SK hynix Liquid Assets and Cash Flow Analysis
SK hynix Liquid Assets and Cash Flow Analysis
I'll start by loading the deep-research workflow, then pull SK Hynix's latest filings and quote data.
Conclusion: SK hynix's cash build is real, but the "tripled" headline needs correcting, and part of the increase is not operating cash. Its liquid assets rose about 4.1x year on year to KRW 114.5T. About KRW 39.1T of the nine-month build came from a one-off ADR-linked share issue earmarked for capex. Excluding that and the shareholder payouts, the liquid-asset build is about KRW 71T. Operating cash generation is strong: nine-month operating cash flow was KRW 145.6T and free cash flow about KRW 113.7T. Investment needs are now rising faster, though. Management guides 2027 capex significantly higher, and unrecognized capex commitments have jumped. I think the memory-driven improvement looks durable through 2027, because volumes are largely contracted. Beyond that, the 80% operating margin is price-driven, and its durability depends on pricing and on capex discipline. I did not check the share price, so this is not a valuation view.
Data as of: company financials through 30 September 2026 (Q3 results released 28 October 2026; quarterly report filed 16 November 2026). I did not use current market quotes.
1. Checking the "tripled" premise
- SK hynix alone: cash and liquid assets rose from KRW 27.85T to KRW 114.46T, "more than quadrupling" (). The ratio is . Formula: 114.46 ÷ 27.85.
- Combined with Samsung: KRW 136.3T to KRW 329.23T (), or . Formula: 329.23 ÷ 136.3.
- So "tripled" fits neither company's figure exactly. The SK hynix-only jump is larger.
- The KRW 114.5T is broader than the cash line. The 30 September balance sheet shows cash and equivalents of , short-term financial instruments of and short-term investment assets of . Together they total .
2. Cash generation (nine months to 30 September 2026)
Formula for FCF: 145.634314 − 30.998642 − 0.960924. Formula for capex intensity: (30.998642 + 0.960924) ÷ 145.634314.
Quality caveats on operating cash flow:
- It includes of dividends received, which is non-operating in nature.
- Income tax paid rose to from .
- Q3 operating margin was on revenue of KRW 104.6T and operating profit of . Calculated, it is . Formula: 83.3338 ÷ 104.6155.
- Q3 net income was , down 58% from Q2. The gap is non-operating: a valuation loss on investments and a FX loss ().
3. Where the cash build came from (31 December 2025 to 30 September 2026)
- Liquid assets went from to , a change of .
- The largest single non-operating item is a paid-in capital increase of . It was a third-party allotment of 17.79 million shares to Citibank, for ADRs, closed 14 July 2026, at a total of (). Press reports say the proceeds go to the Yongin fab, Cheongju packaging and EUV tools. The 6-K itself does not state the use of proceeds.
- Excluding that raise and adding back buybacks and dividends, the build is about . Formula: 79.517349 − 39.142782 + 28.841739 + 1.866227.
- Long-term investment assets rose to from , a increase. The 9M cash outflow for these acquisitions was . I did not verify what these holdings are, so I treat them as an unresolved item.
- Net cash was against borrowings of .
4. Investment needs
- 2026 spending is already large. Korea Herald reports SK hynix spent in nine months, above its full-year 2025 total of . The Q3 PP&E figure differs between sources: (PRISM, citing the DART summary) against (Korea Herald). The definitions differ, so I use the 6-K line items.
- Management says capex in 2027 will "increase significantly" versus 2026, for the Yongin fab ramp and leading-edge equipment (). It gave no 2027 number. It also says it will phase clean-room openings and adjust investment to demand.
- Press reports put unrecognized capex commitments at at September, against at year-end, a increase. This is a secondary figure and I have not verified it against the filing.
5. Shareholder returns
- The 2026 target is about : at least KRW 30T in dividends and at least KRW 60T in buybacks and cancellations. The company says this meets its commitment to return at least 50% of free cash flow.
- Nine-month cash payouts were in buybacks and in dividends. The 2026 target equals 88% of nine-month free cash flow. Formula: 100 ÷ 113.674748.
- The Q3 dividend is , including a from Kioxia gains. The company says most Kioxia gains will be returned, so the special element is non-recurring.
- The Solidigm IPO review was discontinued. That leaves Solidigm capacity to be funded from the group's own balance sheet ().
6. Durability of the memory-driven improvement
Supports durability (through 2027, at least):
- Long-term agreements with 14 major customers cover more than 50% of 2027 and 2028 capacity. Terms are typically five years, with prices set within agreed bands that reflect market conditions, and customer deposits as safeguards ().
- Management expects DRAM and HBM volumes through 2029 to be effectively sold out, including volumes under negotiation. It expects supply to lag demand through 2030. These are management claims, not independent evidence.
- For 2027, management guides to mid-20s% DRAM bit growth and high-teens% NAND bit growth, constrained by supply. It cites hyperscaler capex growth of more than 60% in 2027 from market research ().
Risks to durability:
- Cycle risk from capex. Commitments of KRW 42.7T plus a 2027 capex step-up raise the risk of overbuild if demand slows. This is the classic memory-cycle failure mode.
- Pricing and customer pacing. Analysts pressed on whether AI spending will be sustained. Management says customers are pacing spend but that the trend holds ().
- Demand substitution. Management expects PC and mobile shipments to fall because of higher memory costs. It also flagged HBM stack-height reductions by some customers.
- Financial volatility. Q3 shows large non-operating swings, and FX moves hurt Q3 results.
- Dilution and payout dependence. The ADR issue diluted shareholders by about 2.5%. Special dividends depend on investment gains, so they are not recurring.
My assessment. Cash generation is likely to stay elevated through 2027 because prices are high and volumes are contracted. The more fragile part is the 80% margin and the 2028+ capex-to-cash balance. The signals to watch are: the 2027 capex number when guided; whether LTA coverage stays above 50%; whether unrecognized commitments keep rising faster than revenue; and whether the 2026 payout is sustained from operating cash flow rather than the one-off capital raise and investment gains.
Sources
- SK hynix, 3Q26 results release, 28 Oct 2026: news.skhynix.com
- SK hynix Q3 2026 earnings call transcript (Motley Fool), 28 Oct 2026: fool.com
- SK hynix Form 6-K, quarterly business report to 30 Sep 2026 (consolidated cash-flow and balance-sheet statements), filed Nov 2026: sec.gov
- SK hynix Form 6-K, July 2026 (ADR paid-in capital increase result): sec.gov
- Yonhap News, 17 Nov 2026: en.yna.co.kr
- Korea JoongAng / Korea Herald, 17 Nov 2026: koreaherald.com
- PRISM by Liabooks, 10 Nov 2026 (secondary; cites DART Q3 summary): prism.liabooks.com
- Global Economic (g-enews), 17 Nov 2026 (secondary; commitments and ADR proceeds use): g-enews.com
- Korean-language search results on ADR proceeds (e-focus, K-Stock Briefing, MK), Jul 2026 (secondary).