Every gold mine runs on assay results, the lab test that says how much gold is in a bag of crushed rock. The standard test is fire assay, which melts each sample with lead in a very hot furnace. Chrysos, an Adelaide company built on CSIRO research, replaces it with an X-ray machine called PhotonAssay, and it keeps ownership of every unit. It leases each one out, charges a monthly minimum whether the customer runs one sample or a million, and then charges again for every sample above the line.
That pricing choice is the most useful thing in this week's numbers. On 19 November Chrysos told the market that its usage charges rose 79% while the minimum payments rose 7%, and that nine of its last eleven leases came through laboratory companies rather than its own sales team.
General information only. Not financial advice or a recommendation to buy or sell any security. Data as of 20 November 2026.
What Chrysos does and who pays for it
PhotonAssay fires high-energy X-rays at a jar of crushed rock and counts the gold atoms that respond. The company says a unit analyses about 70 samples an hour, or roughly 40,000 a month, with no furnace, no lead and no chemical waste, according to its trading update presentation. Its customers fall into two groups. Miners put a unit on site to get grade-control results while the shift is still running. Commercial labs such as SGS, ALS, Bureau Veritas, MSALABS and Alfred H Knight install units in their own facilities and charge many miners for the assays.
Neither group owns the box. Chrysos owns each unit (it costs the company about A$4 million to build, it says), installs it in roughly eight weeks, maintains and upgrades it, and signs five-year leases that renew in five-year steps over a unit life of about 20 years. In plain terms, it runs like a finance company with a physics lab attached.
Why it is in the news
The pre-AGM trading update, released at 8:27am on 19 November ahead of the annual meeting on 25 November, reported unaudited revenue of A$36.7 million for July to October 2026, up 27% on the same four months last year. Samples processed rose 43% to 4.91 million, and October set a monthly record of 1.36 million. The company has signed 15 new leases so far in FY27, taking contracted units to 97, while the deployed fleet stands at 51 after one unit in Mali was decommissioned. It reaffirmed FY27 guidance of A$108 million to A$118 million in revenue and A$35 million to A$42 million in EBITDA, on constant currency.
The shares closed at A$6.70 that day, up 9.3% from A$6.13, according to Yahoo Finance daily prices. That is a sharp move for a company that only confirmed guidance, and we read it as a sign that some investors had expected less.
The read below the ticker
The pricing model. Revenue splits into two lines. Minimum Monthly Assay Payments (MMAP) are the floor a customer pays regardless of use. Additional Assay Charges (AAC) are billed per sample above that floor. In the four months to October, MMAP was A$22.6 million (up 7%) and AAC was A$13.7 million (up 79%). AAC is now 38% of PhotonAssay revenue, against 26% a year earlier. In FY26 the floor made up 71% of revenue, which the company puts at about A$1.43 million per unit a year at the FY26 exit rate.
That split explains the gap between sample growth (43%) and revenue growth (27%). The floor grows only when new units go live, and only five net units went live between the August results and November. Usage grows every time miners get busier, and with an elevated gold price they are busy. So the 27% is mostly existing machines working harder, which is good news for margins and a reminder that this half of revenue moves with the gold cycle in both directions.
The channel. Of the 11 leases signed since the August results, two went directly to miners and nine went to labs: four with SGS, four with ALS and one with Alfred H Knight, according to the 19 November trading update. Across all 15 FY27 leases, 11 came through labs. The FY26 results release (11 August 2026) says Chrysos has units deployed with all four major laboratory groups. When a lab takes a unit, Chrysos gets one contract but many end users, because every miner who sends samples to that lab is now a PhotonAssay customer without Chrysos ever pitching to them.
Retention and churn. The company does not publish a renewal rate, so we can't give you one. What the filings do show is the churn that happened: the MSA unit at Bamako in Mali was decommissioned (pre-announced in May) and is being redeployed in the region, a decommissioned MSA unit from Obuasi in Ghana is being redeployed in Morocco, and one MSA contract was cancelled and replaced with the Paramaribo unit in Suriname. Units can move between sites, which softens a lost contract, but a moved unit earns nothing while it travels.
Customer growth. The six units deployed in FY27 went to Bureau Veritas in Kalgoorlie, ALS in Rouyn-Noranda and Surrey in Canada, Aurenne's Mt Ida mine in Western Australia, and MSA in Suriname and Ivory Coast. Six more are being installed, including two Newmont mine sites in Ghana and Suriname. The FY26 results release says the first next-generation "XN" unit went to SGS in Perth.
AI claims. There aren't any to test. Chrysos describes itself as combining "science and software", and the trading update makes no artificial intelligence claims, so there is nothing to check against the numbers. We think that restraint is worth noting at a time when many hardware companies attach the label to anything with a sensor.
Early signals. Manufacturing cadence has risen to 18 units a year, the presentation says, and seven units were built between the August results and November. We found no hiring data or price changes in the filings, so we have left those out.

Chrysos daily closes, 19 Nov 2025 to 20 Nov 2026, with 50-day and 200-day moving averages. Source: Yahoo Finance, data as of 20 Nov 2026; averages calculated by HackrLife.
The numbers
The shares closed at A$6.78 on 20 November, according to the Yahoo Finance quote (Market Index shows the same close), about 22% below the A$8.65 close of a year earlier. The price sits just above its 50-day moving average of A$6.73 and just below its 200-day average of A$6.87, both calculated by us from Yahoo's daily closes to 20 November (Yahoo's own figures are A$6.74 and A$6.87). Both averages have drifted lower since the August results, when the shares jumped from A$6.45 to A$7.56 on 11 August, and the price has since given back most of that rise.
Measure | Figure | As of / source |
|---|---|---|
Share price | A$6.78 | Close, 20 Nov 2026, Yahoo Finance |
Market cap | A$791.5m | 20 Nov 2026, Yahoo Finance |
52-week range | A$5.18 to A$10.13 | 20 Nov 2026, Yahoo Finance |
50-day / 200-day average | A$6.73 / A$6.87 | 20 Nov 2026, HackrLife calculation from Yahoo closes |
Price to sales (trailing) | 9.0x | 20 Nov 2026, Yahoo Finance |
P/E (trailing) | 339.0x on net profit of A$1.79m | 20 Nov 2026, Yahoo Finance |
FY26 revenue | A$88.1m, up 33% | Year to 30 Jun 2026, ASX results, 11 Aug 2026 |
FY26 EBITDA | A$27.2m, up 68% | |
Jul to Oct 2026 revenue | A$36.7m, up 27% (unaudited) | |
Cash | A$19.5m | 31 Oct 2026, ASX trading update |
Debt facility | A$80m drawn, A$120m undrawn of A$200m | 31 Oct 2026, ASX trading update |
Debt to equity | 0.33 (total debt A$64.3m) | 30 Jun 2026 balance sheet, Yahoo Finance |
Free cash flow (trailing) | minus A$25.6m | Year to 30 Jun 2026, Yahoo Finance |
Two of those lines belong together. Operating cash flow was A$17.9 million in FY26, up 103%, according to the FY26 results release, yet free cash flow was negative A$25.6 million, because every new unit is a A$4 million build the company pays for up front. The cash pile fell from A$25.9 million at 30 June to A$19.5 million at 31 October, and drawn debt rose to A$80 million from A$60 million at the time of the August results (which listed A$140 million undrawn). Growth here is bought with borrowed money and repaid by leases.
We could not confirm Chrysos's S&P/ASX index membership from a primary source during this run, so we have left it out.
Who owns it
Ownership is concentrated. The figures below come from the most recent substantial-owner notices as listed on Market Index on 20 November. These notices are filed only when an owner crosses 5% or moves by a full percentage point, and they lag the trade by up to two business days, so a fund can trim within a percentage point without anyone seeing it.
Owner or insider | Stake or trade | Date |
|---|---|---|
Regal Funds Management | 10.46% (down from 12.73% for Regal and associates in an earlier notice) | Notice 12 Aug 2026 |
CSIRO | 9.28% | Last notice listed 3 Aug 2026 |
AustralianSuper | 6.18% | Last notice listed 3 Aug 2026 |
Adrian and Mary Ellen Knowles | 6.05% | Last notice listed 3 Aug 2026 |
Yarra Capital Management | 5.04%, new substantial owner | 6 Oct 2026 (lodged 8 Oct) |
Rob Adamson, chair | Exercised 25,000 options at A$4.50 (Appendix 3Y) | 3 Sep 2026 |
Kerry Gleeson, director | Exercised 7,500 options at A$4.50 (just outside 90 days) | 21 Aug 2026 |
Insiders / institutions (all) | 22.6% / 46.9% | Yahoo Finance, 20 Nov 2026 |
In the last 90 days no director traded shares on market, according to Market Index's director-transaction list (as of 20 November 2026). The only director moves were option exercises at A$4.50, which convert existing entitlements into shares and say little about what a director thinks of today's price. Earlier in the year the picture was mixed: the chair sold 529,020 shares on market at A$5.70 on 30 June, and the chief executive, Dirk Treasure, added 60,000 shares on market at A$5.98 on 10 June. The fund notices show more movement. Regal cut its stake in August, and Yarra crossed 5% in October.
The case others make
The company's own pitch, in the presentation, rests on unit economics. It says each unit earns about A$2.0 million a year in revenue against about A$0.51 million in cost, a gross profit of about A$1.5 million, which pays back the A$4 million build in about 2.7 years and gives a unit-level return on capital of about 37% a year (FY26 figures). It says assay spending is non-discretionary for miners, its gross margin runs at 70% to 80%, and the 97 contracted units already exceed its target of 90 deployed by the end of FY29. Managing director Dirk Treasure put the model in one line in the trading update: "We've built a business model that works throughout the mining cycle."
Independent coverage is more measured. Kalkine noted on 19 November that "a signed contract does not mean the equipment has already been installed or is producing assay revenue", and that cash and an undrawn facility are different kinds of liquidity. Small Caps reported the same figures and the fleet's above-90% availability. We did not find a broker note quoted in the press during this run that we could verify, so we have not cited one.
The main risks
The biggest risk is the gap between 97 contracted units and 51 working ones. Those 46 units earn nothing until they are built, shipped and commissioned, and each needs roughly A$4 million of capital the company mostly borrows. At 18 units a year, reaching 90 deployed units by mid-2029 means adding about 39 units in a little over two and a half years, with no slack for cancellations like the MSA one.
Second, usage charges are now 38% of PhotonAssay revenue, and they ride on gold-mining activity. The company itself cites "an elevated gold price" behind its positive industry outlook. If the gold price falls and exploration slows, the floor stays but the fastest-growing line shrinks.
Third, concentration. A handful of global lab groups now bring most new leases. That is an efficient channel while it works, and a lot of bargaining power in a few hands if it stops. Finally, the shares trade at about nine times trailing sales and about 339 times trailing earnings (Yahoo Finance, 20 November 2026), so the company's market value rests far more on future deployments than on today's profit.
What a builder can take from Chrysos
Chrysos charges a floor that covers its cost of serving a customer and lets usage carry the growth. If your customers' activity swings (seasonal shops, agencies with lumpy client work, creators with launch months), the same structure protects you on the quiet months without capping you on the busy ones. We covered a software version of this in unlimited seats, pay per outcome.
Take a worked example, which is our estimate. Say your AI transcription tool costs you A$40 a month per customer to run. Set the minimum at A$49 a month, including 300 minutes, and charge A$0.10 a minute after that. A quiet month still clears your cost. A busy month earns more without a sales call.
Work out your real monthly cost to serve one customer, including support time.
Set the minimum just above that figure, so a quiet customer never costs you money.
Price usage above the floor per unit the customer already counts (minutes, documents, orders), not per abstract credit.
Find the businesses that already serve your customers in bulk, the way labs serve miners, and offer them the product to bundle or pass on.
Track the share of revenue from usage each month. When it climbs, as Chrysos's did from 26% to 38%, your customers are getting more value than the floor charges for.
The second idea concerns the channel. Chrysos did not knock on every mine's door; it put machines into the labs that miners already trust, and nine of its last eleven contracts followed. The Southern Cross Electrical story made a similar point from the contractor side: the customer you win once can bring the next ten. If your customers are small businesses, your version of a lab is probably an accountant, an agency or a platform they already pay.
Chrysos makes its money when the mine is busy and keeps the lights on when it isn't, and most small software businesses could copy that pricing this quarter.
General information only. Not financial advice or a recommendation to buy or sell any security. Data as of 20 November 2026.

