GPU Cloud Selling to the VP of AI Infrastructure — 60-Min Training

Question

Pulse RevOps · The Machine · Accepted Answer

### Direct Answer

> **GPU Cloud Selling to the VP of AI Infrastructure** is a 60-minute training for AEs running $500K–$50M ACV cycles against **CoreWeave, Lambda Labs, AWS, GCP, Azure, Together AI, Fireworks AI**. Qualify against **VP AI Infra + CFO + Procurement**, run discovery on **GPU capacity + InfiniBand + reserved-capacity economics**. Built on **MEDDPICC** + **Force Management**.

---

## Section 1 — Why GPU Cloud Selling Is Different (5 min)

GPU cloud is **capacity-constrained**. NVIDIA allocation gates availability.

End with **Mark Roberge's** rule: *"Sell capacity availability + interconnect, not generic compute."*

---

## Section 2 — The 60-Minute Discovery (15 min)

> 1. **Opening (3 min):** "Walk me through your AI training + inference workloads."
> 2. **GPU count baseline (10 min):** "Current H100/H200/B200 usage?"
> 3. **Workload mix (10 min):** "Training vs inference split?"
> 4. **InfiniBand requirement (10 min):** "Multi-node training? NDR or HDR?"
> 5. **Reserved-capacity posture (8 min):** "1-year, 3-year commits?"
> 6. **Multi-cloud strategy (7 min):** "Single-cloud or multi-cloud?"
> 7. **Renewal posture (5 min):** "Existing contracts?"

```mermaid
flowchart TD
    A[AE Schedules Discovery] --> B[Pre-Brief Sent]
    B --> C{VP AI Infra + CFO + Procurement?}
    C -->|No| D[Reschedule]
    C -->|Yes| E[GPU Count + Workload 20 min]
    E --> F[InfiniBand + Reserved 18 min]
    F --> G[Multi-Cloud + Renewal 12 min]
    G --> H[Capacity Quote Within 5 Days]
```

---

## Section 3 — The Capacity Negotiation That Wins (15 min)

**Failure modes to ban.** On-demand-only pricing. No InfiniBand commitment. No multi-year reserved discount.

**Wins to coach.** **Reserved-capacity discount math** (30–50% vs on-demand). **InfiniBand topology in writing.** **Capacity SLA per region.**

End with **Andy Paul's** rule: *"Show the customer guaranteed capacity + discount."*

---

## Section 4 — Handling the Incumbent (10 min)

**Counter-move 1 — Capacity wedge.** *"Does your incumbent have B200 capacity in 2027?"*

**Counter-move 2 — InfiniBand wedge.** *"NDR 400 Gb/s or HDR 200?"*

**Counter-move 3 — Reserved-capacity discount wedge.** *"What's your 3-year reserved discount today?"*

---

## Section 5 — Pricing Conversation (10 min)

**Landmine 1 — On-demand-only.**

**Landmine 2 — Multi-year discount.** 3-year 30%; 5-year 50%.

**Landmine 3 — No procurement-only meetings.**

```mermaid
flowchart TD
    A[Joint VP AI + CFO + Procurement] --> B[Reserved Capacity Proposal]
    B --> C{Discount + SLA Aligned?}
    C -->|No| D[Reset]
    C -->|Yes| E[MSA Drafted]
    E --> F{Procurement Solo?}
    F -->|Yes| G[Refuse]
    F -->|No| H[Joint Negotiation]
    G --> H
    H --> I[Provisioning Within 14 Days]
    I --> J[Capacity Live Month 1]
    J --> K[Quarterly Capacity Review]
```

---

## Section 6 — The Trap-Set for Multi-Year Renewal (5 min)

**Trap-set 1 — Reserved capacity utilization 80%+ within 6 months.**

**Trap-set 2 — InfiniBand SLA met monthly.**

**Trap-set 3 — Allocation pipeline for next-gen hardware.**

**Trap-set 4 — Joint VP AI dashboard in QBR.**

Close with **Jeb Blount's** rule.

---

## FAQ

**CoreWeave or AWS?** CoreWeave for AI-first aggressive pricing; AWS for enterprise integration.

**H100 or B200?** Mix based on workload maturity.

**InfiniBand mandatory?** For multi-node training, yes.

**Multi-cloud or single?** Multi-cloud for compliance; single for cost discipline.

**Multi-year discount target?** 30% for 3-year; 50% for 5-year.

## Sources

- NVIDIA — Hopper + Blackwell Datasheets
- CoreWeave — Customer Outcomes Reference
- AWS — EC2 P5 P6 Documentation
- GCP — A3 + TPU Reference
- Azure — ND H100 v5 Reference
- Together AI — Inference Reference
- Force Management — MEDDPICC
- Mark Roberge — Sales Acceleration Formula
- Andy Paul — Sell Without Selling Out
- Jeb Blount — Fanatical Prospecting

GPU Cloud Selling to the VP of AI Infrastructure — 60-Min Training

Direct Answer

Section 1 — Why GPU Cloud Selling Is Different (5 min)

Section 2 — The 60-Minute Discovery (15 min)

Section 3 — The Capacity Negotiation That Wins (15 min)

Section 4 — Handling the Incumbent (10 min)

Section 5 — Pricing Conversation (10 min)

Section 6 — The Trap-Set for Multi-Year Renewal (5 min)

FAQ

Sources

What does the score mean?