Anatomy of a 44% Revenue Collapse and What the Numbers Actually Say
De Beers is heading for a negative underlying EBITDA in the first half of 2026 — the first time the world's most storied diamond company has gone into the red at the half-year mark under these conditions. The disclosure came not in a De Beers statement but buried in the Q2 2026 production report of Anglo American, its 85% owner.
The headline damage, second quarter:
|
Metric (Q2 2026, consolidated)
|
Value
|
YoY change
|
|---|---|---|
|
Rough-diamond revenue
|
$665m
|
−44%
|
|
Consolidated sales volume
|
6.0m carats
|
−11%
|
|
Total sales volume
|
7.1m carats
|
−7%
|
|
Average selling price
|
$110/ct
|
−37%
|
|
Average price index (like-for-like)
|
—
|
−17%
|
|
Sights held
|
3
|
flat
|
And for the half:
|
Metric (H1 2026, consolidated)
|
Value
|
YoY change
|
|---|---|---|
|
Consolidated revenue
|
$1.31bn
|
−23%
|
|
Consolidated sales volume
|
12.4m carats
|
+13%
|
|
Total sales volume
|
14.8m carats
|
+20%
|
|
Average consolidated price
|
$105/ct
|
−32%
|
|
Average rough-price index
|
—
|
−16%
|
|
Production
|
14.9m carats
|
+46%
|
Read that table twice, because it contains the entire story: De Beers sold 20% more diamonds and collected 23% less money for them.
1. The single most important number in the report
Everything else is commentary on this: the 37% price decline in Q2 versus a 17% decline in the like-for-like price index.
That 20-point gap is not a rounding artifact. It is the difference between what the market is paying and what De Beers is selling.
The price index measures like-for-like movement — the same category of goods, period over period. Down 17%. That is the market's verdict on rough diamond values. Bad, but survivable.
The average selling price is simply revenue divided by carats. Down 37%. That number falls both when prices drop and when the goods in the box get cheaper.
The gap is therefore almost entirely sales mix. Anglo American says so directly: an increased proportion of lower-value goods "due to the current inventory mix."
Why this matters for interpretation
There are two very different stories that produce a −37% average price, and the trade needs to be clear about which one it is looking at.
Story A — market collapse. Rough prices are in freefall across the board. This would be catastrophic and would show up in the index. It has not; the index is down 16–17%.
Story B — inventory clearance. De Beers has been working through a stockpile weighted toward cheaper, smaller, lower-quality goods — precisely the categories under lab-grown assault — and the mix shift is dragging the average down mechanically.
The evidence points firmly to Story B, layered on top of a genuine but more moderate Story A. Anglo American's own language confirms the bifurcation:
"Lab-grown diamonds continued to impact demand for lower-value natural diamonds, adding pressure in more price-sensitive categories… However, stronger pricing for higher-value goods supported a stable overall average price index throughout the period."
That is a two-tier market stated in plain corporate English. The top is holding. The bottom is being eaten.
2. The volume paradox, decoded
The most counterintuitive figure in the release is the H1 combination:
Total sales volume +20%
Consolidated revenue −23%
Volume up a fifth, money down nearly a quarter. In a healthy market this combination is impossible. Here it is diagnostic.
De Beers is moving carats, not value. The company is clearing inventory into a soft market, and the composition of what it is clearing is dominated by the categories where pricing power has evaporated. Each additional carat sold in H1 2026 generated dramatically less revenue than the carats sold in H1 2025.
This is what destocking under pressure looks like on a P&L. It is not necessarily the wrong strategy — carrying a stockpile of depreciating small goods into a lab-grown-saturated market is worse — but it is unambiguously margin-destructive in the period it occurs.
The Q1-to-Q2 reversal
Note also the shape within the half. Q1 2026 showed a sales increase, following a price cut. Q2 then fell 44%. The half ended lower despite the Q1 improvement.
The read: the Q1 uptick was price-elasticity response, not demand recovery. Cut the price, clear some goods, book a quarter. When the discount had done its work, Q2 reverted. That is a one-time lever, not a trend reversal — and it has now been pulled.
3. The index disclosure that deserves close attention
Two sentences in the report carry disproportionate analytical weight:
Since Q1 2026, the price index now incorporates the cut-price inventory sales De Beers conducted in 2025.
The index does not include the price reduction implemented at the July 2026 sight.
Take these in order.
On point one: the 2025 discounted inventory sales — the ones that generated persistent trade reporting about whether certain sightholders received special pricing — are now inside the index base. That means the index has absorbed a lower reference point. A −16% move measured against a base that already includes discounted transactions is arguably a harsher signal than the same number measured against undiscounted 2025 pricing. The index has been re-anchored downward.
On point two, and this is the more important one: the July 2026 price reduction is not yet in these figures. De Beers eased rough pricing at the July sight. That cut lands in Q3. Whatever the index does next, it starts from a lower price deck than anything reflected in this report.
The practical implication is straightforward and uncomfortable: the H1 numbers are not the trough on price. They predate the most recent cut.
4. Production: the disconnect that must be closed
While revenue collapsed, output surged:
Q2 production +88% to 7.8m carats
H1 production +46% to 14.9m carats
Anglo American attributes this to three factors:
A favorable comparison base — an extended maintenance shutdown at Orapa (Botswana) depressed H1 2025
Planned mining of higher-grade ore at Jwaneng (Botswana)
Planned mining of higher-grade ore at Gahcho Kué (Canada)
So a meaningful part of the 88% is optical — last year was artificially low. But higher-grade ore is real incremental supply.
The problem with producing more into this market
Mining higher-grade ore in a soft market is operationally rational (it lowers unit cost) and commercially awkward (it adds supply precisely when demand is weak). Combined with destocking, it means more carats chasing less demand — which is the mechanism that produced the mix and price outcome above.
De Beers appears to recognize this, because the second-half plan is explicitly corrective.
The supply-discipline response
Full-year guidance is unchanged at 21–26 million carats. With 14.9m already produced in H1, hitting even the top of guidance requires H2 output of roughly 6–11 million carats — a sharp step down from the first-half run rate.
Three levers deliver it:
The announced two-year production pause at Venetia (South Africa)
Planned maintenance at Orapa (Botswana)
Planned maintenance at Jwaneng (Botswana)
The Venetia suspension is the headline item and the most significant. A two-year pause at a major underground operation is not a scheduling tweak — it is a capital and production retrenchment with employment, community, and sovereign-relations consequences in South Africa. Companies do not idle assets of that scale for two years unless the demand outlook is genuinely poor.
Anglo American's framing — that De Beers "continues to monitor trading conditions to align output with demand" — is the classic language of supply-side management. It is also the oldest tool in De Beers' kit, and its return signals that the company has moved from managing a downturn to structurally resizing for one.
5. The loss trajectory: context on the EBITDA line
The negative H1 EBITDA needs to be read against the recent record:
|
Period
|
Underlying EBITDA
|
|---|---|
|
FY 2024
|
−$25m
|
|
H1 2025
|
−$189m
|
|
FY 2025
|
−$511m
|
|
H1 2026
|
Negative (guided)
|
The pattern is a widening loss over three consecutive reporting cycles. A $25m loss in 2024 was noise. A $511m loss in 2025 was a crisis. The H1 2026 figure has not been quantified, but given revenue down 23% on higher volumes and a further price cut arriving in July, there is no obvious mechanism by which it improves materially on the H1 2025 −$189m base.
Note also the accompanying corporate event: Anglo American halved De Beers' carrying value in connection with the 2025 results. The write-down and the operating losses are two expressions of the same judgment.
6. The sale process, and the elephant in the report
Anglo American CEO Duncan Wanblad was measured:
"Progressing the sale process for De Beers, while concurrently advancing streamlining opportunities to improve cost performance and reduce capital expenditure to minimize the impact from challenging diamond markets."
Decoded: we are selling, and in the meantime we are cutting. "Reduce capital expenditure" and a two-year Venetia pause are the same sentence expressed two ways.
What Anglo American conspicuously did not address: reports that it had selected a consortium led by former De Beers CEO Gareth Penny as preferred bidder.
That silence is notable. The company confirmed the process is progressing while declining to comment on the specific bidder identified in press reports days earlier. Standard practice during live negotiations — but it means the market is being asked to price De Beers' assets on a set of numbers that are deteriorating while the buyer identity remains officially unconfirmed.
The valuation problem
Consider the position of any bidder:
The asset is loss-making and the losses are widening
Its carrying value was halved by the seller
A further price cut hit in July, after the reported period
Its most price-sensitive product categories face structural, not cyclical, competition
It is idling a major mine for two years
The seller is a visibly motivated vendor with a publicly stated intention to exit
Every one of those is a negotiating lever for the buyer. A deteriorating H1 does not strengthen Anglo American's hand. If anything, this production report is a data point in the buyer's favor — which is precisely why the timing of the sale process relative to the earnings trajectory matters commercially.
The counter-case for a buyer: this is a cyclical trough purchase of the single most valuable brand in the category, with supply discipline already implemented by the outgoing owner and a cost base being cut on someone else's watch. Buying De Beers at a halved valuation with Venetia already paused is a meaningfully different proposition from buying it two years ago at full value with full production.
7. The three pressures, separated
Anglo American names three drivers. They deserve separate treatment because they have very different durations.
(a) Geopolitical and macroeconomic uncertainty — cyclical
"The geopolitical and macroeconomic landscape remains uncertain, with the onset of the conflict in the Middle East adding to economic and consumer-confidence risks."
Conflict-driven consumer-confidence shocks are real and they hit discretionary luxury hard. They are also, historically, the pressure that reverses. Diamond demand has recovered from every macro shock in the modern era. This is the component of the downturn that time may solve.
(b) Lab-grown competition in lower-value categories — structural
"Lab-grown diamonds continued to impact demand for lower-value natural diamonds, adding pressure in more price-sensitive categories."
This one does not reverse. Synthetic production capacity is scalable, its costs continue to fall, and it competes most directly at the small, commercial-quality end — exactly the goods weighing down De Beers' inventory mix.
The strategic consequence is severe: a meaningful share of De Beers' production is in categories where the long-run price floor is being set by a manufactured substitute. No amount of supply discipline fixes that, because De Beers does not control synthetic supply. This is not a downturn in those categories. It is a re-basing.
(c) Inventory mix — self-inflicted and self-correcting
The mix drag is the consequence of choices — carrying then clearing a stockpile weighted toward affected categories. It is painful in the period but it is finite. Once the inventory is through, the mix effect stops flattering the downside.
Why the distinction is the whole analysis
The pressures separate cleanly into one that heals (macro), one that does not (lab-grown), and one that ends (mix). Anyone valuing De Beers — bidder, sightholder, or supplier — needs to size the lab-grown component specifically, because that is the portion of the earnings decline that is permanent.
And here the report offers the genuine sliver of good news: "stronger pricing for higher-value goods supported a stable overall average price index." The high end is intact. If the lab-grown threat is concentrated in commercial goods while large, fine, and rare stones continue to firm, then De Beers' long-term franchise sits in the segment that is holding — but the company must shrink and reposition to live there.
8. Implications by constituency
For sightholders and rough buyers
Three sights held in Q2, same as last year — allocation cadence is intact, but value per sight has collapsed
The July price cut is not in these numbers; expect Q3 economics to reflect a lower deck
The two-tier market is now official: plan on continued softness in commercial goods and relative firmness in better goods
Venetia's pause plus Orapa/Jwaneng maintenance means H2 rough availability tightens materially — an argument for securing better-quality supply now
The 2025 discounted-inventory episode is now inside the index base; historical price comparisons need care
For Indian and other cutting centers
Rough is cheaper, which helps input costs — but polished realizations in affected categories are falling in parallel
The squeeze is concentrated in small and commercial goods, the highest-volume, most labor-intensive segment. That is an employment issue, not just a margin issue.
H2 supply reduction may tighten rough availability just as manufacturers were adjusting to abundance
For retailers and brands
Natural diamond cost of goods at the low end is falling — an opportunity to hold price and widen margin, or to compete harder on value
The high end is firming, which supports premium positioning and reinforces the rarity narrative
Consumer confidence risk from the Middle East conflict is a genuine near-term demand variable
For Anglo American shareholders
De Beers is a widening drag: −$25m (2024) → −$511m (2025) → negative H1 2026
The disposal is the strategic answer, and Wanblad's cost and capex actions are designed to limit bleed during the process
The risk is that a deteriorating asset in a live sale process invites a lower clearing price
9. What to watch
The quantified H1 EBITDA loss in Anglo American's interim results — "negative" is guidance; the number is the story. Benchmark against −$189m (H1 2025).
Q3 pricing, including the July cut. This is the first period reflecting the latest reduction. It is the cleanest read available on where rough pricing actually sits.
H2 production versus guidance. Delivering 21–26m carats for the year after 14.9m in H1 requires genuine restraint. Whether De Beers hits it tests the credibility of its supply-discipline commitment.
Confirmation on the Penny consortium — bidder identity, structure, price, and whether Botswana's interests are accommodated.
The price-index gap. If the 20-point spread between average selling price and the index narrows, the mix drag is clearing. If it persists, the inventory problem is deeper than stated.
Whether the high end holds. The entire bull case rests on "stronger pricing for higher-value goods." If that cracks, the two-tier thesis fails and the problem becomes categorical.
Venetia's social and sovereign fallout. A two-year pause at a South African operation of that scale has consequences beyond the P&L.
Marketing spend. In every prior downturn, category demand generation was the lever. A vendor cutting capex ahead of a sale has little incentive to fund it — which raises the question of who invests in natural diamond demand during the interregnum.
Closing view
Strip away the percentages and one sentence remains: De Beers sold 20% more diamonds in the first half of 2026 and collected 23% less money.
That is not a demand story alone, and it is not a price story alone. It is a composition story — a company clearing an inventory weighted toward exactly the goods that a manufactured substitute has structurally re-priced, into a market softened by conflict and macro uncertainty, while its own mines were running higher-grade ore and its owner was preparing an exit.
The response is textbook and, on its own terms, correct: cut supply (Venetia, Orapa, Jwaneng), cut cost and capex, hold guidance, and lean on the segment that is still working — high-value goods, where pricing is described as stronger and the index as stable.
But two facts sit uncomfortably alongside that response. The July price cut is not yet in these numbers, which means H1 is not the floor on price. And the lab-grown pressure is structural, which means a portion of this earnings decline will not come back regardless of how the macro cycle turns or how disciplined supply becomes.
De Beers is being resized for a smaller, higher-value natural diamond business. The open questions are how small, at what price the asset changes hands, and who pays for the demand-building that a company in the middle of a sale process has no incentive to fund.