Two waves, one number
Russian forces spent much of 2021 massing near Ukraine’s borders in two separate waves: an initial buildup in March and April that partly withdrew by June, and a second buildup beginning in October that, by December, had placed more than 100,000 troops on three sides of the country, including Belarus to the north and Crimea to the south. Russian officials denied, repeatedly, that any invasion was planned. One document from that period tried to put a number on where the second wave was headed. It survives without a name on it.
A graphic without a byline
The document, titled “Potential for 175,000 Russian Forces Near Ukraine: Evidence of Recent Artillery, Equipment, Personnel Movements and Planning,” is posted on the National Security Archive’s website as an OCR’d (optical-character-recognized) rendering of an unclassified U.S. intelligence graphic, obtained at the time by The Washington Post. It is built around four DigitalGlobe satellite image captures, dated June 16, November 9, November 16, and November 26, 2021, laid under hand-marked troop and equipment estimates. The cleaned text carries no analyst’s name, no agency seal, and no case or file number. The Archive lists the posting date as November 26, 2021.
What the imagery marks
The graphic identifies newly deployed Battalion Tactical Groups, self-contained Russian army units built around a tank or motor-rifle battalion, at two named sites, Yelnya and Boyevo. Yelnya sits in Russia’s Smolensk region, roughly 75 miles (120 km) from the border with Belarus and about 160 miles (260 km) from Ukraine. The graphic also marks “recently arrived tanks and artillery” near Novoozernoye, in Crimea, and locates more than 50 BTGs positioned along the border and in Crimea overall. Current Russian troop strength, as the graphic states it, was 50,000 to 70,000 personnel.
The projection, not a finding
Beyond that count, the graphic offers a forecast rather than a tally. It puts potential force levels at up to 100 BTGs, or roughly 175,000 personnel, with as many as 50 additional reservist regiments worth about 100,000 more. The wording throughout is conditional: forces “expected,” strength “up to” a stated figure, movements that “may be” underway. One line states that equipment “may be left behind at different training ranges to enable a rapid final buildup.” Another suggests the visible troop movements could be meant “to obfuscate intentions and create uncertainty.” In the document’s own terms, this is a projection built from imagery and pattern, not a settled count of what existed on the ground.
Why an unsigned estimate is not unusual
Threat estimates built from commercial satellite imagery routinely go out without a name attached. Withholding the analyst or office keeps the sourcing method, imagery bought from a commercial provider in this case rather than classified collection, separate from the judgment drawn on top of it. Reporting from the period bears this out: The Washington Post and the Associated Press both cited an anonymous administration official for figures matching the graphic’s own numbers, including confirmation that close to half of the projected 175,000 was already deployed along the border. Ranges phrased as “up to” are likewise standard in open threat assessments, which set a ceiling rather than assert a final count. None of that makes the document unusual. It makes it a working estimate, stated as one.
Two dates on one document
The Archive’s page lists November 26, 2021, as the document’s publication date. The graphic itself, in its own printed caption, is dated December 3, 2021. That means the version on record was, on paper, published before the date it carries. Nothing in the posting explains the gap: no addendum, no note on which date governs, no correction. It is a fact about the paperwork, not evidence of anything beyond it.
No name, no seal
Separate from the date question, the cleaned text names no compiling office, no analyst, and no originating military or intelligence unit anywhere in the document. The Washington Post described the underlying material as unclassified, but no classification marking is legible anywhere in the OCR’d version posted by the Archive. That is silence in the paperwork. It is not, on its own, proof that a name or a marking was ever removed rather than simply absent from the version that survives.
Coordinates that don’t resolve
The OCR’d text attaches coordinate strings to Yelnya and to Boyevo that are internally duplicated and, in places, garbled, with digit groupings that don’t correspond cleanly to either labeled location. Nothing in the source corrects this or explains which string belongs to which site. It may be an artifact of scanning a hand-marked graphic rather than an error in the original briefing. The source gives no way to tell the difference.
What happened to the number
The document does not say. There is no later reference point anywhere in the source, nothing indicating whether the up-to-175,000 projection, or the reservist estimate stacked on top of it, was later confirmed, revised, or borne out by what actually assembled at the border. Broader accounts of the period place later invasion-era estimates in a similar range, but that reconciliation happens outside this document, not inside it. The graphic ends where its own arithmetic ends: a projection, unsigned, dated twice, and never followed up in the record that survives.