The Wagner Group is a Russian private military company that, since 2014, has deployed across multiple African states, providing security, combat, and training services frequently exchanged for resource-extraction rights. ² Its African footprint began in earnest in the Central African Republic: around 175 Russian instructors, described as including Wagner personnel, entered the country in late 2017 after a UN arms-embargo exemption, a figure independently reported by both the Congressional Research Service and Sweden's FOI. ² By early 2018, ACLED records the group operating in CAR under an agreement exchanging Russian military support and weapons for mining concessions. ¹

The entity sits at the intersection of state and private interests. President Putin admitted in June 2023 that the Kremlin provided some 86 billion rubles — approximately $940 million — to Wagner between May 2022 and May 2023, an admission independently reported by the US State Department and RUSI. ¹ The analyst's exchange-rate check (86 billion rubles divided by $940 million yields roughly 91.5 rubles to the dollar) is consistent with the 2022-2023 period. Entities linked to Wagner financier Yevgeny Prigozhin — Lobaye Invest, Midas Ressources, Meroe Gold, M-Invest — recur across the resource-extraction record. ³

Running against this documentation is a consistent wall of official denial. Russia has denied any involvement with Wagner in Africa, saying the companies work on their own, while host governments acknowledge only 'Russian trainers'. ² This report addresses each of the brief's sub-questions in turn — presence, abuses, resource extraction, combat operations, and the official record — before testing the competing explanations that the evidence supports.

The strongest evidence of presence comes where independent origins agree on numbers. In CAR, the ~175 instructors of late 2017 are reported by two independent sources. ² The 2021 peak is convergent: the Congressional Research Service reports some 2,100 personnel while RFE/RL cites up to 2,300, an approximately 9% variance the analyst classifies as order-of-magnitude agreement. ¹ In Mali, three independent sources — FPRI, Inter Populum, and CRS — report a September 2020 agreement for roughly 1,000 Wagner contractors. ² In Libya, both the Jamestown Foundation and CRS relay AFRICOM's June 2020 claim of no less than 2,000 Russian mercenaries fighting alongside the Libyan National Army. ¹

Mozambique illustrates how numeric variance need not mean dispute. CTC Sentinel reports a contingent of some 160; the Euro-Atlantic Council of Slovenia reports around 200 arriving in September 2019. ² The analyst treats the 40-person gap as differing estimation methods, not conflicting accounts, since both agree on a 2019 deployment and subsequent withdrawal after casualties. The Euro-Atlantic Council figure carries a caveat: its source bears a future publication date and is graded REPORTED. ¹

The presence record thins at specific points. The claim that Wagner numbers in CAR reached 1,400 by May 2018 rests on Wikipedia alone and is graded REPORTED. ¹ Placed in the CRS trajectory — 175 in late 2017, then 2,100 in 2021 — the 1,400 figure is not arithmetically contradicted; the analyst notes it would imply roughly 700% growth in the first six months, then slower expansion. ² Country counts also vary by definition: RAND lists a 'clear presence' in six countries in 2025 — Burkina Faso, CAR, Libya, Mali, Niger and Sudan — while the UK Parliament's seven-country count since 2014 includes now-withdrawn operations such as Mozambique. ² The analyst flags that 'presence' and 'operations since 2014' are different thresholds, not competing tallies.

Open: Whether the single-sourced 1,400 CAR figure for May 2018 can be independently corroborated.; Whether Niger's inclusion in RAND's 2025 presence list reflects deployment comparable in scale to CAR or Mali.

The documentation of civilian harm is broad and drawn heavily from the highest-tier sources. In CAR, ACLED records that Wagner mercenaries attacked a Fulani community at Besson in September 2021, killing at least 40, and that since December 2020, 180 civilian-targeting events accounted for 52% of Wagner's engagements in organized political violence. ² The analyst's arithmetic confirms this figure is internally consistent (180 events divided by 0.52 yields roughly 346 total engagements). Human Rights Watch called for ICC and Special Criminal Court prosecution over the killing of 13 civilians near Bossangoa in 2021. ¹ UN bodies documented continuing violations into 2024 and 2025: MINUSCA recorded 147 violations affecting 110 victims in Dembia and Rafaï over one week in October 2024, and an attack near Mboki in January 2025 that killed at least 12 people, including six children. ¹

In Mali, ACLED reports that civilian targeting accounted for 69% of Wagner's engagements as of August 2023. ¹ The analyst cautions against directly comparing this to CAR's 52%: the percentages cover different countries, periods, and conflict environments, and reflect operational context rather than necessarily different Wagner behavior. ³ HRW documented Malian forces and apparent Wagner fighters executing and disappearing dozens of civilians since December 2022, and in 2025 reported the execution of at least 12 Fulani men and forced disappearance of 81 others, including an incident where about 100 men were arrested at Sebabougou in April 2025 and at least 43 bodies were later found near a military camp. ²

The Moura operation of March 2022 is both the most documented incident and the sharpest evidentiary fault line. The UN OHCHR concluded at least 500 people were killed over five days, a finding relayed independently by ACLED, RAND and the UK Parliament, with RAND adding that the UN found the vast majority executed and at least 58 women and girls raped. ² Against this, IRSEM, citing roughly 30 HRW testimonies, reports about 100 Wagner members involved and about 300 civilians killed. ¹ The analyst classifies this as a genuine divergence — a 40% discrepancy for the same event — attributable to methodology: a comprehensive UN investigation versus a testimony-based count. The UK Parliament claim is itself graded CONTESTED for conflating the operation's timeframe. ¹ Readers should hold an uncertainty range of roughly 300 to 500+ fatalities for Moura, without a tiebreak the sources do not support.

Aggregate figures require the most caution. The State Department, citing Economist/ACLED data, reports Wagner killed at least 1,800 African civilians as of August 2023, and a 278% surge in terrorist violence against civilians in Mali since 2021; both are single-origin. ¹ ACLED's CAR-specific count of at least 786 fatalities from 362 incidents between 2018 and 2024 is likewise single-source. ¹ The analyst notes the CAR figure would represent about 44% of the continental total were the periods aligned — but the periods do not align, so the comparison is illustrative only.

Open: Whether primary UN and ACLED data can reconcile the 300-versus-500 Moura death toll or whether the two counts measure different populations.; Whether the single-source continental figure of 1,800 civilian deaths and the 278% Mali surge can be independently verified against primary datasets.

The resource-for-security model is documented across CAR and Sudan through investigative journalism and official sanctions. In CAR, the founding agreement itself exchanged Russian military support and weapons for mining concessions. ¹ Lobaye Invest, established on 25 October 2017, secured at least eight mining permits. ¹ On 17 March 2020 the CAR government awarded Midas Ressources a permit for the Ndassima gold mine, nullifying Canadian company Axmin's 2010 permit; CSIS reports the Wagner-linked shell company then led operations there for three to four years. ² The US Treasury's June 2023 sanctions named Midas Ressources as holding rights to Ndassima, valued at over $1 billion — a Treasury valuation of reserves, which the analyst records as a single-source official figure. ¹ Logging is documented too: African Business reports a 30-year concession over approximately 186,596 hectares allocated to Russian-linked Bois Rouge in February 2021, though both the concession's precise size and the company's March 2019 registry details rest on that single investigative outlet. ¹

Sudan shows the same architecture through leaked documents. OCCRP and C4ADS independently establish Meroe Gold as a 99% Sudanese subsidiary of Prigozhin-owned M-Invest, serving as cover for Wagner's operations. ¹ Both report that Bashir's office in August 2018 ordered the ministry to waive its 30% stake in a gold operation in favor of Meroe Gold. ¹ OCCRP's documents also detail payment terms between M-Invest and a local partner: $100,000 monthly plus $200,000 up front and $500 per person brought into Sudan. ¹

The economic model does not appear to be extraction-only. Inter Populum reports Mali's junta was in talks to pay $10.8 million per month for 1,000 mercenaries — a cash arrangement the analyst notes annualizes to roughly $129.6 million. ¹ Other sources cite approximately $10 million monthly, an 8% variance the analyst treats as rounding or different contract phases, cautioning readers to read the fee as order-of-magnitude ~$10 million. ¹ The critique of these arrangements is voiced from within CAR: former Prime Minister Martin Ziguélé accused Wagner of pillaging resources such as Ndassima without paying taxes to the state or local communities. ¹ A further, uncorroborated dimension comes from i24NEWS, which reports several hundred Wagner fighters running a tramadol-trafficking operation along the upper Oubangui River; the claim is single-source, graded REPORTED, and carries a future-date anomaly. ¹

Open: Whether the Ndassima 'over $1 billion' figure reflects documented extraction and revenue or an estimate of reserves.; Whether the tramadol-trafficking operation attributed to Wagner on the upper Oubangui can be independently corroborated beyond the single source.

Primary-source and high-confidence documentation of combat operations spans several theaters. The UK Parliament's Foreign Affairs Committee assessed with high confidence that Wagner conducted military operations in at least seven countries since 2014, including Ukraine, Syria, CAR, Sudan, Libya, Mozambique and Mali. ¹ In Libya, AFRICOM's June 2020 assessment of no less than 2,000 Russian mercenaries fighting for the LNA is relayed by two sources. ¹ In Sudan, The New Arab reports US officials revealing in April 2023 that Wagner offered surface-to-air missiles to RSF chief Hemedti — a single-source, REPORTED claim attributed to unnamed officials. ¹

Mozambique is the clearest documented instance of operational failure. CTC Sentinel reports the contingent of some 160, deployed in 2019, largely withdrew after failing to make an impact and taking casualties, corroborated by the Euro-Atlantic Council's account of around 200 arriving in September 2019 and beginning to withdraw two months later. ² The deployment followed President Nyusi's August 2019 trip to Russia to secure military help for Cabo Delgado, though that specific trip is single-sourced and graded REPORTED. ¹

Mali's Moura operation is the most consequential combat episode in the record. RAND describes it as a joint FAMa-Wagner five-day operation beginning in March 2022, and the UK Parliament reports a UN fact-finding mission verified Wagner involvement. ² IRSEM's testimony-based account puts roughly 100 Wagner members on the ground. ¹ Mali also generated an information-operations dimension: France accused Wagner in April 2022 of staging a 'French atrocity' with a mass grave near Gossi, a claim carrying France's documented interest in countering Russian influence in Francophone Africa. ¹

Open: Whether leaked documents or visual evidence can confirm the Wagner surface-to-air missile offer to Hemedti beyond the single US-official account.; What primary evidence underlies France's 'staged atrocity' accusation at Gossi.

The most consistent feature of the government record is a recurring semantic distinction: officials deny 'Wagner mercenaries' while acknowledging 'Russian trainers' or 'instructors.' Russia denied any involvement with Wagner in Africa, saying the companies work on their own. ¹ Mali denied the December 2021 deployment while acknowledging Russian trainers were present, a denial reported by Wikipedia, VOA and IRSEM independently. ² CAR's presidential spokesman told Bloomberg that Wagner mercenaries are not operating in the country but confirmed 'unarmed Russian trainers' assist the army. ¹ The analyst characterizes this as a terminological distinction reflecting political positioning, with evidence suggesting the same personnel are involved. ³

Russian spokespeople have paired denial with characterizations of the evidence as fabricated. RFE/RL reports Russian officials repeatedly denied crimes despite substantial evidence from rights activists and UN experts, with Kremlin spokesman Peskov stating advisers 'could not participate and did not participate in murders or robberies. ¹ This is another lie'. ¹ Foreign Ministry spokeswoman Zakharova called abuse allegations 'purely unfounded' fabrications intended to tarnish Russia's image, and envoy Bogdanov dismissed reports of Russian mercenaries in Libya as relying on 'fabricated data'. ² Wagner itself denied operating in Sudan for more than two years in an April 2023 Telegram statement. ¹

The pattern extends to states where Western officials also deny a Wagner role. The Kremlin and President Traoré denied Wagner's presence in Burkina Faso in February 2023, and — in the record's clearest counterexample — unnamed US officials said in April 2023 that Burkina Faso had not employed Wagner. ¹ As of the State Department's report, Mali's transition government had still not publicly acknowledged Wagner's presence. ¹ The official record is not monolithically hostile to Russia: CAR official Fidel Gouandjila credited the Russian defense accord with helping avoid a 'fratricidal war' and saving CAR democracy, and IRSEM cites a Gallup poll finding Mali had the world's highest approval of Russian leadership at 84%, a single-source figure graded REPORTED. ²

Open: Whether internal government communications or contracts explicitly naming Wagner exist to test whether 'Russian trainers' are operationally distinct from Wagner personnel.; Whether the single-source 84% Mali approval poll reflects durable public opinion or a specific survey window.

H1 — Wagner as a de facto extension of Russian state policy — is supported. If Wagner were genuinely independent, one would not expect the Russian head of state to admit direct funding; yet Putin acknowledged some $940 million provided over one year, independently corroborated. ¹ Resource entities cluster around Prigozhin, and the founding CAR arrangement traded Russian state assets — military support and weapons — for concessions. ³ The contradicting evidence is entirely denial-based; these are confirmed as statements made, not as refutations of the funding record. ³ What remains genuinely undocumented is a formal command-and-control structure, which is why the analyst rates this supported rather than settled.

H2 — that civilian targeting is systematic rather than incidental — is supported by convergent documentation. If abuses were isolated errors, one would not expect civilian targeting to constitute 52% of engagements in CAR and 69% in Mali, nor a repeated pattern of mass executions and disappearances across years and countries. ⁴ Russian denials do not engage the documented incidents. ² The discriminating evidence that would strengthen this to a doctrine finding — operational orders directing civilian targeting — is not in the record.

H3 — that governments knowingly contract Wagner while maintaining deniability through wording — is supported. The clearest illustration is Mali agreeing to accept 1,000 contractors while simultaneously denying Wagner and acknowledging only 'Russian trainers'. ³ The single counterexample is Burkina Faso, where US officials themselves said Wagner had not been employed — which suggests the semantic-deniability pattern is common but not universal. ¹

H4 — that Wagner is compensated primarily through resources rather than cash — is only plausible, not supported. Resource concessions are extensively documented in CAR and Sudan, but the Mali contract's roughly $10 million monthly fee shows cash payments are also substantial. ⁴ The analyst concludes both channels operate; the 'primarily resources' framing overstates what the evidence shows, and no full financial accounting exists to weigh the proportions.

H5 — that Wagner withdraws when facing setbacks — is weak. Mozambique fits: the contingent withdrew after casualties and operational failure. ² But CAR presence was sustained through more than 100 gun battles and rising personnel counts, and Mali's deployment persisted despite intense international pressure over civilian targeting. ⁴ Mozambique reads as an exception rather than a rule; the analyst rates the hypothesis weak.

Open: Whether any leaked operational orders or command documents exist to convert H1 and H2 from 'supported' to demonstrated.; What a full financial accounting of Wagner's African revenue would show about the resource-versus-cash balance under H4.

The evidence forces several conclusions. Wagner operated across at least seven African countries since 2014, an assessment made with high confidence by a national parliament and consistent with a separate six-country presence list from RAND. ² The group is funded by the Russian state on Putin's own admission. ¹ Systematic civilian harm in CAR and Mali is documented by UN bodies, ACLED and HRW across multiple independent investigations. ⁴ Resource extraction through Prigozhin-linked front companies is established by leaked documents and official sanctions. ² These findings rest on convergent, high-tier sourcing and are stated declaratively.

Where the evidence is weaker, the wording weakens with it. The single worst incident, Moura, remains a genuine stalemate on scale: the UN's 500+ and HRW-derived 300 cannot be reconciled from the material here, and the report does not adjudicate between them. ³ Several quantitative pillars — the 1,400 CAR personnel, the 786 CAR fatalities, the 1,800 continental deaths, the 278% Mali surge — are single-source and should be treated as indicative rather than established. ³

The official denials, though confirmed as statements, do not survive contact with the documented record. The recurring 'Russian trainers, not Wagner mercenaries' formulation is, on the analyst's reading, a terminological device rather than a factual distinction — but the report notes that Burkina Faso is a real exception, where US officials themselves declined to place Wagner. ³ Speculatively, and labeled as such: if the resource-for-security model documented in CAR and Sudan is Wagner's template, the sustained presence in CAR despite heavy combat suggests deployments track the value of extractable assets more than battlefield success — a reading consistent with, but not proven by, the contrast between resource-rich CAR's persistence and resource-poor Mozambique's withdrawal. ⁴ This inference goes beyond what the claims establish and is offered only as a hypothesis for further investigation.

Open: Whether a definitive, reconciled Moura death toll will emerge from ongoing UN or judicial processes.; Whether the correlation between resource value and deployment durability holds when tested against fuller country-level financial data.