One of the ways could be to make cross-over design – the number of participants stays the same, but the measurement of condition would be twice, thus increasing the power.

That is hard to answer while knowing so extremely little about your trial.

My first resort would probably be to question the budget constraint: Are the endpoints that lack power important enough? If so, then the appropriate budget should be made available accordingly.

As Kastytis mentioned, indeed if you can switch to a cross-over design, that should help a lot.

Perhaps you can also reconsider the effect sizes that you assume for your power analysis.

If you expect that the average cost of a participant is different by trial arm, you may want to consider a different than 1:1 randomization. It is possible that the larger total number of participants that that allows for, more than compensates to loss of power due to changing randomization proportions. (Unless the cost difference is quite large, that will at best lead to a quite modest improvement of power.)

As a last resort you should consider changing the status of the primary end points for which you don't have enough power. When you are at this point, the first resort solution attempt showed that your funder does not think that those end points are important enough, therefore dropping them as primary end point may be appropriate. You may even be able to save some money per participant by dropping some of your end points altogether so that you can increase the sample size and thus save some of your primary end points.

I am sure that there are other possibilities that need detailed knowledge of your trial.

Cross-over designs are problematic because any treatment effect carry-over from one period to the next will invalidate them.

Sample size and power depend, among other things, on the nature of the primary outcome measure. Sample sizes for continuous variables are much smaller than for binary variables,