Predictika · Capacity and price commitments to hyperscale customers

An AI sales agent's storage capacity commitment, checked before it goes out

An AI sales agent prepares a capacity and price commitment for a fictional hyperscale cloud customer: the drive, the exabytes, the quarter, the volume tier and price, the destination and export license, the sales channel, and the free capacity it will tell the customer about. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of capacity by drive family and quarter, customer qualification, the HAMR volume quarter, the contract terms and the compliance rules, and returns a verdict, every reason and the figures it computed.

Scope: Capacity and price commitments for three Western Digital data center drive families (Ultrastar DC HC690 UltraSMR, Ultrastar DC HC590 ePMR, and 44 TB HAMR): capacity by quarter, qualification, contract terms, price band, export license and channel conflicts. See everything this validator covers

This validator covers storage capacity commitments for the three drive families below, not just the scenarios shown. It is about pricing and delivery commitments, not product configuration. What is public and what is modeled is stated here and in the findings.

  • Drives (public Western Digital names) — Ultrastar DC HC690 32 TB UltraSMR, Ultrastar DC HC590 26 TB ePMR, and the 44 TB HAMR drive that Western Digital has said it expects to ship in the first half of calendar 2027.
  • Capacity and timing (modeled) — exabytes free by drive family and quarter, shared across customers, and a HAMR volume quarter (Q2 2027). The free-capacity numbers are illustrative; Western Digital does not publish them.
  • Contract terms and prices (modeled) — each customer's base volume, period minimum and maximum, base and upside price per exabyte with a band around it, and which drive families the customer has qualified. Customers are fictional; prices are in millions of dollars per exabyte and are typical for the industry, not Western Digital prices.
  • Compliance and channel (modeled) — an export license rule for a restricted destination, and a conflicting deal registration on a channel partner account.
  • What the agent states — the price per exabyte and the free capacity it will tell the customer are fields in the request; the validator blocks when they differ from what it computes.

Any capacity commitment for these three drive families that touches these rules is in scope — not only the scenarios pictured here. Not covered: other Western Digital products, the Ultrastar models between these, real capacity, allocation and contract terms (Western Digital does not publish them), and countries other than the destination classes modeled here.

The AI sales agentBuilds the capacity and price commitment from the conversation and says it is ready.
Predictika validatorChecks every rule at once and returns the verdict, each finding, and the computed figures.

7 scenarios in 3 groups · each group runs from the agent’s first attempt to the outcome · click one to switch

What the agent proposes

The AI sales agent says

Illustrative agent statement. The validator receives the structured request below.

Proposing agent: AI sales agent (scripted)

Before sending, the agent must call Recorded run

Predictika validator


      

        
        
        
      

      
Waiting for the call.

      
    

  

  
  

    
Where rules-based CPQ gets it wrong
Click to expand
    
Our assessment of what a typical rules-based CPQ (product rules, price rules and validation rules, as in Salesforce CPQ) does with the same cases. It is a judgment about effort and reliability, not a claim that nothing can be built.

    
What the validator caughtTypical rules-based CPQPredictika
Capacity free by drive family and quarter, shared across customersOnly with workarounds
A rules-based CPQ has no supply table. The free exabytes live in a custom object that someone keeps current, and every commitment has to be netted against it by its own rule.		The free capacity by drive family and quarter is part of the model, and every commitment is checked against it on each call.
A drive that cannot be committed before its volume quarterOnly with care
A date rule per product that blocks the HAMR family before its ramp quarter. It works, but it is a separate rule from the capacity check, so the same cause can show up as two unrelated messages.		The ramp quarter is a property of the drive; the timing and the zero capacity are both found and reported together.
Volume above the contracted base that changes the allowed priceOnly with workarounds
Tagging the excess as upside moves the price band. The tier, the quantity and the price band need chained rules, and a fix to one field creates the next error.		Tier, quantity and price band are checked together, so the tier fix and the price it requires show up in the same answer.
Which customers are qualified on which drive familyOnly with care
A table of qualifications and a validation rule per family. Reliable only while the table is maintained for every new drive.		The qualification is a field of the customer, and the check follows from the drive chosen.
Figures the agent states: the price and the free capacityOnly with care
A validation rule can compare a stated figure with a calculated one only if the stated figure is a field on the quote, which it rarely is.		The stated price and free capacity travel with the request and are compared with the computed values on every call.
Every problem returned in one answerOnly with workarounds
Rules fire as a step is completed, so a salesperson fixes one error and meets the next.		All findings come back together, blocks first, then approvals, whatever order the choices arrived in.

  

  

    
What this validator catches: findings from its test cases
Click to expand
    
Every rule is checked on every call, in any order the choices arrive; one problem never hides another. Blocking findings stop the quote; approval findings need a person to sign off; everything else is computed.

    
131 EB is requested and only 30 EB of Ultrastar DC HC690 32 TB UltraSMR capacity is free in Q2 2027; the commitment cannot be promised.
2This commitment leaves under 3 EB of Ultrastar DC HC690 32 TB UltraSMR capacity in Q2 2027; confirm with supply planning before issuing.
3Larkspur Data Systems is not yet qualified to receive the Ultrastar DC HC590 26 TB ePMR; no commitment can be made on this platform.
4Q1 2027 is before HAMR volume production, which begins in Q2 2027; HAMR cannot be committed yet.
58 EB is below Brightwater Compute's contracted period minimum of 14 EB; confirm the shortfall with sales operations.
627 EB exceeds Brightwater Compute's contracted period maximum of 26 EB; this would commit beyond the agreement.
719 EB tagged as base volume exceeds Brightwater Compute's contracted base of 18 EB; the excess must be priced and labeled as upside volume.
8$16 million per EB is outside the contracted Base tier range of $13 to $15 million per EB for Northgate Cloud.
9This shipment to a restricted destination needs an export license for the Ultrastar DC HC690 32 TB UltraSMR and its status is Not_Filed; do not commit a delivery date.
10The export license is pending; the delivery date must be written as contingent on the grant.
11This account has a conflicting channel deal registration; the partner commitment cannot be issued until it is resolved.
12The draft states 28 EB free in Q2 2027, but the supply table shows 30 EB.

  

  

    
Scenarios and results
Click to expand
    
The recorded result of the validator for each scenario on this page: the request the AI agent sent, the verdict, and every finding, in the order they appear above.

    
1. HAMR for a hyperscale customer: 44 TB HAMR for Brightwater Compute in Q1 2027, shipping to a restricted destination

    
The agent's request. Please commit 30 EB of the 44 TB HAMR drive to Brightwater Compute for Q1 2027, tagged as base volume at $15 million per EB. It ships to a restricted destination and the export license is not yet filed. I will tell the customer that 12 EB of the 44 TB HAMR drive are free in Q1 2027. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Capacity short in the quarter. 30 EB is requested and only 0 EB of 44 TB HAMR drive capacity is free in Q1 2027; the commitment cannot be promised.
    • 
      
  • HAMR committed before its volume quarter. Q1 2027 is before HAMR volume production, which begins in Q2 2027; HAMR cannot be committed yet.
    • 
      
  • Commitment above the contract maximum. 30 EB exceeds Brightwater Compute's contracted period maximum of 26 EB; this would commit beyond the agreement.
    • 
      
  • Base volume above the contracted base. 30 EB tagged as base volume exceeds Brightwater Compute's contracted base of 18 EB; the excess must be priced and labeled as upside volume.
    • 
      
  • Export license missing for a restricted destination. This shipment to a restricted destination needs an export license for the 44 TB HAMR drive and its status is Not_Filed; do not commit a delivery date.
    • 
      
  • Stated free capacity does not match. The draft states 12 EB free in Q1 2027, but the supply table shows 0 EB.
    • 
    

    
2. HAMR for a hyperscale customer: The agent resubmits after fixing the errors

    
The agent's request. Please commit 22 EB of the 44 TB HAMR drive to Brightwater Compute for Q4 2027, tagged as upside volume at $15 million per EB. It ships to a restricted destination and the export license is pending. I will tell the customer that 32 EB of the 44 TB HAMR drive are free in Q4 2027. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Stated price outside the tier band. $15 million per EB is outside the contracted Upside tier range of $17 to $19 million per EB for Brightwater Compute.
    • 
      
  • Export license still pending. The export license is pending; the delivery date must be written as contingent on the grant.
    • 
    

    
3. HAMR for a hyperscale customer: The agent resubmits after fixing the errors

    
The agent's request. Please commit 22 EB of the 44 TB HAMR drive to Brightwater Compute for Q4 2027, tagged as upside volume at $18 million per EB. It ships to a restricted destination and the export license is granted. I will tell the customer that 32 EB of the 44 TB HAMR drive are free in Q4 2027. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Commitment summary. 22 EB of 44 TB HAMR drive in Q4 2027 at $18 million per EB; commitment value $396 million.
    • 
    

    
4. Channel partner order: 44 TB HAMR through a channel partner in Q3 2027

    
The agent's request. Please commit 8 EB of the 44 TB HAMR drive to Tessera Distribution for Q3 2027, tagged as base volume at $15 million per EB. It is sold through a channel partner, and another partner has a conflicting deal registration on the account. I will tell the customer that 20 EB of the 44 TB HAMR drive are free in Q3 2027. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Customer not qualified on the drive. Tessera Distribution is not yet qualified to receive the 44 TB HAMR drive; no commitment can be made on this platform.
    • 
      
  • Base volume above the contracted base. 8 EB tagged as base volume exceeds Tessera Distribution's contracted base of 6 EB; the excess must be priced and labeled as upside volume.
    • 
      
  • Conflicting channel deal registration. This account has a conflicting channel deal registration; the partner commitment cannot be issued until it is resolved.
    • 
    

    
5. Channel partner order: The agent resubmits after fixing the errors

    
The agent's request. Please commit 6 EB of the Ultrastar DC HC590 26 TB ePMR to Tessera Distribution for Q3 2027, tagged as base volume at $15 million per EB. It is sold through a channel partner. I will tell the customer that 18 EB of the Ultrastar DC HC590 26 TB ePMR are free in Q3 2027. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Commitment summary. 6 EB of Ultrastar DC HC590 26 TB ePMR in Q3 2027 at $15 million per EB; commitment value $90 million.
    • 
    

    
6. Approvals: UltraSMR for Northgate Cloud, shipping to a restricted destination

    
The agent's request. Please commit 12 EB of the Ultrastar DC HC690 32 TB UltraSMR to Northgate Cloud for Q2 2027, tagged as base volume at $14 million per EB. It ships to a restricted destination and the export license is pending. I will tell the customer that 30 EB of the Ultrastar DC HC690 32 TB UltraSMR are free in Q2 2027. It is ready to send.

    
Verdict: Needs approval.

    
    
      
  • Commitment below the contract minimum. 12 EB is below Northgate Cloud's contracted period minimum of 20 EB; confirm the shortfall with sales operations.
    • 
      
  • Export license still pending. The export license is pending; the delivery date must be written as contingent on the grant.
    • 
      
  • Commitment summary. 12 EB of Ultrastar DC HC690 32 TB UltraSMR in Q2 2027 at $14 million per EB; commitment value $168 million.
    • 
    

    
7. Approvals: The agent resubmits after the approvals are addressed

    
The agent's request. Please commit 24 EB of the Ultrastar DC HC690 32 TB UltraSMR to Northgate Cloud for Q2 2027, tagged as base volume at $14 million per EB. I will tell the customer that 30 EB of the Ultrastar DC HC690 32 TB UltraSMR are free in Q2 2027. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Commitment summary. 24 EB of Ultrastar DC HC690 32 TB UltraSMR in Q2 2027 at $14 million per EB; commitment value $336 million.
    • 
    

  

  

    
Narration transcript
Click to expand
    
The text of the narrated walkthrough, step by step.

    
    
      
  1. Welcome to a demo of Predictika's Validator. An AI sales agent prepares a capacity and price commitment for a fictional hyperscale cloud customer: the drive, the exabytes, the quarter, and more. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of capacity by drive family and quarter, customer qualification, the HAMR volume quarter, the contract terms and the compliance rules, and returns a verdict, every reason and the figures it computed.
    2. 
      
  2. I will start with the first request by the agent.
    3. 
      
  3. The agent says. Please commit 30 EB of the 44 TB HAMR drive to Brightwater Compute for Q1 2027, tagged as base volume at $15 million per EB. It ships to a restricted destination and the export license is not yet filed. The rest of the request is on screen.
    4. 
      
  4. Before it goes out, the agent calls the validator.
    5. 
      
  5. Blocked, with six findings, all in one answer. 30 EB is requested and only 0 EB of 44 TB HAMR drive capacity is free in Q1 2027; the commitment cannot be promised. Q1 2027 is before HAMR volume production, which begins in Q2 2027; HAMR cannot be committed yet. 30 EB exceeds Brightwater Compute's contracted period maximum of 26 EB; this would commit beyond the agreement. The other three are on screen.
    6. 
      
  6. The validator returned six blocking findings in one answer. A check that reads one field at a time would pass most of this; it shows only when the choices in the quote are checked against each other.
    7. 
      
  7. What this shows: for a drive that cannot be committed before its volume quarter, a rules-based CPQ is error-prone for this.
    8. 
      
  8. Now the agent resubmits.
    9. 
      
  9. The corrected request. It changes the delivery quarter to Q4 2027, the commitment to 22 EB, the volume tier to upside, the export license to pending and the stated free capacity to 32 EB.
    10. 
      
  10. It calls the validator again.
    11. 
      
  11. Still blocked, with one finding and one approval, all in one answer. $15 million per EB is outside the contracted Upside tier range of $17 to $19 million per EB for Brightwater Compute.
    12. 
      
  12. Six findings cleared, but the change brings a different blocking finding: stated price outside the tier band.
    13. 
      
  13. The agent resubmits once more.
    14. 
      
  14. The revised request. It changes the export license to granted and the stated price to 18.
    15. 
      
  15. It calls the validator once more.
    16. 
      
  16. This time the validator clears it.
    17. 
      
  17. The validator also returns the line-item quote, on screen.
    18. 
      
  18. Both earlier findings are cleared, and the validator's figure matches the quote.
    19. 
      
  19. Six problems caught in one call, fixed in two passes, and a clean commitment. That is the value of Predictika's validator: catching mistakes before they reach the customer.
    20.