Two things are true about the VA claims system in 2026, and holding both at once is the whole game. The backlog of aging claims has fallen dramatically, from more than 400,000 in early 2024 to around 70,000 now, which is real progress by any measure. And an August investigation by NOTUS, drawing on VA employees, internal data, board attorneys, and veterans' advocates, reports that the speed is coming with a cost: rising denial rates, basic errors, and an appeals pipeline swelling with decisions that shouldn't have been made the way they were.
We're not here to referee the politics; the VA disputes the accuracy criticism and points to its own improving error metrics. We're here for what the numbers mean for the decision letter sitting on your table.
What the reporting found
The load-bearing numbers from the investigation. Denial rates climbed from 35% of claims in fiscal 2024 to 38% in 2025 and 39% in the first quarter of fiscal 2026, and in one recent quarter the VA completed nearly 30,000 more claims than two years prior while granting 9,000 fewer. Pending higher-level reviews, the lane for veterans seeking a second look, have roughly doubled. At the Board, a union-commissioned survey reported 97% of the Board's attorneys saying new production quotas increased the likelihood of factual or legal errors, and 93% saying they no longer have time for a completely fresh review. At the court above it all, appeals hit a record in fiscal 2025, and of the more than 10,000 cases the court received, 9,600 contained errors.
The VA's response, for fairness: it says its internal accuracy metrics have improved alongside the speed, with error-free rates rising from roughly 79% to 83%, and that it measures success by service to veterans. The inspector general's audit office, meanwhile, told NOTUS its reviewers find higher error rates than the VA's internal checks do. Both of those things are on the record; weigh them as you like.
What this means for your decision letter
Strip out the institutional fight and a practical picture emerges for the individual veteran, and it's one we've been building toward all year.
A fast denial deserves slow reading. The investigation's most humane detail is also its most instructive: a veteran denied with reasoning that referenced medical issues that weren't his, copy-paste artifacts from someone else's case, and a supplemental medical opinion denied within days without the new evidence being assessed. Those are exactly the failure modes
the five-check reading we laid out is built to catch: reasons that don't match your file, evidence lists missing what you submitted, issues never addressed at all. In a speed-first environment, the decision letter is where the corners that got cut become visible, and the veterans who read theirs closely are the ones who catch it.
A denial is a first draft, now more than ever. If the system is producing more denials under production pressure, then the population of wrong denials is growing with it, and wrong denials are what
the review lanes exist to fix. The court numbers make the point better than any argument: when the overwhelming majority of decisions that reach independent review contain errors, walking away from a denial without a hard look is statistically the wrong move.
Expect the second round to be slower, and protect your dates. The doubling of pending higher-level reviews means the correction lanes are congested, which makes two disciplines matter more: choosing the right lane the first time rather than reflexively grabbing one, and
keeping your effective date protected by acting inside the deadlines while the queue does its thing. Time in line costs nothing if your date is protected. It costs everything if it isn't.
Build claims that are hard to deny fast. The reporting describes adjudicators moving at speed through whatever is in front of them. You can't control their clock; you can control what's in front of them. A
complete, organized file with the elements laid out plainly is harder to wrongly deny in five minutes than a thin one, which has always been true and is truer under quota pressure. It's also, frankly, the design logic behind
our free initial-claims representation: the claim built right at the front end is the one that survives a rushed reviewer.
The bottom line
None of this means your specific denial was wrong; plenty of denials are correct, fast or slow. It means the base rates have shifted in a direction that makes checking worth it. A system optimizing for throughput will produce more decisions like the ones in the reporting, and the veterans who treat their decision letters as documents to be audited rather than verdicts to be accepted are the ones the current moment rewards.
Where we come in
We're a veteran-led firm, and this environment is precisely our job description: reading fast decisions slowly, finding the errors production pressure leaves behind, and choosing the review lane that actually fixes them. If your denial came back quick and reads thin,
reach out and we'll tell you honestly whether it holds up.
