Hopper
H100
The baseline high-end accelerator that became a core reference point for AI compute pricing. Key idea: strong general-purpose AI capacity.
Learn
H100 vs H200 vs B200
How accelerator generations change performance, supply, and cost.
H100, H200, and B200 are NVIDIA data-center accelerators used in advanced AI workloads. They belong to different points in the product cycle, and the differences between them matter because chip generation affects memory, performance, workload fit, pricing, and availability.
Same market, different chipsCompare
All three are high-end AI accelerators, but they do not deliver the same capacity.
Hourly price is not enoughCaution
A more expensive GPU-hour can still be cheaper per completed workload.
At a glance
The simple comparison
Hopper
H200
A Hopper-generation step-up with much larger and faster memory for memory-heavy AI workloads. Key idea: better fit for larger models and memory-sensitive workloads.
Blackwell
B200
The next-generation Blackwell accelerator, pushing the performance and memory frontier higher again. Key idea: a new generation that can shift workload economics and market expectations.
Example
Why the newer chip can be cheaper in practice
A GPU that costs more per hour can still be cheaper per completed workload if it finishes the job faster, supports a larger model more efficiently, or reduces the number of chips required.
Takeaway
Price per GPU-hour is only one input.
Cost per useful work is the better comparison.
Chip generations
What changes from H100 to H200 to B200
- Memory capacity and bandwidth increase, which matters for larger models and memory-heavy workloads.
- Architecture advances can improve throughput and efficiency.
- New generations can change which workloads are practical, how many chips are needed, and what buyers are willing to pay.
- Availability and supply mix also change as the market moves from one generation to the next.
Why it matters
Why chip generation matters in the market
- Providers price different accelerators differently because they deliver different value.
- Buyers compare not just hourly rental rates, but workload fit and total cost to complete a job.
- Supply can shift as newer chips enter the market and older chips remain in service.
- A market index has to distinguish between chips instead of treating all GPU-hours as identical.
Common mistake
Do not compare chips by hourly price alone
A lower hourly rate does not automatically mean lower compute cost. The right comparison is whether a chip can complete the required workload at the needed speed, scale, and total cost.
Price
Hourly price
What access costs per unit of time.
Performance
Performance
How much useful work the chip can complete.
Fit
Workload fit
Whether the chip is well suited to the model and task.
Keep learning
Related lessons
Concept
What is AI compute?
The basic resource behind training and running AI models.
Unit
What is a GPU-hour?
The basic unit behind compute pricing.
Market
What is compute cost?
How the market price of AI compute capacity is expressed and compared.