Every charger bill of materials contains at least one part that is hard to replace: a specific GaN device, a controller IC, a transformer wound to a particular specification. When that part goes on allocation, a programme either has an approved alternative or it has a delay. The difference is decided months earlier.
Why “we have alternatives” is not a policy
An alternative that has not been tested is not an alternative, it is a hypothesis. Swapping a controller changes the switching behaviour, the thermal profile and often the EMC result — which means the substitution can invalidate a certification that was paid for and issued. A real second source is a part that has been designed in, sampled, measured and where necessary added to the tested configuration.
What a workable policy looks like
For each critical part: a primary, a qualified secondary, and the test data showing the secondary meets the same measurements — output accuracy, ripple, temperature rise, EMC margin. Where the secondary changes the EMC profile, the certification scope has to account for both configurations, which is a decision made at design time rather than during a shortage.
What it costs
Engineering time, and in some cases a slightly higher unit cost because the second source is bought in smaller quantities to keep it active. Against that: a programme that keeps shipping during an allocation event, and a customer who does not have to explain a stockout to their own channel.
What buyers should ask
Ask which parts in the bill of materials are single-sourced, what the qualification status of the second source is, and what the factory would do in the first four weeks of an allocation notice. The answer tells you more about the supply chain than any capacity figure. If there is no answer, the risk is yours, and it appears at the worst possible moment — during a launch.