Salesforce Data Cloud Starter lists at $108,000 per year, Foundations at $216,000, and Enterprise scales by credit consumption with prices ranging from $0.0008 per credit to $0.014 per credit depending on workload type. The base bundle includes 250,000 unified profiles, 1 billion records, and 100 million segmentation rows. Mid-market deployments routinely cross $400,000 to $1.4M per year within 18 months of go-live. Enterprise deployments at scale reach $2M to $8M. This page is the 2026 reference for Data Cloud tier mechanics, credit consumption math, segmentation pricing, profile counts, the consumption overshoot pattern, and the negotiation levers that contain Data Cloud spend.
Inside This Pillar
- Data Cloud 2026 tier matrix
- Starter at $108,000 per year
- Foundations at $216,000 per year
- Enterprise consumption model
- The credit pricing math
- Profile and record limits
- Segmentation and activation rows
- Einstein AI and Agentforce credits
- The consumption overshoot pattern
- Discount bands by deal size
- Data Cloud versus Snowflake and Databricks
- Negotiation levers for 2026
- How to reduce Data Cloud cost
Data Cloud 2026 tier matrix
Salesforce Data Cloud follows a tier plus consumption model in 2026. The platform fee sets the base capacity envelope. Credit consumption fills out usage above the included entitlement. Three named tiers cover most deployments.
| Tier | Annual list | Included entitlement |
|---|---|---|
| Data Cloud Starter | $108,000 | 250,000 unified profiles, 1 billion records, 100 million segmentation rows, 1,000 credits per hour |
| Data Cloud Foundations | $216,000 | 500,000 unified profiles, 2 billion records, 250 million segmentation rows, 2,000 credits per hour, expanded ingestion connectors |
| Data Cloud Enterprise | Custom, starts $432,000 | 1 million plus unified profiles, multi-tenant data spaces, full credit catalogue, Calculated Insights at scale |
| Data Cloud for Marketing Cloud | $108,000 add-on | Activation to Marketing Cloud Engagement and Account Engagement |
| Data Cloud for Service Cloud | Included in Einstein 1 Service | Profile unification scoped to Service workloads |
The platform fee buys an entitlement envelope. Customers exceed the envelope through any combination of profile growth, record ingestion, segmentation evaluation, activation, AI inference, or calculated insight computation. Each overage class consumes credits at a different rate.
Starter at $108,000 per year
Starter is the entry point for organisations bringing first-party customer data into Data Cloud. The 250,000 unified profile entitlement covers mid-market customer bases and pilot deployments within larger enterprises. The 1 billion record cap usually binds before the profile cap in retail, financial services, and telecommunications scenarios where event volume is high.
Starter includes the connector library for Marketing Cloud, Service Cloud, Sales Cloud, and Commerce Cloud. It includes core identity resolution. It includes 1,000 credits per hour of segmentation, calculated insight, and AI workload capacity. The hourly credit rate is the binding constraint for most Starter customers within the first six months.
Starter does not include the Snowflake Zero Copy data sharing connector at full bidirectional capacity. It does not include the Databricks Zero Copy connector at full capacity. It does not include cross-region data residency. These constraints push more demanding deployments to Foundations or Enterprise.
Foundations at $216,000 per year
Foundations doubles the Starter envelope and adds the Snowflake and Databricks Zero Copy bidirectional connectors at production scale. The 500,000 profile and 2 billion record entitlement covers a mid-large enterprise customer base. The 2,000 credit per hour ceiling supports modest AI and segmentation workloads in parallel.
Foundations is the most common landing point for B2C enterprises with 1 million to 5 million customers. It is also the most common tier where the consumption overshoot pattern begins to bite, because the AI and segmentation workloads added to a Foundations deployment frequently exceed the included credit envelope within 12 months.
Enterprise consumption model
Data Cloud Enterprise is the named tier for deployments above 1 million unified profiles or with regulatory data-residency requirements. Pricing starts at $432,000 per year for the base Enterprise platform fee and scales by negotiated credit commitments. Most Enterprise customers commit between 50 million and 500 million credits per year on multi-year terms.
Enterprise includes data spaces for multi-brand or multi-region segregation. It includes cross-region replication. It includes priority ingestion support. It includes the full Calculated Insight catalogue at production volume. The biggest commercial difference versus Foundations is that Enterprise customers negotiate credit unit prices in volume bands rather than paying per-overage at the menu rate.
The credit pricing math
Data Cloud credit consumption rates are workload-specific. The rate card differs by activity class. The seven main credit categories are listed below at illustrative 2026 list rates.
| Workload class | Credit consumption | Typical scale |
|---|---|---|
| Batch ingestion | 1 credit per 100,000 records | 10 to 50 million credits per year for a mid-market deployment |
| Streaming ingestion | 1 credit per 10,000 records | Higher rate reflects real-time processing |
| Identity resolution | 1 credit per 100 profiles unified | Bound by profile count and reconciliation cadence |
| Segmentation evaluation | 1 credit per 1,000 segment members | Heaviest credit consumer in most deployments |
| Activation | 1 credit per 100 records activated | Driven by campaign cadence |
| Calculated Insight computation | 10 credits per insight refresh per 1 million records | Scales with insight catalogue size |
| Einstein AI inference | 5 to 50 credits per prediction | Highest per-event consumer, scales with use case maturity |
The credit menu rate above the included entitlement starts at approximately $0.014 per credit and falls to $0.0008 per credit at multi-hundred-million credit commitments. A customer activating segmentation for 50 million members across weekly cadence consumes 2.6 billion credits per year. At menu rate that is $36M. At commit rate it is $2M to $6M. The negotiation discipline between menu and commit is the single largest cost lever in Data Cloud.
Profile and record limits
Unified profile counts measure the distinct customer or constituent records produced after identity resolution. The included entitlement of 250,000 or 500,000 profiles is rarely the binding constraint for mid-market organisations but binds quickly for B2C enterprises. Record counts measure ingested raw events: web behaviour, transactions, support cases, product telemetry. The 1 billion or 2 billion record ceiling binds for high-event-volume industries before the profile ceiling.
Customers exceeding the entitlement pay overage at the credit menu rate by default. A 600,000-profile deployment on Foundations (which includes 500,000) consumes additional identity resolution credits for the surplus. The discipline is to negotiate the platform fee tier high enough that the included entitlement covers the production volume, rather than chronically paying overage.
Segmentation and activation rows
Segmentation is the heaviest credit consumer in most Data Cloud deployments because customers refresh segments more often than they intuitively expect. A daily refresh on a 5 million segment is 1.8 billion segment-evaluation rows per year. The included segmentation envelope of 100 million or 250 million rows is exhausted within weeks at typical enterprise activation cadence.
Activation is the second-heaviest credit consumer. Sending 5 million records per month to Marketing Cloud Engagement activates 60 million records per year. At 1 credit per 100 records activated that is 600,000 credits, modest in itself. The combined load of segmentation and activation drives the headline credit consumption number that exceeds the platform fee for most deployments by Year 2.
The two-engine cost pattern: In 80 percent of Data Cloud deployments, segmentation evaluation and AI inference together account for more than 70 percent of credit consumption. The platform fee is 25 to 40 percent of total spend in Year 1 and falls below 20 percent in Year 2 as AI use cases compound. The negotiation focus should be the credit unit price and the unit price escalator, not the platform fee.
Einstein AI and Agentforce credits
Einstein and Agentforce inference is the fastest-growing credit consumer in 2026 deployments. A single Agentforce conversation typically consumes 50 to 200 Data Cloud credits between profile retrieval, calculated insight refresh, and predictive ranking. A million conversations per year (modest for enterprise contact centres) is 50 million to 200 million credits, often equivalent to or larger than the platform fee.
For comparable customer profile workloads, see Agentforce pricing and ROI, which details the per-conversation cost model that pairs with Data Cloud credit consumption. The Agentforce conversation fee is separate from the Data Cloud credits that fuel each conversation.
The consumption overshoot pattern
Consumption overshoot is the most common commercial pathology in Data Cloud deployments. The pattern: Year 1 negotiated commit set at expected Year 1 use, modest buffer. Actual Year 1 use exceeds commit by 30 to 70 percent due to underestimated segmentation cadence, AI use case proliferation, and untracked ingestion growth. Overage at menu rate creates a multi-hundred-thousand dollar surprise invoice. Year 2 renewal positions the customer in a weaker negotiating posture because the actual consumption number is already on the table.
The defence is dual: instrument credit consumption from week 1 of production, and negotiate a true-up at fair value rather than overage at menu rate as part of the original contract terms. The same defensive pattern applies to Snowflake, Databricks, and other consumption-based platforms.
Discount bands by deal size
Data Cloud platform fee and credit unit discounts follow these advisor-observed benchmarks for 2026 renewals and new logos.
| Annual commit | 1-year discount | 3-year discount | 5-year discount |
|---|---|---|---|
| $100k to $500k | 5 to 12 percent | 10 to 18 percent | 15 to 22 percent |
| $500k to $2M | 10 to 18 percent | 18 to 28 percent | 24 to 34 percent |
| $2M to $5M | 18 to 28 percent | 28 to 38 percent | 32 to 44 percent |
| $5M plus | 25 to 38 percent | 35 to 48 percent | 42 to 55 percent |
Discount on the credit unit price is more important than discount on the platform fee for Year 2 onward. Many enterprises win 35 percent off the platform fee but only 12 percent off the credit unit, then discover that credits account for 70 percent of the bill within 18 months.
Data Cloud versus Snowflake and Databricks
Data Cloud occupies a different position in the data stack than Snowflake or Databricks. Data Cloud is purpose-built for customer profile unification, segmentation, and activation. Snowflake is a general-purpose data warehouse. Databricks is a lakehouse plus AI platform. Most enterprise deployments use Data Cloud alongside Snowflake or Databricks via the Zero Copy connector, with the warehouse holding the system of record and Data Cloud holding the activation layer.
The economic comparison: Snowflake credit pricing covers compute on the warehouse side. Data Cloud credit pricing covers the activation side. A combined deployment commonly runs $1.5M to $4M on Snowflake plus $800k to $3M on Data Cloud. See Snowflake versus Databricks versus BigQuery for the underlying warehouse decision.
Negotiation levers for 2026
Seven levers move Data Cloud renewal pricing.
Credit unit price discount. The most important single lever. Negotiate the unit price at the volume band you expect to reach by Year 2 or Year 3, not Year 1. A 15 percent improvement on credit unit price often beats a 30 percent improvement on platform fee.
True-up at fair value. Contractual clause that overage above the commit converts to commit-rate pricing at the next true-up, rather than burning at menu rate during the term. This single clause has saved customers seven figures on Year 2 invoices.
Multi-year commit with escalator cap. Three-year and five-year commits deliver 10 to 18 points of additional discount. Cap the annual unit price escalator at 3 to 5 percent.
Bundled commit across Marketing Cloud, Service Cloud, and Data Cloud. Customers consolidating their Salesforce footprint can negotiate cross-product credit pooling that improves utilisation.
Consumption forecast accuracy. Buying short on Year 1 and topping up later is more expensive than buying close to forecast. Invest in pre-deployment consumption modelling.
Competitive alternative documented. Snowflake plus a third-party customer data platform (Treasure Data, Tealium, Adobe Real-Time CDP) is the credible alternative. Documented vendor selection moves Data Cloud pricing materially.
Phased AI rollout commitments. Rather than committing 100 percent AI capacity from day one, structure the contract with phased increases that the customer signs off as use cases mature. Reduces stranded credit.
Negotiation pattern that works in 2026: A 3.2 million customer profile B2C enterprise renewed Data Cloud Foundations plus 180 million negotiated credits per year on a 4-year term. The credit unit dropped from $0.011 menu to $0.0021 commit, a 17 percent discount on the volume band rate. Annual escalator capped at 4 percent. True-up clause converts overage to commit rate at each anniversary. Total annual commitment $570,000 versus the $1.8M list-rate equivalent on prior consumption.
How to reduce Data Cloud cost
Pre-renewal optimisation (12 to 18 months ahead). Instrument credit consumption by workload class. Identify segmentation refresh cadences that exceed business need (daily refreshes that could be weekly, hourly insights that could be daily). Identify orphaned segments that still evaluate but no longer feed an active campaign. Audit calculated insight catalogue for unused or duplicate insights. Most Data Cloud deployments have 18 to 32 percent recoverable credit consumption from this housekeeping alone.
At renewal. Apply the seven levers above. Lock the credit unit price at the right volume band. Add the true-up at fair value clause. Negotiate multi-year with escalator cap. Document the competitive alternative.
Mid-term. Push the consumption forecast review every quarter, not annually. Adjust workload patterns when consumption diverges. Build the Calculated Insight library on the right cadence rather than the default daily refresh.
For the broader Salesforce commercial framework, see the Salesforce licensing guide, Salesforce pricing 2026, Sales Cloud pricing, Service Cloud pricing, and Salesforce renewal strategy. The Salesforce vendor hub aggregates the cluster. Engagement starts at cloud contract negotiation for consumption-model deals or software licensing advisory for whole-portfolio Salesforce strategy.