The IBM 308X line, launched in 1980, was the last family of mainframes that were sold in a climate of scarcity. Customers were lined up to get the machines as fast as IBM could pump them out. IBM held lotteries to assign delivery positions. An early delivery was worth a fortune. One American lessor got lucky. He ordered a machine and was assigned a very early slot. He found a bank that got a poor position and wanted desperately to upgrade from a 3033. The bank paid $1m over IBM list for the lessor’s 3081. Then the bank unloaded its 3033 for a lot more than it would have been worth several months later, recovering a chunk of its investment. This wasn’t an isolated incident. Other lessors devoted most of their energy to playing the positions market. During the 308X generation, IBM changed its practices, not only for assigning machines to customers but for production planning, too. IBM’s chairmen – Cary at first, then Opel poured concrete all over the Hudson Valley in New York and at its other mainframe production plants around the world. By the time the 3090s were in full swing, there was no such thing as mainframe scarcity. Markets being, well, markets, the increased supplies of equipment led to a more rapid decline in prices than might have otherwise occurred. The simultaneous rise of alternatives to the mainframe compounded the effect, as did the dedication of IBM’s mainframe rivals.
By Hesh Wiener
When the world’s industrial economies took a nosedive a few years ago, IBM’s mainframe production capacity was not just ample, it was excessive. And IBM’s decision to invest so heavily in facilities that could make only mainframes looked like a poor strategy. There’s a shortage of IBM mainframes again. We understand that a 9021 ordered today can’t be supplied until September, maybe later. IBM seems willing to accept this situation, even if it means that some extra business will go to Amdahl, Hitachi, Comparex or Olivetti. We suspect that IBM cannot easily increase production. The company has the plant capacity but not the personnel. Temporarily hiring back former employees is apparently not practical. And even if IBM could build more mainframes, it might not want an oversupply to put pressure on prices of the new equipment or on the residual values of machines leased by its financing group. If customers hold their old machines longer, IBM still gets software and maintenance revenue, so the impact is bearable. And if an IBM-compatible machine goes where an IBM one might have, IBM gets the software revenue from that customer, too. The situation may not change for a year or more. IBM’s 967X hardware won’t be fully developed nor equipped with all the right software for a long time. But the scarcity might not bring about a rise in mainframe prices, not a long-lasting one, anyway. And the new (or actually very old) IBM strategy that leads to shortages may get a few customers concerned about the future availability of 967X machines. IBM might be planning skinny production runs for these, too. This is not the sort of situation that builds customer confidence – not in IBM, not in mainframes and not in their own ability to formulate strategies.
From Infoperspectives International, June 1994 Copyright (C) 1994 Technology News Ltd London NW1 8JA.
