An enterprise runs a central analytics warehouse in BigQuery with predictable, 24/7 business intelligence queries and continuous data transformations across multiple departments. The company stores 600 TB of uncompressed data that achieves a high compression ratio in BigQuery columnar storage, with minimal daily table modifications or data overwrites. The chief financial officer (CFO) requires predictable monthly compute expenses and wants to minimize the overall total cost of ownership (TCO) for both query compute and storage. Which combination of BigQuery compute and storage billing models should the data engineering team implement?

Implement capacity-based compute billing using slot reservations and commitments, and set the dataset storage billing model to physical storage.: What the Solution Is This solution pairs BigQuery capacity-based compute pricing (using slot reservations and optional commitments) with the physical (compressed) storage billing model at the dataset level to provide cost predictability and optimize total storage expenditure. How It Meets the Requirements Predictable Monthly Compute Expenses: Capacity-based pricing allows organizations to purchase slot commitments and allocate pools of slots into reservations (e.g., for ELT and BI). This eliminates variable per-byte query charges and provides a fixed, predictable monthly compute budget. Optimizes Compute TCO for Steady Workloads: For continuous, 24/7 data processing and BI querying, baseline slot commitments offer substantial discounts compared to pay-as-you-go or on-demand rates. Minimizes Storage TCO with High Compression: The physical storage billing model charges based on the compressed bytes stored on disk. Because the enterprise data exhibits high columnar compression ratios and low daily churn (which keeps time travel and fail-safe charges minimal), physical storage significantly undercuts standard logical storage costs. Key Advantages and Capabilities Budget Predictability: Compute costs are strictly bounded by configured reservation slot limits and baseline commitments. Workload Isolation: Multiple departments can share dedicated compute pools or take advantage of idle slot sharing across reservations in an administration project. Reduced Storage Footprint: Leverages Capacitor columnar compression so billing reflects compacted physical data rather than raw logical bytes. No Performance Trade-off: Switching between logical and physical storage billing models alters only billing metrics without impacting underlying read/write performance. Why This Is Optimal Steady-state 24/7 analytical workloads suffer from unbounded cost volatility and high unit costs under on-demand scanning models. Similarly, highly compressible data with low mutation rates is financially penalized under logical storage billing. Choosing capacity reservations alongside physical storage delivers both maximum cost governance and lowest overall TCO.