Semiconductor Socket Wars – II

I’ve concluded the prequel to this article by listing out the following critical pre-requisites for fighting a socket war –

1. Know the type of war you are in
2. Know the terrain
3. Know your position

In this article I will list out the strategies that may work best in different scenarios during an Invader Defender type of war (the most common).

Invader Defender type of war

The offerings of a semiconductor company can be broadly classified into 2 categories –

1. Proprietary
2. Commodity

Simply speaking a proprietary chip is a chip that is manufactured by a single supplier and has no identical competition in the market. There may be alternate methods of achieving the functionalities of a proprietary chip but the design would be different. In other words replacing a proprietary chip sitting in a socket with an alternate chip requires a board redesign. When the same chip (same functionality and geometry) is manufactured by multiple suppliers then the chip is said to be a commodity. Replacing an incumbent commodity chip in a socket can happen seamlessly without a redesign. Any manufacturer would be aware of the proprietary or commodity nature of its offerings. Be rest assured the manufacturers strategy for fighting a socket war would depend a lot on the nature of the product!

A commodity chip sitting in any socket would be inevitably vulnerable to an invader-defender type of war. If a customer doesn’t need to make design changes to change the chip in a socket they are likely to continuously engage with multiple competing suppliers – not just to get the best price but also to secure supply. In fact in some cases customers are reluctant to design in proprietary chips, as they believe that having a single supplier for a product is very risky! Consequently in some cases manufacturers of proprietary chips do welcome competition. For example let’s consider the case of Everspin Technologies – the world’s only producer of MRAM (Magnetoresistive Random Access Memory) chips. In a recent interview, Everspin’s CEO, Philip LoPresti was asked whether he was worried about the fact that big companies like Toshiba, Micron and Samsung were working on MRAM. LoPresti’s response was interesting – “We would actually like other people in the market. If there are more people shipping the same product, then customers are more prone to engage in high-volume applications. This is because they don’t have to worry about being sole-sourced. Big companies don’t like it when one of the most important parts in their systems comes from only one company or one fab. So long term, we would like to proliferate MRAM to the biggest memory companies and so on. We don’t see that as a bad thing.” LoPresti’s opinion however cannot be extrapolated to be the industry’s standpoint. A lot of companies do thrive on creating innovative chips and love their position of the monopolist in their market-spaces.

While an Invader-Defender war is routine for a commodity chip, it cannot be concluded that a proprietary chip is totally shielded from such an attack. Consider a hypothetical example: You are selling a proprietary chip to a socket at a price of $5/unit to meet an annual demand of 1 Million units. Now a competitor comes in with an offer of $2/unit. The customer knows that designing for the competitor will be time consuming and expensive. However the customer also sees a benefit—$3M worth of annual savings! Not reacting in such a situation may cause damage to the incumbent supplier’s business!

To fight an Invader-Defender type of war it is critical to know the terrain and one’s position in the terrain. If the terrain is flat then more often than not the war is down to a price-war. When is a terrain flat? When multiple competitors are selling identical chips and there is no “hassle” at all for the customer to replace one chip with another, the terrain can be said to be flat. In such a flat terrain war the only differentiator becomes the price i.e. the lower priced chip gets more business share. When firms are desperate for revenue they are likely to drop prices more than usual to retain business. Customer’s know that and love that! Thus customers’ strategy would be to ensure that the terrain stays as flat as possible. They ensure the same by –

1. Preferring commodity chips over proprietary chips wherever possible
2. Qualifying multiple chips for the same socket (even identical chips from different vendors need to go through qualification process)
3. Splitting share between multiple suppliers

On the other hand for a supplier the key to winning a war would be to identify the unevenness in the terrain. In case of a proprietary chip it is an easy guess –A proprietary chip seated in a socket enjoys a position similar to that of the army positioned at the hill-top. To replace the chip with an alternative there are a whole bunch of barriers involved – most important one being the need to redesign the board.

In the case of commodities there may not be any need for redesign but there are other factors involved that make the terrain uneven. The 2 most important barriers are as follows –

1. Qualification – There are two levels of qualification. First a supplier has to be qualified and then most importantly the competing chip needs to be qualified for the socket. For an incumbent it is definitely an advantage if the competitor is either not qualified as a vendor or their chip is not qualified for the socket.

Qualification is an easier process in comparison to redesign. Thus a qualification advantage is lesser than a proprietary advantage.

2. Alignment with customer’s customer – There are cases when the chip customer is a B2B player manufacturing components for an end OEM. Consider the case of automotive electronics. The chip maker’s customer would be the automotive tier-1 module/component maker (e.g. Bosch) who would be selling their products to the carmakers (e.g. GM, VW). If the tier-1 wants to change any chip in their module they would require approval from the carmaker. Approval may not be readily available because a carmaker may consider a chip level change risky.

While defending a socket it is important to be cognizant of all the above mentioned factors that could help in case of an invasion. More often than not an invasion is a price attack—thus not being aware of the factors may cause a company to either leave money on the table or to lose the socket (if the threat is not acknowledged with seriousness).

While attacking a competitors socket the rules of the game reverse. The invader needs to plan out the attack in phases. Most importantly it needs to understand the key-drivers that could convince the customer to switch. At a base level, the following would be the goals at various phases of the attack:

1. Convince on Value proposition – Applicable when you are attacking a conventional solution or a proprietary chip with your proprietary offering. It is most important to educate the customer on your solution. When education is complete the discussion needs to focus on individual sockets. Find out from the customer what his existing solution is and then pitch how you can help him out.
2. Push for qualification – Once you are done convincing your value proposition you need to focus on getting yourself qualified as a vendor and then getting your chip qualified as an alternative for the target sockets. For proprietary chips this may involve getting an alternate design approved.
3. Final attack – Once the customer knows well about you and also has qualified your products you need to launch the final attack – the price proposal. It is important to ensure that the price you quote achieves both objectives – Keeps the customer interested and also leaves you with room to move in case of multiple negotiation rounds. The incumbent supplier, your competition, is bound to react to any invasion—thus having a room to move is important.

When you are selling a commodity chip the attack is easier. For big volume projects customers do prefer multiple sources to mitigate supply risks and thus getting qualified may not be challenging.

As mentioned, when multiple sources are qualified for the same board the fight tends to become a price war. In such cases the best strategy would be to get qualified in multiple sockets and then negotiate with the customer on the overall business and not socket by socket.

Other War types

There are 2 other war-types that I had mentioned in the prequel –

1. War between the new world conquerors
2. Alien attack

I will discuss both in the sequel to this article (the concluding one).