Desk · CA · Archetype: The Exit

Published 7 Feb 2026

CA API Management is one of the older API gateway products in the enterprise market and is the one we see most frequently evaluated for exit in 2026. The buyer's typical motivation is a renewal quote that has run well above the inflation band for two or three cycles in a row, often coupled with the perception that the product roadmap has slowed since the Broadcom transaction. The buyer asks a small set of destination vendors to quote a replacement. The opening destination quotes typically run 40 to 60 percent below the CA API Management renewal on a like for like throughput basis. The buyer reads the gap as the saving. That reading is wrong. The all in economics across a 36 month window land somewhere between break even and a 25 percent saving in our 2024 to 2026 sample, and the variance inside that band is wider than the headline number suggests.

The purpose of this article is to set out the components of the all in math that buyers under count, the typical magnitudes we see on each component, and the secondary value of a credibly priced alternative even when the buyer ultimately chooses to stay. The intent is not to defend CA API Management. The product has real limitations and the renewal economics in 2026 are not friendly. The intent is to make the migration math defensible so that the buyer's decision is grounded.

What the comparison usually shows on paper

The opening like for like comparison frames the destination platform's per transaction or per gateway price against the CA renewal. Kong, Apigee, and Mulesoft are the three destinations we see most often. The price band on the destination, in our sample, ran 38 to 57 percent below the CA renewal on a comparable throughput envelope. The number is real on the surface. It is also not the full comparison. The CA contract typically carries scope that the buyer has not been consuming. The destination quote is scoped to the throughput the buyer says it wants, which is the consumed throughput. The two are not equivalent. Normalising the comparison shrinks the headline by 10 to 20 points.

The migration cost the buyer underestimates

The largest cost line in the migration is the policy and integration rebuild. CA API Management deployments accumulate policies over five to ten years. Authentication, rate limiting, transformation, routing, monetisation policies have been built into the gateway with deep dependencies on the surrounding integration estate. A meaningful subset of those policies does not transfer cleanly to the destination platform. The rebuild is engineering work. The labour cost runs $1.4M to $4.2M in our sample for a buyer with roughly 800 to 1,600 published APIs. The variance is driven by the complexity and documentation quality of the existing policy set.

"The destination price is the surface. The policy rebuild, the parallel run, and the integration test cycle are the body of the work. Buyers who arrive at the migration with a policy inventory complete the move profitably. Buyers who do not, do not."CA API Management Engagement Lead, The Desk

The second cost line is the parallel run. The migration cannot be done in a single weekend cutover. The two gateways run side by side while consumer applications are repointed one API at a time. The parallel run in our sample ran six to fifteen months. The dual licensing cost during the parallel run is material. Most buyers also carry duplicated operational tooling during the period, which adds an operations cost the migration budget rarely captures.

The third cost line is the integration test cycle. APIs that have been in production for several years are consumed by application code that the API team often does not have full visibility into. The cutover risk runs through whether the destination gateway behaves identically to the CA gateway under every consumer call pattern. The test cycle to validate parity is non trivial and is rarely budgeted as a discrete line. Buyers underestimate the cycle by 35 to 60 percent on the first pass through the migration plan.

What the all in math looks like on a 36 month basis

When we run the all in math on the buyers in our sample over a 36 month window, the migration moves from a 38 to 57 percent surface saving against the CA renewal to a 0 to 25 percent all in saving against the same baseline. The factors that push a buyer to the higher end of the band are policy documentation quality, a tightly scoped consumer base, and engineering capacity available to run the rebuild on the original timeline. The factors that push a buyer to the lower end are sprawling consumer integration, documentation gaps, and the buyer's tendency to discover during the migration that the CA gateway is doing more than the policy inventory captured.

The secondary value buyers underestimate

The secondary value of a priced alternative is often larger than the primary value of the migration. The buyer who arrives at the CA renewal with a credible Kong, Apigee, or Mulesoft quote, a documented policy rebuild estimate, and a board level statement that the migration is on the table moves the CA renewal discount band materially. In our 2025 to 2026 sample the median further reduction available on the CA renewal when the alternative had been credibly priced was 19 to 28 percent on top of whatever else the buyer had been negotiating. The Broadcom deal desk treats a credibly priced alternative differently than a renewal in place. The credibility of the alternative does not require a migration decision. It requires the work to be visible.

The numbers

CA API Management migrations and evaluations reviewed 2024 to 202611

Median surface saving on destination quote (Kong, Apigee, Mulesoft)38 to 57%

Median policy rebuild labour cost$1.4M to $4.2M

Median parallel run duration9 months

Median integration test cycle underestimate on first plan pass35 to 60%

Median all in 36 month savings range0 to 25%

Median further discount available on CA renewal when alternative is credibly priced19 to 28%

What we have seen on live deals

A global insurer evaluated a Kong migration in late 2024. The destination quote was 49 percent below the CA renewal on like for like throughput. The buyer engaged us to read the all in math before signing the destination contract. The policy inventory produced 1,140 published APIs across 280 active policies. Roughly 70 percent of the policies transferred cleanly to Kong. The remaining 30 percent required rebuild. The all in math, including a 12 month parallel run and a $3.4M policy rebuild, produced a 14 percent saving against the CA baseline over 36 months. The buyer used the priced alternative to renegotiate the CA renewal and closed at 31 percent below the original opening quote. The total economic outcome was better than the migration would have produced.

A government entity completed a Mulesoft migration in 2023 and is now 36 months into the post migration operating period. The 36 month all in saving against the projected CA baseline has landed at 22 percent, which is at the high end of the band we see in our sample. The factors that produced the better outcome were a small consumer integration footprint, a well documented policy set, and an engineering team that was already trained on the destination platform from an adjacent project. None of those factors are universally present. The case is a useful upper bound rather than a typical expectation.

A national bank ran the same evaluation against an Apigee quote and chose to remain on CA API Management. The internal policy inventory exercise identified that the bank's gateway estate carried 14 percent of policies it no longer needed and that a CA gateway consolidation could be negotiated against the renewal. The reorganisation produced a 17 percent renewal reduction without any migration. The bank captured the secondary value of the alternative pricing exercise without paying the migration cost.

The takeaway

The surface destination quote on a CA API Management migration almost always overstates the all in saving. The policy rebuild, the parallel run, and the integration test cycle between them convert a 38 to 57 percent headline saving into a 0 to 25 percent all in saving over 36 months in our 2024 to 2026 sample.
The variance inside the all in savings band is driven by the documentation quality and complexity of the existing CA policy set. Buyers with a tightly scoped, well documented gateway complete the move profitably. Buyers with a sprawling, undocumented gateway often do not.
A credibly priced alternative typically delivers more value as a negotiation position against the CA renewal than as an executed migration. In our sample the median further discount available when the alternative had been priced was 19 to 28 percent, on top of whatever else the buyer was negotiating.