30 minutes to under five: De Beers ships SynthDetect 2
The screening instrument cuts at least 25 minutes off a full tray scan and needs less operator input, De Beers Group Science and Technology says. No price is published.
By The Diamond Desk
What changed
De Beers cut the scan. SynthDetect 2 takes a full tray scan to under five minutes from about 30, the company says, with more automation and less operator input.
What it means
- If you screen parcels in volume: the company claims at least 25 minutes off every full tray.
- If you run a grading lab: the instrument reports automatically and takes software updates.
- What this is not: a measured result. Every figure here is the company's own claim and no independent test is published.
Key figures
| Full tray scanSynthDetect 2 | under 5 min |
|---|---|
| Full tray scanthe previous version | about 30 min |
| Time removedat least, per full tray | 25 min |
| First SynthDetectthe original instrument | 2017 |
| SynthDetect XLthe second in the line | 2018 |
Source: De Beers Group Science and Technology release, reported 15 September 2026 · the company's own stated figures, not measured by Carat Capital
The claim is time
Six times faster. A full tray moves from about 30 minutes to under five, which is at least a six-fold cut on the company's own figures.
Labour comes out too. De Beers says increased automation reduces the need for operator input, which is where most of the cost in screening sits.
| Instrument | Released | Full tray scan | What it added |
|---|---|---|---|
| SynthDetect | 2017 | not stated | first of the line |
| SynthDetect XL | 2018 | not stated | tests set jewellery without a probe |
| The previous version | — | about 30 minutes | the company's own comparison point |
| SynthDetect 2 | 2026 | under 5 minutes | at least 25 minutes faster |
Notes on this table
Source. De Beers Group Science and Technology, as reported by Mining Weekly, 15 September 2026.
The 30 minutes. is attributed in the release to previous versions, not to SynthDetect XL. It is carried the same way here rather than pinned on a named instrument.
No price, tray capacity or throughput. is published, so a cost per stone cannot be computed.
The time saving. is the company's claim; no independent test was found.
Nine years since the first
2017, then 2018. The first SynthDetect shipped in 2017 and SynthDetect XL in 2018, both able to test set jewellery without a probe.
Eight years on. SynthDetect 2 follows the XL by eight years, the longest gap inside this line of instruments.
No price, no throughput
Nothing on cost. The release carries no price, no tray capacity and no throughput figure, so a cost per stone cannot be worked out.
Shown in Hong Kong. The instrument was demonstrated at the De Beers Show and Tell during the Hong Kong Gem and Jewellery Show.
What to watch
- Q4 2026First shipments, and whether a list price is published with them.
- 2027An independent scan-time test from a grading lab, which no source carries yet.
The story so far
Go deeper
What would change this call
An independent scan-time measurement, or a published price. Either would move these figures from a company claim to something a buyer can check.
Screening is not grading
A screening instrument sorts a parcel into stones that are certainly natural and stones that need a further test. It does not grade and it does not issue a certificate. The value of a screener is therefore throughput per hour and the false-referral rate, because every stone it cannot clear has to go to a slower and costlier instrument. A scan time on a full tray is the headline figure a manufacturer uses, but it means little without the tray's capacity, which this release does not give.
Method
Two reports of the same company release were read, Mining Weekly's of 15 September 2026 and African Mining Market's; the De Beers release itself was not fetched from the company's own domain and no address was guessed for it. Arithmetic, Carat Capital's: 30 minus 5 is 25 minutes, and because the new figure is stated as UNDER five minutes the saving is written as at least 25 minutes rather than exactly 25; 30 divided by 5 is 6, so the cut is at least six-fold and is written that way. From 2018 to 2026 is eight years, against one year from 2017 to 2018, which is why eight years is called the longest gap in this line; the claim is scoped to the SynthDetect line and to nothing else De Beers makes. Sarandos Gouvelis, the division's senior vice president, is quoted in the release; the line runs past this paper's 15-word cap and is paraphrased nowhere here because it carries no fact the figures do not. Archive test, both periods named: a grep of the 375-article archive returns zero hits for SynthDetect. Carat Capital's 7 September diamonds piece covers HRD Antwerp opening a Mumbai GRADING lab at 2,000 stones a day; this item covers a De Beers SCREENING instrument and its scan time. Different company, different function, different period, so NEW.