I woke this morning with my iPod Touch alerting me to Oracle buying Sun. It took me a few minutes to get it through the sleep haze. When things cleared the entire concept brought a big smile to my face.
I've know for many years that Oracle has been sniffing around the semiconductor world, trying to find a profitable way in. They've looked at Sun avariciously through that entire time because it had the greatest profit potential, but it was either too expensive or had too many suitors. The stock crash solved the expense issue and one by one the suitors went away with IBM finally bowing out. I hadn't thought they would move this fast but they must have had everything ready to go for this moment.
There will be lots of people wonder what this will mean, but I have a pretty good idea. The design and building of semiconductors is primarily a database issue, but one that is decades old technologically, incredibly fractured and extremely inefficient. Who better to take it over than a company that has essentially defined the database industry. The problem, however, is that IC design is also a highly specialized field and Oracle did not have that expertise at had, nor the management chops to deal with that kind of technology. In buying Sun, they have purchased both.
Sun and the EDA world have long had a love/hate relationship. The nature of IC design requires big server farms and Sun provided much of that hardware, both as a service and as product, but the EDA world kept trying to find ways to break their dependence on big Sun iron and many tools now work on souped up laptops. That didn't sit well with Sun over the years. And since Sun was, in itself, a semiconductor company as well, it made EDA pay for it's infidelity when buying EDA tools. Oracle adding its database knowledge to Sun's semiconductor experience makes is possible for Oracle to radically reinvent, and maybe even replace, the struggling EDA world.
EDA is hampered by many things but there are three issues I see as the greatest problems, they refuse to effectively market their technology because they believe that the engineers that use their products are immune to marketing (like it's some sort of black magic). Second, they don't really try to sell their technology outside of their ever decreasing circle of customers. Third, they are stuck in the technology paradigm that birthed them in the 1980s and refuse to consider standards and technology that could make their tools more efficient, therefore less expensive... ah, and less profitable ... even though it might help their customer base to grow.
Oracle has none of that problem. They are a marketing behemoth and ruthless competitors in their markets. They don't cut their prices to push competitors out of the market but make their products so feature rich and valuable to customers that it is almost a no brainer to buy them. And if that doesn't work, they buy the competition. But what is even more important is they keep reinventing themselves. If they need to obsolete a technology to make the market grow, they do it.
All this makes the purchase of Sun by Oracle a very dangerous outcome for established EDA companies who expect to continue competing with this big dog and a very welcome move for all the little EDA companies trying to compete against the hidebound EDA leaders. Oracle now has the resources to completely change how chips are designed and, at the same time, design the chips themselves. They will market the hell out of their approach and bury the big four if not consume them. And they will actively encourage innovation through startups, investing in them and acquiring the best of the lot.
A few months ago, Rajeev Madahavan at Magma predicted that there would only be two EDA firms left before five years are gone. I disagreed with him to a certain extent, but I might now agree with him. There will be Oracle... and the EDA Consortium made up of a whole bunch of startups. I think we might also see the end of EDA conferences, like DAC and DesignCon. All of that will be moved into OracleWorld.
That's my take. An I'm sure a lot of you are going to say I'm completely nuts. We'll find out in about five years. In the meantime, Big EDA might want to start putting together their contingency plans for when the Oracle M&A ream comes knocking on their doors.
