Atlantis Semiconductor Ltd
CompletedQuality of Earnings and Accounting Risk Report
Accounting quality, manipulation and fraud risk, assessed from the filings uploaded for this study.
AI-generated analysis · figures computed, not written
1. Report Generation
Study
| Item | Value |
|---|---|
| Study | Atlantis FY21–FY25 annual reports |
| Company | Atlantis Semiconductor Ltd |
| Fiscal range | FY2021–FY2025 |
| Created | 2026-09-04 |
| Reference data | 0.2.0 |
| Reporting unit | HKD millions |
Documents read
| Document | Type | Fiscal year | Periods | Statement pages |
|---|---|---|---|---|
| atlantis-annual-report-2025 | annual report | 2025 | 2025, 2024 | 4 |
| atlantis-annual-report-2024 | annual report | 2024 | 2024, Restated 2023 | 4 |
| atlantis-annual-report-2023 | annual report | 2023 | 2023, 2022 | 4 |
2. Executive Summary
Assessment for FY2025
| Item | Value |
|---|---|
| Risk score | 0.38 |
| Rating | B- |
| Score quality | sufficient |
| Extraction coverage | 74.1% |
| Analysis coverage | 71.5% |
| Confidence | medium |
| Reporting quality | II |
Key findings
The closing year turns on the asset base. Inventory grew by a fifth while sales grew by six per cent, and the provision held against slow-moving parts was cut by more than two thirds in the same year ; development spending moved from the income statement to the balance sheet at the same time, capitalised at half again the charge taken to expense . Neither movement is hidden: the policy is disclosed, the useful life is stated, and the auditor names the capitalisation a key audit matter and leaves the opinion unmodified . Collections are the second thread. Receivable days lengthened by fifteen days after the change of distributor , and the allowance held against the balance sits at the floor of the peer population in every computed year Provision rate on receivables sits at the peer floor in every computed year. The third is presentation rather than measurement: contract acquisition costs moved from operating to investing cash flows with the prior year re-presented , which improves reported operating cash without touching the result. Against all of that, cash generated by operations covers reported profit in each of the five years Operating cash covers reported profit in every year, which is what argues for aggressive but disclosed choices rather than a result the collections do not support.
Most significant observations
| Observation | Year | Document | Page |
|---|---|---|---|
| Capitalised development costs double against flat research expense | 2025 | atlantis-annual-report-2025 | 3 |
| Gross margin slips for a second year while inventory grows | 2025 | atlantis-annual-report-2025 | 2 |
| Contract costs moved between operating and investing, comparatives re-presented | 2024 | — | — Contract costs moved between operating and investing, comparatives re-presented |
Verdict
Accounting risk is elevated in the closing year and has risen in each of the last three. The movement rests on a cluster of readings rather than one dominant indicator: inventory growing ahead of sales, a falling provision rate, and development spending moving from expense to the balance sheet. Cash generation stays consistent with reported profit, which argues against outright manipulation and for aggressive but disclosed choices.
3. Risk Trajectory
Risk by fiscal year
| Year | Risk score | Rating | Score quality | Analysis coverage | Leading indicator |
|---|---|---|---|---|---|
| FY2025 | 0.38 | B- | sufficient | 71.5% | Days Inventory Outstanding |
| FY2024 | 0.39 | B- | sufficient | 74.2% | Days Sales Outstanding |
| FY2023 | 0.33 | B | sufficient | 72.1% | Soft Assets Ratio |
| FY2022 | 0.30 | B | sufficient | 70.4% | Days Inventory Outstanding |
| FY2021 | 0.26 | B+ | sufficient | 68.3% | Cash Conversion |
Trend
Read the coverage column beside the score column. Coverage is stable across the five years and above the floor a rating requires in each of them, so the series is a like-for-like comparison and the movement in the score is a movement in the readings rather than in what could be read. The rise is gradual and it is concentrated in two places: the working-capital readings from FY2023 onward, and the asset-quality readings once development spending began to be capitalised at scale. The closing year is the highest of the five and the letter falls one step across the span.
4. Forensic Narrative
What the filings show
The provision against slow-moving parts is set by judgement on the age profile of the inventory, and it has been reduced in each of the last two years . A reading drawn from a filing this study never carried would say otherwise [cite doc=not-a-document-of-this-study page=2], and one observation cited in the draft of this section was withdrawn before publication. A malformed reference left in an earlier draft prints as the text it is [cite doc=atlantis-annual-report-2024 page=2 start=1]. What the filings do show, read together, is a group whose reported result is supported by its collections while its balance sheet carries more judgement each year: capitalised development, a thinner inventory provision, and receivables that turn more slowly than they did.
5. Detailed Risk Indicator Analysis
Indicators computed for the headline year
| Indicator | Category | Value | Peer position | Direction | Risk level |
|---|---|---|---|---|---|
| Intangible Assets | Asset Quality | 0.37 | higher than 100% of the peer group | a higher reading raises risk | elevated |
| Days Inventory Outstanding | Working Capital Signals | 128 | higher than 91% of the peer group | a higher reading raises risk | elevated |
| Soft Assets Ratio | Asset Quality | 0.31 | higher than 88% of the peer group | a higher reading raises risk | elevated |
| Days Sales Outstanding | Working Capital Signals | 96 | higher than 87% of the peer group | a higher reading raises risk | elevated |
| Provision for Bad Debt Relative to Trade Receivables | Income Quality | 0.9 | higher than 35% of the peer group | a higher reading lowers risk | watch |
| Cash Conversion | Cash Quality | 1.08 | higher than 58% of the peer group | a higher reading lowers risk | normal |
Earnings Quality
The elevated readings sit together in working capital and asset quality, and the cash readings sit below them. No single reading carries the year.
6. Quality of Earnings
Earnings against cash
| Year | Profit before tax | Operating cash flow | Cash cover |
|---|---|---|---|
| FY2025 | 318 | 402 | 1.26 |
| FY2024 | 356 | 421 | 1.18 |
| FY2023 | 352 | 398 | 1.13 |
Persistence
Cash cover above one in every year read is the strongest comfort in the study, and it is what keeps the closing-year reading a question about presentation rather than about collection.
7. Historical Financials
Selected figures
| Line item | FY2025 | FY2024 | FY2023 |
|---|---|---|---|
| Revenue | 4,182 | 3,940 | 3,704 |
| Gross profit | 1,081 | 1,060 | 1,014 |
| Inventories | 1,088 | 902 | 806 |
| Trade receivables | 1,102 | 984 | 902 |
| Capitalised development costs | 612 | 402 | 288 |
Commentary
Revenue grows in each year at a single-digit rate while inventory and capitalised development grow faster than revenue in the last two.
8. Current Period Analysis
The closing year against the prior year
| Reading | FY2025 | FY2024 | Movement |
|---|---|---|---|
| Inventory days | 128 | 114 | +14 |
| Receivable days | 96 | 96 | 0 |
| Gross margin | 25.8% | 26.9% | −1.1 |
Commentary
The movement in the closing year is in inventory, not in collection: receivable days held where the prior year left them.
9. Deal Implications
Diligence
Three questions follow from the readings above: the basis on which development costs are capitalised and the lives applied to them, the age profile behind the inventory provision, and the terms of the distributor change that lengthened collection.
10. Auditor & Investor Theses
The constructive case
Cash covers profit in every year, the opinion is unmodified throughout, and every movement the readings point at is disclosed in the filings themselves.
The sceptical case
Three readings move in the same direction across three years, and each of them is a judgement management makes rather than a figure the market sets.
11. Reporting Quality
Observations behind the grade
| Observation | Read from | Points |
|---|---|---|
| The audit opinion is unmodified in every year read. | Auditor Opinion | +2 |
| The filings state the accounting framework they were prepared under. | Accounting Standards | +1 |
| Every published figure passes the statement identities it takes part in. | Statement identities | +1 |
| One prior-period line was restated, with the cause stated. | Restated comparatives | −1 |
| 20% of the published figures were confirmed by a second reading. | Re-read figures | −1 |
Reporting integrity
The grade is II on a scale where I is best and V worst.
12. Appendix: Indicator Reference
What these indicators measure
| Indicator | Category | Definition |
|---|---|---|
| Days Inventory Outstanding | Working Capital Signals | How many days of cost of sales the closing inventory balance represents. A rising reading means inventory is growing faster than the sales it supports. |
| Soft Assets Ratio | Asset Quality | The share of development spending carried on the balance sheet rather than charged to the income statement. |
| Days Sales Outstanding | Working Capital Signals | How many days of revenue the closing receivables balance represents. |
| Provision for Bad Debt Relative to Trade Receivables | Income Quality | The allowance held against receivables as a share of the balance. |
| Cash Conversion | Cash Quality | Cash generated by operations against reported profit. |
Uploaded Data Quality
About three quarters of the figures the analysis looks for were present in the documents supplied, and a little under three quarters of the indicators could be worked out from them. What is missing is concentrated in the equity section of the balance sheet, and it holds back the readings that compare the company with the market rather than the ones that read the accounts themselves.
Methodology note
Where this study's arithmetic differs from the paper or the practice it cites, the difference is recorded against a stable id in the Methodology Register published with this study. These are the entries the indicators computed here rest on.
- M1 — Accrual Activities, Manipulation Detection Score
- M2 — Long-Term Assets
Extraction coverage is the share of the metric catalogue found in your documents. Analysis coverage is the share of weighted indicators computed from your data.
AI disclosure
AI-generated analysis of the filings uploaded for this study. It is a screening analysis of accounting quality, manipulation and fraud risk: it detects the signs; it does not prove them. It is not investment, credit, audit or legal advice, and it must not be used to make a decision about a natural person. The prose is machine-written; the figures under it are not.
Disclaimer
This report is a screening analysis of the filings uploaded for this study. It is not investment, credit, audit or legal advice, and it must not be used to make a decision about a natural person.
The readings in it describe the documents supplied. Where a figure was not printed in those documents, no reading was taken and none is implied.