The Cost of Compute is a practical B2 business English lesson that teaches you to explain the factors that drive AI and cloud costs. It includes workplace examples, guided rehearsal, and a next-step exercise you can apply to a real meeting, message, interview, or customer conversation.
In this lesson
Explain the factors that drive AI and cloud costs
Use financial terms like "CapEx," "OpEx," and "Unit Economics"
Discuss strategies for cost optimization in the cloud
The Cost of Compute
In the AI era, "Compute" is the new oil. It is the most significant cost for many tech companies. Understanding the economics of the cloud is essential for any leader or senior engineer.
CapEx vs. OpEx
CapEx (Capital Expenditure): Buying a server and putting it in a room. You own it, but it's expensive upfront.
OpEx (Operating Expenditure): Renting a server from AWS. You pay as you go.
Key phrase: "Moving to the cloud shifted our infrastructure from CapEx to OpEx, giving us more flexibility."
The Token Economy
AI models don't read words; they read tokens.
Rule of thumb: 1,000 tokens is about 750 words.
Pricing: You are charged for input tokens (what you send) and output tokens (what the AI sends back).
Key phrase: "We need to optimize our prompts to reduce token usage and lower our monthly bill."
GPU Scarcity and Cost
Training and running AI requires GPUs (Graphics Processing Units), primarily from NVIDIA.
H100s: The current "gold standard" GPU. They are incredibly expensive and hard to get.
Key phrase: "Our biggest bottleneck right now isn't code; it's GPU availability."
Cost Optimization Strategies
Reserved Instances: Committing to use a server for 1-3 years for a massive discount.
Spot Instances: Using "spare" cloud capacity for a 90% discount (but the cloud provider can take it back at any time).
Model Distillation: Using a large, expensive model to train a smaller, cheaper model.
Summary Table
| Term | Definition | Business Impact |
| --------------------- | ------------------------- | ------------------------- |
| Compute | Processing power | The primary cost driver |
| Tokens | Units of AI text | How you are billed for AI |
| Unit Economics | Cost per user/request | Profitability |
| Over-provisioning | Buying more than you need | Wasted money |
When you talk about "Price/Performance," you are showing that you care about the company's bottom line, not just the technology.
Apply this lesson
Build a rehearsal brief for work you have this week.
This stays on your device. Bring the brief to Alex, a live session, or the conversation itself.
Key takeaways
Cloud costs are OpEx (Operating Expenses), meaning you pay for what you use
AI costs are driven by "tokens" and "compute hours"
Optimization is about balancing performance with cost (Price/Performance ratio)
Check your understanding
Practical questions
The Cost of Compute FAQ
What does the The Cost of Compute lesson teach?
It teaches you to explain the factors that drive AI and cloud costs.
Who should use this The Cost of Compute lesson?
This lesson is for engineering leaders, technical managers, senior developers working in English across teams, customers, or markets.
What should I be able to do after this lesson?
You should be able to cloud costs are OpEx (Operating Expenses), meaning you pay for what you use.
How can I practice the cost of compute?
Adapt one example to your current work, say it aloud, then use the rehearsal brief to practice a realistic response with the AI coach or voice lab.
Discuss this lesson
This academy is updated continuously, and recommendations are welcome. Every submission is checked before it appears. Contact details are never published.