- Intel will not share its hard won advance to 14nm manufacturing tech with Qualcomm.
- In the most optimistic case scenario, Intel will convince the two largest OEMs to use Intel chips despite a difference in instruction set for Apple.
- If that fails, Intel will move to market equilibrium point of production and extinguish profit at the point where marginal revenue is equal to marginal cost.
- Intel will win mobile customers in 2015.
- Assuming massive adoption of Intel SoCs, Intel will break even and marginalize Qualcomm. In the best case scenario, a 50/50 duopoly emerges resulting in Nash Equilibrium.
Intel (NASDAQ:INTC) faces an uphill challenge with mobile when compared to Qualcomm (NASDAQ:QCOM), and while the question of a roadmap has been sufficiently addressed on Seeking Alpha. I think the question of whether or not Intel will become a legitimate third-party fabrication and compete further up-stream with companies like Taiwan Semiconductor Manufacturing (NYSE:TSM), Samsung (OTC:SSNLF) and Global Foundries hasn't been addressed in nearly enough detail.
If I were to adhere to the laws of microeconomics, Intel will focus on competing directly in mobile, and will attempt to use its economic moat to win market share.
The current trajectory of the foundry model
The economics of the foundry model are rapidly changing according to McKinsey & Company:
A McKinsey analysis shows that moving from 32nm to22nm nodes on 300 mm wafer causes typical fabrication costs to grow by roughly 40 percent. It also boosts the costs associated with process development by about 45 percent and with chip design by up to 50 percent. These dramatic increases will lead to process-development costs that exceed $1 billion for nodes below 200mm. In addition, the state-of-the-art fabs needed to produce them will likely cost $10 billion or more. As a result, the number of companies capable of financing next-generation nodes and fabs will likely dwindle.
In an environment of improving scaling density, Intel will most likely try to hoard this advantage in order to secure more mobile OEMS rather than share this scaling density with its number one competitor in mobile, Qualcomm.
Intel assumes that it will have a 45% scaling density improvement, even if Taiwan Semiconductor Manufacturing were to operate at 10nm nodes. I couldn't wrap the logic around why Intel's 10nm nodes would be better than TSM's. But in a recent article from Seeking Alpha, TSM's counter argument was way more qualitative whereas Intel created a function that models the processing tech improvements at different nodes. With outward looking projections, nothing is completely certain, and in the past we've seen semiconductor companies make very huge claims about what their next product generation would do (remember the Intel and AMD saga from 2006-2014).
However, there's one thing investors, industry on-lookers and insiders can basically agree upon. Intel will arrive at 14nm first. What no one really agrees on is what Intel will actually do when it reaches 14nm from a business standpoint.
Intel needs significant share
As many of you are well aware, Intel develops the silicon and also does the manufacturing of chips in-house. Some have speculated that Intel may be willing to allow Qualcomm chips to take advantage of Intel's latest fabrication facilities, but I'm becoming increasingly doubtful of this, as the cost of developing the latest generation technology will require significant market share to break even.
According to Forbes, "Qualcomm has 50% market share in the smartphone application processor market. Qualcomm has 86% market share in LTE cellphone modems." In other words, Qualcomm has a near monopoly on modem and operates an oligopoly in the application processor market. Qualcomm's product roadmap seems fairly compelling, despite a slower shift to better manufacturing tech, and it seems that many are willing to stick with Qualcomm through 2014. The upcoming Samsung Galaxy S5 will come with the Snapdragon 801, so if anything, Intel has to generate significant business wins in mobile, otherwise all the money put toward foundry and mobile would have been meaningless. Ashraf Eassa mentions that the mobile division loses Intel $3 billion or more per year.
Source: Economics online
Using, Ashraf's earlier mentioned figure, for Intel to break even on mobile, a combination of application processors and modems would need to total to $3 billion in revenue for Intel to reach the point of average total cost along an oligopoly supply and demand curve (maybe more to cover marginal production cost). If you look at the above figure closely, oligopolies tend to generate revenue well above the point of marginal cost of production, and tend not to over-produce in favor of maximizing the elasticity of demand at the specific quantity of supply. In this case, Qualcomm already operates along the Q and P curve.
How this economic model plays out in the real world is difficult to fully explain. From the looks of it, the number of competitors in mobile silicon may be limited to two. This means that Intel needs to convince pre-existing companies that are already plenty happy with their own in-house developments (Apple (NASDAQ:AAPL) and Samsung) that an Intel application processor can be both cost-effective and give an added edge. Assuming Intel is able to convince Samsung and Apple which have in the avenue of 50% market share collectively, Intel can operate well above the marginal cost curve and earn reasonable margins on mobile.
However, if Intel is unable to convince Apple and Samsung to go with an Intel based SoC (system on chip), Silicon Valley we have a problem. In this worst case scenario, Intel can no longer absorb the safe and predictable flow of income that Qualcomm currently generates. Instead, Intel will have to move to Q1, and P1, according to the economic model for oligopolies. This indicates that Intel will produce as much 14nm wafers as it can until wholesale pricing can no longer meet marginal cost. In other words, Intel will price itself so that the amount of profit earned per unit will continue to decline along the curve, until there's no more profit. Another way of thinking of this is that as each 14nm SoC is produced, the amount of profit earned per unit will decline until there's no more profit.
How this works, in the case of Intel, is that it will price its high-end solutions lower than Qualcomm, mid-end solutions lower, and so forth. Eventually, Intel will saturate some of the high-, mid-, low-end. At the low-end marginal production halts as lower pricing will result in no more economic profit. In other words, Intel will reach as low as possible until wafer cost reaches parity with the low, low, low-end pricing of SoCs. Think in terms of single digit dollar amounts.
As a result of extremely low pricing in the low-end, Intel will operate closer to market equilibrium and in turn increase the aggregate demand for mobile modems and application processors. In this case scenario, the consumer wins, because those who were priced-out when Qualcomm was the primary distributor of SoCs are now able to afford entry-level devices. This is of course assuming OEMs pass the cost efficiencies onto consumers.
Intel's modem and processor solutions may become superior to Qualcomm by 2015. In this optimistic scenario, Intel has to move aggressively, by lowering pricing to market equilibrium, and in so doing out competes Qualcomm for the remaining 50% of the application processor market.
However, in a more optimistic case scenario, Intel will secure a deal with Apple and Samsung (highly unlikely, but sometimes miracles do happen), and supply the modem and application processors for all mobile and tablet devices. This will allow Intel to return to the point of profit maximization, which would allow Intel to operate well above the marginal cost curve. Assuming Nash Equilibrium, Qualcomm and Intel will each supply 50% of the mobile market, and will price products at the point of profit maximization rather than at the point of equilibrium.
Shareholders will have to pay close attention as to whether or not Intel's mobile market share improves in 2015. This will be the primary metric by which investors will grade Intel's success on. A rapid improvement in market share would indicate that Intel has better pricing and a superior product. If demand for Intel chips is soft, it means Intel's pricing is way too high despite having a marginally better product.