Samsung E&A Co., Ltd.'s (KRX:028050) cash flow statement for FY2025 shows investing activities consuming ₩1,709,116,071,579.
Purchases of property, plant and equipment in the same year: ₩53,492,598,716.
So ₩1.66tn of investing outflow bought something other than plant, at a company whose entire property, plant and equipment balance is ₩507.5bn.
The balance sheet resolves it. Total assets were ₩10,013,337,796,132 at the end of FY2024 and ₩10,039,557,572,475 at the end of FY2025 — up ₩26.2bn. Nothing left.
What changed was the composition. Cash and equivalents fell from ₩2,595.7bn to ₩861.8bn, down ₩1,733.9bn. Non-current assets rose only ₩309.0bn. Trade receivables fell ₩586.5bn. Put those together and roughly ₩2tn moved into other current assets.
That is a treasury operation, not an investment. Money went from demand deposits into term deposits, short-term instruments and similar holdings that sit in current assets but no longer count as cash and equivalents. Under IFRS, moving cash into an instrument with an original maturity beyond three months shows up as an investing outflow even though the company is no less liquid.
An investor who looked at the cash line falling by ₩1.7tn and concluded that Samsung E&A had spent its money would have been wrong. The notes to the balance sheet name what the cash became, and that page is worth reading before drawing conclusions from the headline.
It partly reversed in 2026. Cash was ₩1,528.9bn at the end of March and ₩1,575.5bn at 30 June, with first-half investing outflows of only ₩204.7bn.
The relevant number is not the movement but the level. Samsung E&A holds liquid assets amounting to a large fraction of its ₩4,720.8bn of total equity, at a company whose fixed assets are ₩524.4bn.
The reason is FY2015. That year the company reported an operating loss of ₩1,454.3bn and ended with total equity of negative ₩323.3bn. It survived through a rights issue that took share capital from ₩200.0bn to ₩980.0bn. What kills a contractor in that situation is not the loss itself but the liquidity squeeze that follows — banks pull guarantee lines, clients demand more security, and the company cannot post the bonds needed to keep bidding.
A management team that lived through that keeps cash. It is expensive insurance and it is rational insurance.
The other reason for the buffer is visible in the volatility of everything else.
Operating cash flow: negative ₩459.8bn in FY2023, positive ₩1,635.8bn in FY2024, positive ₩254.4bn in FY2025, and ₩971.5bn across the first half of 2026 on the cumulative basis Korean interim statements use. A ₩2.1tn swing between two consecutive years.
Trade receivables move just as sharply. They were ₩3,504.8bn at the end of FY2024, ₩1,353.4bn at the end of September 2025, ₩2,918.2bn three months later, ₩2,370.8bn in March 2026 and ₩2,686.3bn at 30 June. A ₩2.2tn range inside eighteen months.
That is what percentage-of-completion accounting on large EPC contracts looks like from outside. Revenue is recognised as work progresses; billing happens at milestones; the gap between the two sits in receivables and moves by trillions depending on where a handful of projects are in their schedules.
Current liabilities of ₩5,402.7bn out of ₩5,776.9bn total tell the other half — customer advances and payables to subcontractors, which fund the work and unwind as it completes.
A company with those swings on both sides of working capital needs a large liquidity buffer to avoid ever being a forced seller of anything.
Holding ₩2tn-plus in deposits is not free. At a Korean policy rate of 3.00%, that money earns perhaps ₩60bn a year before tax. The same capital deployed in the business, or returned to shareholders, would do more.
Finance income was ₩241.7bn in FY2025 against ₩278.6bn in FY2024 — meaningful, and roughly 30% of operating income, which tells you how large the cash pile is relative to the operating business.
The company has started to release some of it. Dividends went from effectively nothing in FY2024 to ₩129.4bn in FY2025, with ₩169.3bn already paid in the first half of 2026, under a policy that moved the payout from 17% to 25% of controlling-interest net income.
Order intake gives it a reason to keep the rest. Samsung E&A booked ₩7.6tn of new work in the first half of 2026 against a ₩12tn annual target, including ₩3.4tn of chemical plant contracts. Mobilising on that many projects consumes working capital before any of it comes back.
The defence writes itself and it is strong.
A contractor's balance sheet exists to reassure clients and bonding banks, not to optimise return on assets. A Middle Eastern national oil company awarding a multi-billion-dollar EPC contract looks hard at whether the contractor can survive the project. Samsung E&A's liquidity is part of what wins the work, and after 2015 it needs to be visible.
The FY2025 movement was also, on the evidence, entirely benign. Assets flat, liquidity intact, and the cash line partly restored within two quarters. Nothing was lost; a line item moved.
And the company is releasing capital in the right order: first rebuild equity from negative ₩323.3bn to ₩4.72tn, then start a dividend, then raise the payout ratio when results beat guidance. That sequence is hard to criticise.
The breakdown of current financial assets in the half-year report notes — the maturities and instrument types of whatever the ₩2tn became. Short-dated deposits are one answer; anything with duration or credit risk is another.
Second, whether the payout keeps rising. A company holding this much liquidity against a rebuilt balance sheet has room to go past 25% of net income, and the FY2026 decision will show whether the board thinks the insurance can be reduced.
Third, the receivables balance at the year end. It has swung by more than ₩2tn in eighteen months. A December figure above ₩3tn, with operating cash flow weak, would mean the ₩7.6tn of new orders is being funded out of the buffer — which is exactly what the buffer is for, and also the point at which it stops being spare.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.