SAP HANA migrations are structurally different from every other workload class in cloud migration. The SAP licence cost dominates total cost of ownership. The SI partner labour dominates one-time migration cost. The downtime tolerance constrains the cutover approach. And the SAP commercial relationship (Enterprise Agreement, Universal Subscription, RISE with SAP) determines the destination and the cost in ways no other workload does. This page is the 2026 cost reference for SAP HANA migration to cloud.
SAP HANA workloads can only run on SAP-certified compute. Each hyperscaler publishes a certified instance list maintained by SAP themselves on the public SAP Note 2235581. The certified instance families on each hyperscaler:
SAP-certified HANA instance families (May 2026)
| Hyperscaler | Instance family | Max HANA memory | Approx on-demand $/hour (US East) |
|---|---|---|---|
| AWS | X2iedn (24xlarge) | 4.1 TB | $33.00 |
| AWS | X2iedn (32xlarge) | 8.2 TB | $44.00 |
| AWS | U-7i (high-memory) | 24 TB | $60 to $200 |
| Azure | M192ms_v2 | 4.1 TB | $30.50 |
| Azure | M832ixs_v2 | 12 TB | $95.00 |
| GCP | m2-megamem-416 | 5.7 TB | $31.20 |
| GCP | m3-ultramem-128 | 3.9 TB | $30.50 |
Pricing varies by region and changes regularly. The headline rates are broadly comparable across hyperscalers on like-for-like memory sizes. Reserved instance and savings plan discounts of 30 to 50 percent are available on commitment, which is the typical pattern for production HANA workloads with stable memory requirements. Spot or preemptible instances are not suitable for production SAP HANA.
RISE with SAP is SAP's bundled offer that includes the S/4HANA Cloud Private Edition licence, the underlying hyperscaler infrastructure, the SAP-managed application services (HEC, the Hyperscaler Edition Cloud), and basic-tier support. Customers select the hyperscaler (AWS, Azure, or GCP) at contracting; SAP manages the infrastructure on the customer's behalf and bills a single per-FUE per-month subscription rate.
The commercial implication is significant. Customers on traditional perpetual SAP licences plus maintenance pay a one-time licence fee plus 22 percent annual maintenance. RISE converts that to a per-FUE per-month subscription, typically at 30 to 50 percent of the equivalent annualised cost in year one but converging over years two and three. The hyperscaler infrastructure is bundled, so the customer does not see a separate hyperscaler bill. The trade-off is loss of control over the infrastructure layer and lock-in to SAP's managed service tiers.
For migration cost modelling, RISE simplifies the model by removing the SAP licence commercial change from the migration programme. The migration becomes a lift to RISE with SAP, paid through a different mechanism. The migration partner labour is unchanged. The one-time conversion cost from ECC to S/4HANA (where applicable) is unchanged. The headline saving on RISE versus self-managed S/4HANA is typically 15 to 25 percent over a 3-year horizon for customers who do not have specialised SAP operations capability in-house.
RISE versus self-managed
Most SAP customers still run ECC (the predecessor to S/4HANA) on premises or in a managed hosting environment. SAP's mainstream maintenance for ECC ends in 2027, with extended maintenance available to 2030 at a premium. The forcing function for cloud migration is this maintenance cut-off, which makes the migration decision a question of when and how, not whether.
Two paths exist. Lift-and-shift moves ECC to a hyperscaler with no functional change; the customer continues to run ECC on certified compute and storage with no SAP version change. This path is fast (9 to 18 months) and relatively low risk but defers the S/4HANA decision. Conversion moves ECC directly to S/4HANA on the same hyperscaler, with the data model conversion (Simple Finance, Simple Logistics) and custom code remediation done as part of the migration. This path is slow (18 to 36 months) and higher risk but produces a steady-state S/4HANA environment.
The cost difference is material. Lift-and-shift typically costs $1.5M to $6M for a mid-enterprise estate. Conversion typically costs $4M to $20M. The conversion premium comes from the application validation, custom code adaptation, end-user retraining, and the longer parallel-running window. Many enterprises do lift-and-shift first then conversion second, spreading the cost across two programmes. This pattern has lower peak cost and lower risk concentration but adds the cost of running the lift-and-shift environment in the gap year.
SAP migrations are almost always SI partner-led. The big four (Accenture, Deloitte, EY, KPMG) and the SAP specialist firms (Capgemini, NTT Data, T-Systems, Wipro, TCS, Infosys) dominate the partner landscape. Day rates for SAP-certified consultants typically run $1,200 to $2,400 per day for offshore-blended teams and $2,000 to $3,800 per day for onshore-only teams in the US and UK.
A typical mid-enterprise SAP HANA migration team comprises 15 to 35 SI consultants across Basis, Functional, Custom Development, Integration, and Programme Management roles, over 12 to 24 months. The blended cost typically lands at $3M to $12M of SI labour, depending on team mix, location strategy, and scope. The largest line item in any SAP migration is almost always SI labour, often exceeding all other costs combined.
A representative cost build for a mid-enterprise SAP migration: ECC to S/4HANA on Azure (RISE with SAP commercial path), 8 TB HANA database, single production landscape, 18-month programme, blended onshore-offshore SI team.
Worked SAP HANA migration cost build, ECC to S/4HANA on Azure RISE, 18 months
| Cost line | Low estimate | Typical estimate | High estimate |
|---|---|---|---|
| Assessment, blueprint, project preparation | $400,000 | $700,000 | $1,200,000 |
| SI partner labour, blueprint to go-live | $2,800,000 | $5,500,000 | $11,000,000 |
| Custom code remediation (Z-program rework) | $300,000 | $700,000 | $1,800,000 |
| Integration rework (PI/PO to BTP, third-party) | $250,000 | $600,000 | $1,400,000 |
| RISE subscription premium (18 months at FUE rate) | $1,500,000 | $2,800,000 | $5,200,000 |
| Parallel ECC running, 12 months on existing | $600,000 | $1,200,000 | $2,400,000 |
| BTP integration and extension build | $200,000 | $500,000 | $1,400,000 |
| Test cycle (SIT, UAT, performance, regression) | $350,000 | $800,000 | $1,800,000 |
| End-user training (S/4HANA UX, Fiori) | $120,000 | $280,000 | $650,000 |
| Cutover and hypercare (8 weeks) | $200,000 | $450,000 | $1,100,000 |
| Programme management and PMO | $300,000 | $700,000 | $1,500,000 |
| Contingency at 15 percent | $1,200,000 | $2,100,000 | $4,500,000 |
| Net total estimate | $8,220,000 | $16,330,000 | $33,950,000 |
The typical-column number, $16.3M, is materially higher than typical mid-market cloud migration scenarios because SAP is structurally more expensive than other workloads. SI partner labour alone is $5.5M of the total, before any technology cost. The cost discipline that brings the typical column down is scope discipline: limiting custom code remediation, deferring non-critical integrations, and accepting a leaner test cycle than historical SAP programmes have required.
All three major hyperscalers have credible SAP HANA offerings. The choice between them is rarely a technology decision and usually a commercial decision. Three factors typically decide the destination. First, existing hyperscaler relationship and Enterprise Agreement leverage; customers with a Microsoft EA frequently choose Azure for SAP because the commercial terms align. Second, partner ecosystem; specific SI partners have stronger practices on specific hyperscalers and the partner choice often drives the hyperscaler choice. Third, peripheral service preference (analytics on BigQuery favours GCP, AI services aligned with Azure OpenAI favour Azure, the broadest service portfolio favours AWS).
For RISE with SAP, the customer chooses the hyperscaler at contracting and SAP delivers on that hyperscaler. SAP does not currently offer hyperscaler choice changes mid-contract, so the destination decision is a multi-year commitment. The cost difference across hyperscalers for the same RISE FUE count is typically under 5 percent; the commercial terms with SAP dominate.
SAP HANA migration is the largest and most expensive shape of cloud migration that most enterprises undertake. The cost is dominated by SI partner labour and SAP licensing. Hyperscaler choice rarely changes the bottom line by more than 5 to 10 percent. The discipline that produces an on-budget outcome is custom code rationalisation, integration scope discipline, and aggressive test cycle planning. The teams that succeed treat the migration as a business transformation programme with a cloud destination, not as a cloud migration with SAP attached.
A. Mid-enterprise SAP HANA migrations (5 to 15 TB HANA database, ECC or S/4HANA, single production landscape) typically cost $3M to $12M over 12 to 24 months, all-in including SAP-certified instance cost during the parallel run, partner consulting, RISE with SAP commercial change if applicable, and validation testing. The cost is dominated by the SAP-certified compute (bare-metal or VM with SAP HANA certification) and the SI partner labour, with smaller programmes landing under $3M and large multi-landscape S/4HANA conversions running above $20M.
A. RISE with SAP is SAP's bundled offer of S/4HANA Cloud Private Edition (the SAP-managed version of S/4HANA), running on one of the three hyperscalers (AWS, Azure, GCP) chosen by the customer. RISE bundles the SAP licence, the underlying infrastructure, and the SAP-managed services into a single per-FUE (Full Use Equivalent) per-month subscription. For most customers RISE eliminates the need to manage the SAP-certified infrastructure separately, at the cost of locking in SAP's ongoing service fees. The break-even versus self-managed S/4HANA on hyperscaler typically lands at 30 to 50 percent of the equivalent self-managed cost in year one but converges over years two and three.
A. All three hyperscalers offer SAP-certified instance families: AWS U-series and X2iedn, Azure M-series and Mv3, GCP M2 and M3. Pricing is broadly comparable on like-for-like instance sizes. Azure typically wins for customers with existing Microsoft Enterprise Agreement leverage. GCP typically wins on raw compute price for very large HANA databases (above 6 TB). AWS typically wins on partner ecosystem breadth and the U-series bare-metal range for the largest deployments (up to 24 TB HANA). The hyperscaler choice is rarely decided on raw price; it is decided on customer-specific commercial leverage and the existing partner relationship.
A. S/4HANA conversion (from ECC to S/4HANA on the same hyperscaler) typically costs $4M to $20M over 18 to 36 months for mid-enterprise estates. The cost is dominated by SI partner labour and custom code remediation. Migration (lift-and-shift ECC to a hyperscaler without conversion) typically costs $1.5M to $6M over 9 to 18 months. Most enterprises do migration first then conversion second, spreading the cost across two programmes and reducing risk; some do them together (the Selective Data Transition approach) which compresses timeline but increases risk concentration.
A. SAP Business Technology Platform (BTP) is SAP's PaaS offering for extensions, integrations, and side-by-side applications. Most S/4HANA Cloud deployments use BTP for at least integration and identity, often more. BTP pricing is consumption-based with a Cloud Platform Enterprise Agreement (CPEA) discount available at scale. Mid-enterprise BTP spend typically lands at $200K to $800K per year. The cost is rarely included in initial migration cost estimates and frequently surprises programme budgets.
A. ECC to ECC-on-cloud migrations (lift-and-shift) take 9 to 18 months for a single landscape, longer for multi-landscape estates. S/4HANA conversions take 18 to 36 months. RISE with SAP transitions take 12 to 24 months and align with the SAP commercial cycle. The dominant timeline factor is downtime tolerance and the testing cycle, not raw data movement; HANA databases migrate in days, the application validation and integration testing takes months.
Oracle to AWS ->
Another high-licence workload
On-prem to AWS ->
MAP, EC2, Snowball
On-prem to Azure ->
Hybrid Benefit, FastTrack
On-prem to GCP ->
Sustained-use, BigQuery
Data centre exit ->
Multi-workload programme reality
Mainframe modernisation ->
Another long programme
Strategy cost tables ->
7Rs framework
10 hidden costs ->
Detailed playbook
Updated 2 May 2026