Tera Microsystems Inc, the two-year-old Santa Clara, California designer of building block chip sets for Sparcsystem manufacturers, has run out of cash and was closed last week by its financial backers. Tera claimed that its 50MHz MicroCore chip set was the only 50MHz set to be production worthy so far. Tera founder Henri Uehara, re-installed as president a few weeks ago after a management restructuring ousted the incumbent, Perry Considine, says he was faced with a shortfall of funds and no time to develop a story to take to the investment community. He was unable to secure even bridge financing from existing backers Philips Signetics, Sequoia Capital, Mitsubishi and Vertex Management, the Silicon Valley-based venture capital arm of Singapore Technology. Together they have put at least $6.48m in the firm over two rounds. Empty pockets dashed Tera’s hopes of bringing out and shipping in April an unannounced MicroServer, a $3,000 two-slot Sbus card developed over the last two quarters, that it was convinced could generate $20m to $30m by the end of 1993. Its 50MHz chips were reportedly also just coming out of fab and its operating system implementation is finished. Last week Uehara was working with only a skeleton staff hoping to find a buyer for the company’s MicroServer technology as well as the MicroCore chip sets more commonly associated with the company. The rest of Tera’s 30 employees were laid off. Uehara said too many investors are questioning the Sparc marketplace because Sparc cloners are dropping out. Observers, however, attribute Tera’s demise to flaws in its basic design which eschewed using either Sun’s Sbus or Mbus, but went for a proprietary bus instead. They argue that this factor limited Tera’s utility, making its chip appropriate only for laptops. Last October, Uehara said, Tera began to re-evaluate the market, realising that just selling chips wasn’t enough to make it successful. That epiphany led Tera to a strategy that would also take it into board and subsystem products beginning with the MicroServer and followed by two or three others this year. The MicroServer would also have been the basis of a line of systems Tera put out itself, adding to the $20m to $30m in revenues it was projecting for the board alone. Tera’s re-evaluation of the market also apparently led it to conclude that Considine, with his semiconductor background, was not the right man at the helm. While he admits he needed a major round of financing, Uehara says he could have brought out MicroServer with $500,000. He describes it as twiceas fast in input-output transactions as Sun’s existing Sparcstation 2s but four times cheaper and reckons he could have done full-blown MicroServer systems at the same low overheads as MicroServer boards.
