It was SEMICON China last week, and I've written a couple of posts about it this week. Talking about semiconductor manufacturing equipment a lot recently reminded me of a friend who used to work in the semiconductor equipment industry (and I did some consulting for him years ago). When I was having lunch with him, we got to discussing why EDA and semiconductor equipment sales processes are so different. At first glance, there are a lot of parallels between semiconductor equipment and EDA. Most notably: many of the same customers, the same technology treadmill, and a small number of large suppliers. There is also not a huge amount of differentiation in the product offerings, at least in terms that are easily assessed by the customer. To a first approximation, all plasma etchers etch, and all Verilog simulators simulate Verilog. A better assessment requires deciding more precisely what you want to do, and then running detailed tests. But there are also big differences, the two big ones being lead time and cost. Lead Time and Cost Semiconductor companies know that they need both new equipment for the fab and new design tools for their design groups in order to bring a new process node online. In general, the most advanced fabs work very closely with the equipment vendors on the spec of new equipment, and then on ensuring that the equipment works properly in the new environment, and they also work closely with EDA companies for the tool capabilities (and often IP) that they will require. If you think it is hard to get your hands on a netlist for a next-generation design, try getting your hands on some test wafers when most of the equipment does not yet exist. Finally, when the equipment is ready for production, the fabs have no expectation that they will get it for free in return for this work, though they will certainly drive for deep discounts. As my friend said, sometimes the customers think “JDP” stands for “jumbo discount program.” One big difference from EDA is the way equipment is sold. Of course, it is hardware, not software, which means that neither the salesperson nor the buyer knows the exact cost of goods, and so what the profit margin is at any particular price, although a few customers actually invest in “should cost” programs to work out what they think a piece of equipment should cost for negotiating leverage. Another big difference about hardware is that it has lead time. If you want to open your fab by such-and-such date, then the equipment needs to be ordered by a much earlier deadline. This makes the negotiation much more balanced: the equipment vendor can delay knowing that the clock is ticking. Yes, they want the order but the fab absolutely has to close a deal by a given day. This really doesn't occur with software since the buying company knows that they can purchase on the last day of the quarter without a problem. The other difference about equipment is that it really is a one-time buy, a true “permanent license.” You buy a piece of equipment this year and you pay for it this year. Next process generation, you don’t “rebuy” all your existing equipment with just a soupçon of new stuff such as better optics. But with software you do. So even though a new piece of equipment may contain a lot of the previous generation in its design, especially in the software, the semiconductor company doesn’t expect to get that bit for free, on the basis that they already paid for it in the previous generation. EDA The way EDA works, even the old days when EDA still had a hardware business model and sold permanent licenses, there was always a debate at the customer as to how much of a new product was incremental (thus expected to be included as part of maintenance) or was a new tool (thus required a new license to be purchased). Today, with time-based licenses, much of a salesperson’s quota may be “re-selling” the existing capability. When so much is riding on just keeping the customer on-board using the existing tools, the salesperson can become very risk-averse about selling new products. Unless the customer insists on buying, it is only a small amount of incremental revenue for possibly a large amount of incremental disruption. Having sold a house last year, I noticed a similar dynamic with realtors: they make their money by closing a sale contract, but while an extra $10,000 was important to me as the seller, it barely changed the realtor's commission. Their incentives are to close any deal, and not risk the deal for a little $10,000 incremental upside. It is almost surprising that realtors do seem to put a lot of effort into getting a good price (but you can also read Why a Real Estate Agent Might Skip the Extra Mile ). I joined Cadence in 1999 with the acquisition of Ambit, and Cadence salespeople of that era largely behaved like this, and were not interested in selling a few Ambit synthesis licenses as part of a larger deal. The incremental revenue was minimal, the risk to closing the deal over a product that the salesperson was unfamiliar with was high. Plus, the Cadence sales teams were used to doing deals that were, in effect, "we'll supply all your EDA needs except synthesis" since almost every company got their synthesis tools from, shall way say, another vendor. This behavior was despite the fact that aggressively selling synthesis was (and is) a strategic imperative, in some ways more so now since the integration is tighter than just reading and writing Verilog netlists. An equipment salesperson is more like an EDA startup salesperson. If he or she doesn’t sell new equipment, there isn’t anything else to sell on the price list. The older equipment may still be on the price list but there is not much demand. Very little ramping of production goes on except in the latest processes. When you hear of new designs being done in 130nm, say, then they are running in fabs built years ago, not that some foundry decided that was an unmet opportunity if only the built a new 130nm fab. That may be changing a little since fabs with process nodes off the leading edge are being refurbished, and perhaps expanded, rather than decommissioned. According to SEMI, the market for second-hand 8" equipment, which used not to exist, has gotten hot. But I don't think it extends to anyone building new old-node fabs. More like adding electronic ignition to an old car. Emulation One part of all of the big three EDA companies' product lines is emulation (and FPGA prototyping, just assume emulation means all that). This is much more like the semiconductor equipment market. It is hardware, it has lead time, it has non-zero cost, it is expensive. Also, from a finance perspective, it is not sold at one moment, with the revenue recognized over future years (except probably some maintenance). Hardware bookings turn to billings and revenue the second the hardware ships. This is one reason that financial analysts are always asking lots of questions on earnings calls about hardware sales. They want to know how much of the revenue came from one-time bookings that will not be repeated automatically next quarter, as opposed to ratable business, where the revenue recurs for many quarters to come. The smart analysts know that selling a few emulation platforms (revenue recognized immediately) has the potential to cover up a weakness in some other part of the business for a time, and they are paid to find that sort of thing out before it is obvious to everyone else. Sign up for Sunday Brunch, the weekly Breakfast Bytes email.Image may be NSFW.
Clik here to view.![]()
Clik here to view.