It's an order of magnitude estimate of the worst case per acre, multiplied by a very course estimate of the number of acres affected, with little or no accounting for the length of the outage, the timing of the outage relative to specific crop-planting times in specific regions, and similar factors. It relies on manufacturer's estimates of the benefit of using their technology to assess the impact of not being able to use the technology. In other words, it's a SWAG.

Consider: the outage may have meant that some equipment would be slower, less accurate, or missing some features, causing time delays and other losses that may or may not be equivalent to using the non-GPS equipment. Or it could have rendered the equipment entirely unusable during the outage. It could even have permanently damaged the equipment (e.g. causing it to drive over an embankment or something). There is no apparent attempt to differentiate these outcomes, which could have significant impact on the estimate.