The mechanism worth studying here is the moment of assignment rather than the statute itself. When a product category arrives without a law, it goes to whichever existing authority can be stretched to reach it, and the stretching tends to be done quickly, by staff, with litigation already in the room. The reasoning recorded by the Works in Progress essay was that genomes are made of chemicals and novel chemicals were already regulated, so modified genomes were covered too [7]. An answer like that becomes durable through inertia rather than quality: once a category has a home, moving it requires an agency to volunteer for work nobody asked it to do, and the constituency lobbying for that volunteering consists of firms whose products do not exist yet.
The arithmetic of the freeze is worth stating precisely. Researchers filed more than 240 applications between 1987 and 2018, of which one, in 1997, was approved for widespread use [8]. That is 31 years [18], roughly eight filings a year [19], and a widespread-use approval rate below half a percent [20]. Eight filings a year might look like modest demand rather than a bottleneck, but a venue with one clearance in three decades does not attract speculative filings in the first place, so thin volume is what a closed door looks like from the outside. The record is also thin in one respect an operator should notice: the essay reports a single widespread-use approval, not how many of the remainder were denied, withdrawn, or cleared for something narrower. We do not know the composition of that denominator, and it matters for whether the binding problem is refusal or attrition.
What the case shows most clearly is that the trigger is definitional rather than physical. FAST-PETase, built at the University of Texas at Austin, is 98 percent identical to the natural enzyme and digests PET in under 24 hours, against six weeks for wild microbes cultured on a thin PET film at 30 degrees centigrade [12][13]. On the essay's own comparison that is at least a fortyfold improvement [21], though it sets a cultured organism beside an engineered enzyme, so the multiple is directional rather than measured. The engineering has been done. Putting the improved gene back into the organism it came from is the step that brings the Toxic Substances Control Act into play [14]. The heavy-metal sensors sit in the same position: E. coli reading cadmium at five parts per billion, and a 2010 student project reading mercury at seven nanomolar, seven times below the EPA's contamination threshold, all of it confined to the lab because the strains carry intergeneric DNA [15][16][17].
None of this moved this week, which is the point. The useful distinction for anyone allocating capital is between the decade-scale question and the quarterly one. The decade question is whether engineered organisms ever get a venue designed for them, and the essay's forty-year record suggests that is not a planning assumption [9]. The quarterly question is narrower and answerable: does the value of the programme you are funding depend on a boundary term you cannot move, such as intergeneric DNA [17], and if it does, is there a contained or cell-free formulation that captures part of the return without crossing it. Firms that ask that while the science is still in progress retain choices, and firms that ask afterwards inherit whatever venue the category was assigned decades before they existed.
