Generative design
Generative design is a computer-assisted process that creates and evaluates design options against goals and constraints set by a person. Engineers, architects and product designers use it to explore trade-offs such as strength, weight, cost and available space.
How it works
The designer defines what a part or structure must achieve: loads it must carry, areas it must fit around, materials and permitted manufacturing methods. Software generates alternatives and evaluates their performance, often through simulation.
A person compares the results and selects candidates for further development. The output is not automatically ready for production: it still needs engineering review, manufacturing checks and appropriate physical testing.
A bracket that is 30% lighter but costs more
A hardware startup compares an existing bracket with a generated alternative. In this hypothetical example, both pass the same load tests. The manufacturing costs are illustrative, not market benchmarks.
| Measure | Existing bracket | Generated candidate |
|---|---|---|
| Mass per bracket | 500 g | 350 g |
| Manufacturing cost per bracket | $12 | $18 |
| Cost for 10,000 brackets | $120,000 | $180,000 |
The mass reduction is (500 − 350) ÷ 500 = 30%.
But production costs rise by ($18 − $12) × 10,000 = $60,000.
The generated shape wins on weight, not cost. For a drone startup, lower weight might justify the expense through longer flight time. For a stationary machine, the same trade-off might make no commercial sense.
That is why the goals matter: optimizing only for low weight can produce an expensive answer to the wrong business problem.
Questions people ask
What does "generative" mean in simple terms?
It means producing new options. Here, software proposes designs rather than merely checking one that a person has drawn. Generative design does not necessarily use generative AI: optimization and simulation can produce alternatives without a chatbot or image-generation model.
