Cadence greener grass

The traditional model
In a traditional market sector, consolidation generates two positive effects: better price control, and lower costs. Combined they generate higher profits which translate in stronger stock, which generates additional investments. Consolidation means less competition, of course, since there are less producers and thus supply can be better regulated to stay a bit below demand, a sure way to raise prices.

Consolidation normally creates redundancies in the administrative and often in the production and support sectors of the resulting organization: thus it is possible to reduce staff positions and cut costs, especially overhead, or indirect, costs.

In the case of Cadence and Mentor administrative, marketing, sales, and support organizations can be significantly smaller than the sum of the present two, or at least this is the reasoning by the financial analysts.

The fact that there are product redundancies is also not seen as a waste, but as an opportunity for divesting, at a profit of course, those products, either through an outright sale or through a spinoff.

So when EDA editors write against the proposals, they are seen either as biased, ignorant, or idealists. And often a combination of some of the above.

Why it will not work
The EDA industry is not like any other industry. It is a service industry and the relatively small number of significant buyers have alternatives not present in traditional industries.

The number of customers dealing with leading edge problems, the ones that purchase licenses that generate the largest profits margins, is becoming smaller with each fabrication node, normally an indicator that supply consolidation is desirable. But in our industry the best customers are also the ones with the greater access to alternative solutions.

Assuming that Cadence is successful and that it can impose a new pricing model on the industry, and supposing that Synopsys falls in line, this would mean two things: Magma will increase its share of the markets in all sectors with the exception of digital simulation and verification where it does not have a presence, and smaller companies would also see greater opportunities to expand. The result would be that in order to be competitive, Cadence would have to match what the customers want to pay, not the other way around.

Cadence is locked in too many standards, and of course it would be very difficult to undo Open Access, a reasonably straight forward tool for startups to "augment" capabilities in the installed Cadence base. The jump from "augment" to "replace" is not very long when the financial stakes are high enough. Cadence does not have unique technology in any sectors of EDA: it competes directly with at least one among Synopsys, Mentor, and Magma, in every one.

Demand can also be mitigated by IDMs and by Foundries by developing internal tools. Consolidation in the supply sector would mean that there would be a higher supply of well qualified engineers willing to work on proprietary tools. By the time it would take Cadence to settle the ripples from the merger, say two years, its customers would be in a position to counter any of its pricing moves. They are not going to wait and see, they are going to react immediately on all three fronts: roll your own, switch to a direct competitor, and nurture one or more startups.

So, although I have heard from a few financial analysts who think the acquisition is good for Cadence and for the industry, I continue to believe it will not work even if classical investment theory says it should.

See also:
Cadence bids to buy Mentor Graphics
M&M Would Be Sweet
Cadence-Mentor deal makes sense, says analyst