Acquiring Verigy: from memory to SoC, and the world test lead (2011)

Binding the volume of memory to the function of SoC

At the heart of this decision was that a company standing at the front of the world in memory placed a non-memory tester of a wholly different character inside itself, whole, by acquisition rather than alliance. Tester technology divides in two: the volume of memory, measuring great numbers of identical devices fast, and the function of SoC, measuring different functions apart on a single machine. Advantest had mastered the first but had not reached the front in the second. To take in Verigy for about $1.1B (¥91bn) was a decision not to buy the time to grow the missing half in-house, but to acquire the bearer of non-memory testing together with its Western customer network. The lead was born where the two technologies came into one company’s hands.

Yet assembling the two technologies and earning from them are separate things. The first year of the acquisition sank into a cyclical trough and a loss; the fruit did not ripen at once. It ripened only when the age arrived in which AI demanded unprecedented testing of both computation and memory. The 2011 choice to bind the volume of memory to the function of SoC was a hedge that would drop nothing whichever of them grew. Not the assembling of the vessel in itself, but whether one can lay out in advance a form that catches demand wherever it swells — the roughly $1.1B (¥91bn) Advantest paid was its answer to that question, the price of holding the lead in an industry where the semiconductor’s protagonists keep changing.