This post will explore why the best mobile processor maker (Apple) does not have the largest market share.
As the forum knows, SOCs can integrate many functions essential to success in the marketplace. The crucial triad is processor, graphics, and modem.
There are many ARM processor licensees (in the hundreds), fewer graphics core vendors (in the tens), but when it comes to LTE modem vendors, it is down to two, just Intel and Qualcomm. The LTE modem concentration is interesting and I think market shaping.
LTE modems have proliferated to the point that even low-cost smartphones have them now. They are indispensable even at the volume, $200 smartphone price point.
Apple 2017 iPads come with and without LTE functionality. This unbundling is highly interesting, since you can buy an iPad without LTE for $329 and with LTE, $459 (thus, LTE costs $130). This price delta is consistent. And as we know Apple makes their own processors, bakes Imagination Technology graphics into their SOC, but must purchase the LTE chip from either Qualcomm or Intel.
The discrete LTE modem increases the $329 iPad cost by 40%. This is interesting for several reasons. Apple has split the product stack with LTE and non-LTE. This must be painful for Apple since it runs counter to their usual practice (which is to decide what the best customer experience would be, build only that device, and expect Apple customers to understand regardless of the cost). LTE seems to be a special case.
Another indication that LTE is a special case is Apple’s dual sourcing and underspecing of LTE modems. The best LTE modems have (on paper) Gigabit speed (Qualcomm X16). Apple’s latest products spec 600 MB/s though, which opens the door for dual sourcing (since Intel can only manage 600 MB/s). And the iPad modem spec is 150 Mb/s (which is 2014 spec).
Given Apple’s reputation for driving vendors hard to obtain the very best components, the LTE underspecification, and the split product stack, is highly interesting. I will now discuss a market segment that doesn’t exist, that I think should exist, and link it back to these issues.
Qualcomm has the largest smartphone market share, with the best modem, but mediocre processor and graphics technology. Apple has the best processor. In the PC market, having the best processor delivers the largest market share. How is Qualcomm able to prevent their market share from collapsing?
As we’ve seen, the discrete LTE modem adds $130 to a device cost. Qualcomm, unlike Apple, can integrate their modem in the Qualcomm SOC (such as Snapdragon 820, floorplan picture here) which does not increase the SOC cost by $130, because the modem takes up a relatively small part of the floorplan. Integration of the modem creates a huge value proposition for the mid and low end of the smartphone range.
In a $200 high-volume smartphone, the $130 discrete LTE cost is unsustainable, giving Qualcomm leverage. Qualcomm’s offering of a modem that includes processor and graphics, all in one chip, enables this market. There is no other competitive option. (Note: I’m aware that Mediatek and Hisilicon have offerings in the $200 smartphone market. While Hisilicon has a competitive processor, graphics and modem in a single integrated die, on a modern process, they are offering their product internally to Huawei only, indicating that they can compete only without competition and would likely not stand alone. And Mediatek offers such low value processors, graphics, modem and process to obtain the $200 price point that they will essentially never be competitive with Qualcomm. I think this proves my point.)
It makes sense that Qualcomm can achieve dominance by offering the full integrated package (3 out of 3; processor, graphics, and modem) where competitors can only offer 2 out of 3. This indicates the value the LTE modem integration offers Qualcomm. I believe this explains why Qualcomm dominates the smartphone market instead of Apple; it is the optionality of the $130 LTE discrete modem cost leveraged over phones with $100-500 price points.
Qualcomm’s modem is a gold mine. It gives them a moat. The moat is so high I think they could stop or slow modem, processor and graphics development without repercussion. Qualcomm appears to desire to improve their offerings but they are under no obligation.
Qualcomm’s SOC processors offer the minimum cost, rather than maximum performance. They have no incentive to offer maximum performance.
I would argue that in a slightly different world, a world without LTE modem integration optionality, most smartphones, even the lowest cost, would have Apple rather than Qualcomm inside. I think consumers would be better served in this world. Imagine the 9.7 inch 2017 iPad (at $329) scaled down to 5 inches, with an integrated LTE modem adding a nominal cost.
I would argue that it is reasonable to expect Apple, not Qualcomm, to have the dominant smartphone market share, absent LTE-related leverage. In the PC market, which doesn’t have the bundling of LTE function optionality, the best processor (Intel) achieves the highest market share (90%). Apple processors are to Intel processors as Qualcomm is to AMD. In this analogy Apple market share should be like Intel’s.
In conclusion, it seems likely that Qualcomm’s LTE-related leverage restricts Apple’s market share and changes the market. Better processor and graphics options are restricted by the $130 discrete LTE modem cost. Qualcomm doesn’t have to compete on processor performance and consequently just licenses stock ARM cores and mildly tweaks them. Qualcomm is lucky not evil. Intel has a parallel position to Qualcomm, in LTE modems, in most respects, and could disrupt Qualcomm but chooses not to.
The LTE modem seems to be a king-maker in mobile. I could imagine in the future Apple, Intel, or Qualcomm will change the status quo in some way (perhaps through acquisitions) to unlock the value of Apple’s dominant processor, so that more value is created, similar to the PC market.
As the forum knows, SOCs can integrate many functions essential to success in the marketplace. The crucial triad is processor, graphics, and modem.
There are many ARM processor licensees (in the hundreds), fewer graphics core vendors (in the tens), but when it comes to LTE modem vendors, it is down to two, just Intel and Qualcomm. The LTE modem concentration is interesting and I think market shaping.
LTE modems have proliferated to the point that even low-cost smartphones have them now. They are indispensable even at the volume, $200 smartphone price point.
Apple 2017 iPads come with and without LTE functionality. This unbundling is highly interesting, since you can buy an iPad without LTE for $329 and with LTE, $459 (thus, LTE costs $130). This price delta is consistent. And as we know Apple makes their own processors, bakes Imagination Technology graphics into their SOC, but must purchase the LTE chip from either Qualcomm or Intel.
The discrete LTE modem increases the $329 iPad cost by 40%. This is interesting for several reasons. Apple has split the product stack with LTE and non-LTE. This must be painful for Apple since it runs counter to their usual practice (which is to decide what the best customer experience would be, build only that device, and expect Apple customers to understand regardless of the cost). LTE seems to be a special case.
Another indication that LTE is a special case is Apple’s dual sourcing and underspecing of LTE modems. The best LTE modems have (on paper) Gigabit speed (Qualcomm X16). Apple’s latest products spec 600 MB/s though, which opens the door for dual sourcing (since Intel can only manage 600 MB/s). And the iPad modem spec is 150 Mb/s (which is 2014 spec).
Given Apple’s reputation for driving vendors hard to obtain the very best components, the LTE underspecification, and the split product stack, is highly interesting. I will now discuss a market segment that doesn’t exist, that I think should exist, and link it back to these issues.
Qualcomm has the largest smartphone market share, with the best modem, but mediocre processor and graphics technology. Apple has the best processor. In the PC market, having the best processor delivers the largest market share. How is Qualcomm able to prevent their market share from collapsing?
As we’ve seen, the discrete LTE modem adds $130 to a device cost. Qualcomm, unlike Apple, can integrate their modem in the Qualcomm SOC (such as Snapdragon 820, floorplan picture here) which does not increase the SOC cost by $130, because the modem takes up a relatively small part of the floorplan. Integration of the modem creates a huge value proposition for the mid and low end of the smartphone range.
In a $200 high-volume smartphone, the $130 discrete LTE cost is unsustainable, giving Qualcomm leverage. Qualcomm’s offering of a modem that includes processor and graphics, all in one chip, enables this market. There is no other competitive option. (Note: I’m aware that Mediatek and Hisilicon have offerings in the $200 smartphone market. While Hisilicon has a competitive processor, graphics and modem in a single integrated die, on a modern process, they are offering their product internally to Huawei only, indicating that they can compete only without competition and would likely not stand alone. And Mediatek offers such low value processors, graphics, modem and process to obtain the $200 price point that they will essentially never be competitive with Qualcomm. I think this proves my point.)
It makes sense that Qualcomm can achieve dominance by offering the full integrated package (3 out of 3; processor, graphics, and modem) where competitors can only offer 2 out of 3. This indicates the value the LTE modem integration offers Qualcomm. I believe this explains why Qualcomm dominates the smartphone market instead of Apple; it is the optionality of the $130 LTE discrete modem cost leveraged over phones with $100-500 price points.
Qualcomm’s modem is a gold mine. It gives them a moat. The moat is so high I think they could stop or slow modem, processor and graphics development without repercussion. Qualcomm appears to desire to improve their offerings but they are under no obligation.
Qualcomm’s SOC processors offer the minimum cost, rather than maximum performance. They have no incentive to offer maximum performance.
I would argue that in a slightly different world, a world without LTE modem integration optionality, most smartphones, even the lowest cost, would have Apple rather than Qualcomm inside. I think consumers would be better served in this world. Imagine the 9.7 inch 2017 iPad (at $329) scaled down to 5 inches, with an integrated LTE modem adding a nominal cost.
I would argue that it is reasonable to expect Apple, not Qualcomm, to have the dominant smartphone market share, absent LTE-related leverage. In the PC market, which doesn’t have the bundling of LTE function optionality, the best processor (Intel) achieves the highest market share (90%). Apple processors are to Intel processors as Qualcomm is to AMD. In this analogy Apple market share should be like Intel’s.
In conclusion, it seems likely that Qualcomm’s LTE-related leverage restricts Apple’s market share and changes the market. Better processor and graphics options are restricted by the $130 discrete LTE modem cost. Qualcomm doesn’t have to compete on processor performance and consequently just licenses stock ARM cores and mildly tweaks them. Qualcomm is lucky not evil. Intel has a parallel position to Qualcomm, in LTE modems, in most respects, and could disrupt Qualcomm but chooses not to.
The LTE modem seems to be a king-maker in mobile. I could imagine in the future Apple, Intel, or Qualcomm will change the status quo in some way (perhaps through acquisitions) to unlock the value of Apple’s dominant processor, so that more value is created, similar to the PC market.