Analysts lift Samsung Electro-Mechanics' third-quarter profit forecast 30% despite a stronger won

The consensus is already net of the currency. When the won-dollar rate falls, dollar revenue converts into fewer won and exporters' reported profit shrinks ^[5], so the 605 billion won in FnGuide's figure ^[1] is what brokerages expect after the stronger won has taken its share. Their judgement that the AI orders win, or rather that they win by whatever part of the currency loss the brokerages have modeled, rests on a net estimate 344.7 billion won above last year's result ^[3]. The report does not give the size of the won's move, the share of sales billed in dollars, or how long the 1.0722 trillion won MLCC contract runs ^[8].

Extra volume explains only part of the revision. Lines running at 91% at the end of June ^[9] can add about 10% more output by running flat out ^[4], while the profit estimate for July to September rose about 29% ^[2]. Operating leverage lets profit grow faster than volume, and price and mix supply the rest. An AI server takes more MLCCs than a conventional server, and higher-capacity ones, and the report ties that to a global shortage and rising prices ^[7]. FC-BGA demand is moving toward larger substrates with more layers, where the market expects better prices and margins ^[12].

Currency is the first risk to the forecast. If the won keeps strengthening after the estimates were set, the quarter prints below 605 billion won, and KB Securities' 660.1 billion won and iM Securities' 650 billion won ^[4] become the top of a range; KB sits about 9% above consensus ^[5]. Supply is the second. The Philippine plant and new MLCC lines in Busan together lift capacity by up to 20% a year ^[11], aimed at the same shortage that produced this quarter's price increases ^[7]. KB Securities' case for durability is contractual: it expects package substrates to stay in severe shortage and the company to sign long-term agreements with advance payments and margin protection clauses ^[13].

I think the September quarter lands close to consensus, since lines this full and a record contract leave currency as the main near-term risk. The exposure starts once Calamba is producing ^[10]. The company has also chosen where its scarce output goes. A widening share of FC-BGA and MLCC sales sits under long-term agreements with big technology firms ^[12], so it sells less of the shortage at spot prices and takes advance payments, cash received while it is still building ^[13]. The counter-case is that those contracts cap what the spot market would have paid until the new lines arrive. The view is wrong if this year's agreements come without the margin protection KB Securities expects ^[13], or if MLCC prices fall once Calamba runs ^[10].