OKX changes how much margin hedge-mode positions require

What is changing and who is affected

OKX is changing initial and maintenance margin calculations in cross-margin hedge mode for Futures and Multi-currency Margin accounts. Instead of adding the two sides’ requirements, it selects the larger. Isolated margin is excluded. The phased rollout starts September 22, with completion expected by October 12, 2026; account and API figures determine what currently applies.

For each side, positions and opening orders are included. Initial margin divides their combined value by leverage; maintenance margin multiplies it by the applicable ratio. The final requirement is the larger of the separately calculated long-side and short-side values.

What the two types of margin mean

OKX’s general margin guide describes initial margin as the amount needed to open exposure and maintenance margin as the minimum needed to sustain it. Cross margin shares collateral across positions, while isolated margin assigns it to individual positions. Those definitions help explain the update, but the newer rollout notice governs the changed hedge-mode calculation.

The general guide still shows the older summed hedge-mode formula in its earlier material. That is why the date and scope of the announcement matter. Reading a general explanation and a newer change notice together requires identifying which rule the notice replaces, rather than treating every formula on both pages as simultaneously applicable.

An equal-side example

Imagine that the long-side initial requirement has already been calculated as 100 and the short-side requirement as 100. Adding them gives 200. Selecting the larger gives 100. The difference is 100, which is half of the original sum. These are invented requirement values, not prices or an account quote.

The important word is “requirement.” Nothing in that subtraction says the account earned 100. It says two ways of combining the same inputs produce different collateral totals. A reduction in the amount required is not the same event as a realized gain from a completed trade.

It also does not follow that every account sees a fifty-percent reduction. The equal inputs create that proportion in this example. To understand a different account, you would first need its own side requirements and the rule actually applied to it during the rollout.

Unequal sides produce a different result

Now use 100 for the long side and 40 for the short side. The sum is 140, while the maximum is 100. The difference is 40, about 28.6 percent of the sum. With 100 on one side and zero on the other, both methods produce 100 and there is no difference.

These three cases reveal the pattern: when both inputs are nonnegative, the difference between their sum and their maximum is the smaller input. Equal sides make the difference look largest as a share of the sum. A much smaller second side produces a smaller proportional change.

The same reasoning can be applied to two already-calculated maintenance requirements, but that does not make the initial and maintenance inputs identical. Keep each calculation in its own row. Combining an initial requirement from one side with a maintenance requirement from the other would mix different quantities before the comparison even begins.

How to review the change in an actual account

A useful review record would contain the account mode, observation time, each side’s requirement and the combined figure shown by the platform. Include opening orders in the investigation, because a position-only note can omit part of the announced calculation. If the result differs from your expectation, first establish which method is active rather than assuming that the rollout reached every account at once.

The examples below compare sums and maximums without estimating liquidation prices. They do not model changing position values, costs or the account’s full collateral situation. Use them to understand the arithmetic, then rely on the actual account record for the applied requirement. The most accurate description of the update is a changed margin rule with a phased implementation, not a guaranteed saving or a promise of lower trading risk.