6 ms·

The article is very illuminating on how the licensing on the technology side. But I am confused. First it said: This approach had advantages for everyone invol

The article is very illuminating on how the licensing on the technology side. But I am confused. First it said: This approach had advantages for everyone involved. It meant licensees and Qualcomm didn't have to scrap over which parts of the phone did or didn't use Qualcomm technology, so they could just go ahead and focus on the more important task of developing and selling these hip new gadgets. And then: Qualcomm's argument was that no matter what went in the phones, they wouldn't work at all without its technology. So, was the current licensing agreements put in place because it was advantageous for everyone or was it because Qualcomm argued that the company's technology enabled mobiles? The reasoning is important because if it was former then wouldn't removing the current licensing structure remove the intended "simplicity". But, if it was the latter then I find it surprising why weren't people fighting an obviously dubious claim from Qualcomm.

crusso 9y ago
It was put in place because Qualcomm owned the most important piece of the package and leveraged their strength to get a percentage of the whole phone's value. Everyone got locked into the uniform Qualcomm licensing agreement and despite the change in the cost structure of modern phones, no one has had the leverage needed to force Qualcomm to change their licensing.
mdasen 9y ago
I think looking at it through a couple lenses might help. For the sake of simplicity, let's say that the average royalty Qualcomm gets is $15/phone and they're charging 5% of the MSRP (manufacturer's suggested retail price). These are made-up royalty rates since, as the article notes, royalty rates are a closely guarded secret. 1) Let's look at it from Apple's perspective. Their current device lineup costs $700, $800, $850, $950, $1,000, or $1,150 (iPhone 8, iPhone 8+, iPhone 8 256GB, iPhone 8+ 256GB, iPhone X, iPhone X 256GB). Qualcomm is taking $35 to $57.50 per phone. Increasing the storage from 64GB to 256GB means Qualcomm gets an additional $7.50 for doing nothing. They didn't make the LTE faster on the devices with more storage. It's the exact same technologies, why should it cost more? In fact, Apple is paying 2.3-3.8x more per phone than the average. So, you can imagine that Apple wants to pay less. They're a manufacturer that only makes high-end phones. 2) Let's look at it from LG's perspective. They sell a variety of phones and let's say they have a $100 phone, a $300 phone, and a $509 phone and they sell equal amounts of each. They have to pay a $25.45 royalty on the $509 phone and only a $5 royalty on the $100 phone, but they're paying on average $15/phone in royalties so it doesn't really matter. Before Apple, most device manufacturers created a variety of devices at a variety of price points so it didn't really matter if it was a percentage or a fixed cost. In fact, being a percentage meant that you didn't have to worry as much going down-market. If you create a cheap device, you could capture some of that market without as much worry about what the royalties would do to your margins. I think the article is a bit fuzzy there. It's hard to argue that a percentage is really better for everyone. It does offer a certain flexibility in a market where people might be feeling things out. For example, let's say I want to make the next IoT LTE-connected thing - you want sensors around your home connected by LTE measuring micro-humidities and you'll need at least 100 of them per house. I'm planning on selling them for $10/sensor. It's going to be the next-gen of home comfort automation! If Qualcomm just has a straight royalty fee of $15, I can't get my business of the ground. New, innovative uses of LTE can't come to the forefront because while a $15 royalty is totally manageable for something like a smartphone with a high selling price, it's unduly burdensome for something like my sensors. You can also say that the value-add of LTE in a smartphone is higher than the value add of LTE in my sensors. Why not just use WiFi for the sensors? Even if LTE is needed, they're going to be transmitting a lot less data, don't really need fast data or low latency (old GPRS would provide enough bandwidth to send the tiny updates). LTE is adding less value to these sensors than it does to a smartphone. I think that's part of the argument from Qualcomm's perspective. They want me to be able to create my micro-humidity sensing IoT with affordable royalty rates while still raising enough money to pay for the technology development. If they had to charge $0.50/device to everyone (5% of my $10 selling price), they wouldn't make back their investment in developing new wireless techniques. If they had to charge me $15/device, I couldn't make a viable business model. So, in that way, it can benefit everyone and make it simple. There's no arguing about how much value-add that Qualcomm is adding to my sensors vs your smartphone. I think most people would agree that the value-add proposition of LTE in a smartphone vs WiFi-only is way above my sensors having LTE or WiFi only. I think one way one could measure the value-add is by how much data is consumed by the device. If the device's user is using more LTE data, they're getting more value from the Qualcomm IP than a device that's always on WiFi. If I agree to turn off LTE/CDMA/UMTS and just use WiFi, should I get a rebate of the Qualcomm royalty? While that's impractical, it makes sense - I'm not getting any value add from that IP. Ultimately, we only have imprecise approximations. As phone prices have skyrocketed in recent years, the fact that the royalty is a percentage has become a bigger issue. For years, the prices of phones would have been going down as technology became cheaper. Smartphones changed that as phones went from something where a good one was $100-200 to where a decent smartphone started at 5 to 10x that. I'm guessing that's part of the big issue now. People are looking at Qualcomm as getting fat off of their smartphone innovations rather than Qualcomm's wireless innovations. A better display or more memory doesn't make Qualcomm's IP better. Of course, to play Qualcomm's side of the argument, a better camera means that fast data is a bigger value-add since you don't want to be waiting forever for uploads; a better display means more pixels and more data to send over the network; a better phone means a device you're going to be using more data with. I tend to take Apple's side in this argument, but I can also see how Apple's side is self-serving. I think FRAND (fair, reasonable, and non-discriminatory) should be FRAND. Everyone should pay the same price and, in fact, hidden royalty prices should be enough evidence alone that a company is violating FRAND. Apple making better software or a better display doesn't mean Qualcomm should get to leach off that. Of course, that's kinda self-serving for Apple: a non-percentage royalty rate would mean royalties would have to rise on cheaper, non-Apple devices. It would also stifle innovation in areas we don't even know LTE might be useful yet. Maybe there's some hot new use for the technology just waiting to be invented, but if it requires a fixed fee, it won't get off the ground. Should a farm sensor that uploads maybe 1MB/day really require the same royalty as an iPhone where I slog through a couple hundred MBs/day? Do we want to say, "sorry, IoT is cancelled due to royalties"? Maybe we just need categories. Phones cost $15/device, sensors cost $0.50/device, etc. based on some notion of value add.
- dman 9y ago
  Wonder if Apple will take the same stance on the cut it takes off apps. Ie if you look at the value add that Apple provides by running the app store it has a fixed cost, whether the app costs $2 or $200.
- kemiller 9y ago
  One angle missing is accounting. Percentage makes it really easy — Tell me revenue earned on licensed devices in aggregate, without even revealing breakdown, and we know the royalty.
- kalefranz 9y ago
  It seems that FRAND in this case is a farse. Would be interested in informed individuals’ opinions on how the regulatory framework is falling over in this case.
- dis-sys 9y ago
  Apple doesn't pay Qualcomm directly, thus the retail price you see in apple stores means nothing here. Companies like Foxconn pay Qualcomm to make those phones, because they are the M (Manufacturer in MSRP) here. Open data such as custom export data indicates most iphones shipped by companies like Foxconn are valued at around half of the retail price you see in apple stores.
  - icebraining 9y ago
    Yeap. In fact, Apple pays Foxconn which pays Qualcomm which pays Apple! https://appleinsider.com/articles/17/09/21/qualcomm-loses-two-key-rulings-in-its-patent-royalty-fight-with-apple https://appleinsider.com/articles/17/09/21/qualcomm-loses-tw...
- tooltalk 9y ago
  1) Apple does not have a license with Qualcomm. 2) Apples' MSPR is irrelevant here since it's Foxconn who pays Qualcomm with their existing licensing agreement from pre-iPhone era. Apple outsourced their manufacturing and "licensing" to cheaper manufacturers in Asia from the getgo. Foxconn's royalty in turn is based on their manufacturing cost, not retail cost of, Apple iPhones. So the prices for the current lineup would be more like $200, $230, $250, $300, $350 and Qualcomm's 5% (without any volume discount or other "rebates" agreed upon) is probably no more than $17 and change per device at max, or less than 1% or 1.5% of Apple's MSRP, which you grossly overestimated to make your lopsided argument. The way I see it, Apple's problem here isn't that Qualcomm's licensing rates are unreasonable or something out of ordinary, but that Apple believe that while their patents are worth ~$30 per device for a handful of frivolous utility and design patents (see Apple vs Samsung 2012), everyone else's, especially those that are essential and committed to SSO, are worth $0.0000000X per device. That is why they challenged pretty much every single wireless patent holders (eg, Nokia, Ericsson, Samsung, Moto, etc) whenever their licensing contract was up for renewal (but lost or settled every single one of them). 3) >why should it cost more? I think most of us with smartphones would agree that Qualcomm's underlying tech is the single most important driver for market demand. We know from the significantly lower sales of Apple iPod, which is a perfect substitute good without Qualcomm's IP's, that Qualcomm's IP's adds significantly more value and drive market demand for Apple's iPhones. 4) >Before Apple, most device manufacturers created a variety of devices at a variety of price points so it didn't really matter... No, it always mattered and regulators in other countries have challenged Qualcomm's licensing practices to lower their rates in the past. This time around we have a domestic company with significant political connections and influence. 5) I think you are compounding a few issues that aren't quite related. IMO, it's fair to charge per device at whatever rate Qualcomm is charge if Qualcomm's IP's indeed add significant value and drive market demand for end-products. For decades, while there were always some chatter that Qualcomm's gain was unfair -- just in the same way that some see Apple's profit as excessive and unfair today -- it helped catapult the mobile communication market from the 90's to what it is today. I expect there would be a different pricing tier, with perhaps different licensing models, for IOT's and other products (ie, cars or airplanes) that aren't communication devices and whose value and market demand are not driven significantly by Qualcomm's IP's. This is common sensical stuff. 6) As I replied to your other misguided comment about FRAND, the problem here isn't that Qualcomm's licensing is discriminatory -- many would probably argue that it's equally too high. Standard Settings Organization or FRAND says nothing about what precise royalty rates or basis ought to be* -- it's to be determined by market participants by trial and error, and through negotiation -- Apple has been the least willing participant in this, having developed or contributed nothing to offer in return, other than what Apple sees as "exorbitant" licensing fees. FRAND isn't certainly about making Apple or Apple fanbois happy and changing their rates as they see it fair, reasonable, fit to improve Apple's profit margin. * note that IEEE made it an official policy to use the smallest salable patent practicing unit that embodies IP's as the royalty basis a few years back, but it is the only SSO to announce such policy.

crusso 9y ago

It was put in place because Qualcomm owned the most important piece of the package and leveraged their strength to get a percentage of the whole phone's value. Everyone got locked into the uniform Qualcomm licensing agreement and despite the change in the cost structure of modern phones, no one has had the leverage needed to force Qualcomm to change their licensing.

mdasen 9y ago

I think looking at it through a couple lenses might help. For the sake of simplicity, let's say that the average royalty Qualcomm gets is $15/phone and they're charging 5% of the MSRP (manufacturer's suggested retail price). These are made-up royalty rates since, as the article notes, royalty rates are a closely guarded secret. 1) Let's look at it from Apple's perspective. Their current device lineup costs $700, $800, $850, $950, $1,000, or $1,150 (iPhone 8, iPhone 8+, iPhone 8 256GB, iPhone 8+ 256GB, iPhone X, iPhone X 256GB). Qualcomm is taking $35 to $57.50 per phone. Increasing the storage from 64GB to 256GB means Qualcomm gets an additional $7.50 for doing nothing. They didn't make the LTE faster on the devices with more storage. It's the exact same technologies, why should it cost more? In fact, Apple is paying 2.3-3.8x more per phone than the average. So, you can imagine that Apple wants to pay less. They're a manufacturer that only makes high-end phones. 2) Let's look at it from LG's perspective. They sell a variety of phones and let's say they have a $100 phone, a $300 phone, and a $509 phone and they sell equal amounts of each. They have to pay a $25.45 royalty on the $509 phone and only a $5 royalty on the $100 phone, but they're paying on average $15/phone in royalties so it doesn't really matter. Before Apple, most device manufacturers created a variety of devices at a variety of price points so it didn't really matter if it was a percentage or a fixed cost. In fact, being a percentage meant that you didn't have to worry as much going down-market. If you create a cheap device, you could capture some of that market without as much worry about what the royalties would do to your margins. I think the article is a bit fuzzy there. It's hard to argue that a percentage is really better for everyone. It does offer a certain flexibility in a market where people might be feeling things out. For example, let's say I want to make the next IoT LTE-connected thing - you want sensors around your home connected by LTE measuring micro-humidities and you'll need at least 100 of them per house. I'm planning on selling them for $10/sensor. It's going to be the next-gen of home comfort automation! If Qualcomm just has a straight royalty fee of $15, I can't get my business of the ground. New, innovative uses of LTE can't come to the forefront because while a $15 royalty is totally manageable for something like a smartphone with a high selling price, it's unduly burdensome for something like my sensors. You can also say that the value-add of LTE in a smartphone is higher than the value add of LTE in my sensors. Why not just use WiFi for the sensors? Even if LTE is needed, they're going to be transmitting a lot less data, don't really need fast data or low latency (old GPRS would provide enough bandwidth to send the tiny updates). LTE is adding less value to these sensors than it does to a smartphone. I think that's part of the argument from Qualcomm's perspective. They want me to be able to create my micro-humidity sensing IoT with affordable royalty rates while still raising enough money to pay for the technology development. If they had to charge $0.50/device to everyone (5% of my $10 selling price), they wouldn't make back their investment in developing new wireless techniques. If they had to charge me $15/device, I couldn't make a viable business model. So, in that way, it can benefit everyone and make it simple. There's no arguing about how much value-add that Qualcomm is adding to my sensors vs your smartphone. I think most people would agree that the value-add proposition of LTE in a smartphone vs WiFi-only is way above my sensors having LTE or WiFi only. I think one way one could measure the value-add is by how much data is consumed by the device. If the device's user is using more LTE data, they're getting more value from the Qualcomm IP than a device that's always on WiFi. If I agree to turn off LTE/CDMA/UMTS and just use WiFi, should I get a rebate of the Qualcomm royalty? While that's impractical, it makes sense - I'm not getting any value add from that IP. Ultimately, we only have imprecise approximations. As phone prices have skyrocketed in recent years, the fact that the royalty is a percentage has become a bigger issue. For years, the prices of phones would have been going down as technology became cheaper. Smartphones changed that as phones went from something where a good one was $100-200 to where a decent smartphone started at 5 to 10x that. I'm guessing that's part of the big issue now. People are looking at Qualcomm as getting fat off of their smartphone innovations rather than Qualcomm's wireless innovations. A better display or more memory doesn't make Qualcomm's IP better. Of course, to play Qualcomm's side of the argument, a better camera means that fast data is a bigger value-add since you don't want to be waiting forever for uploads; a better display means more pixels and more data to send over the network; a better phone means a device you're going to be using more data with. I tend to take Apple's side in this argument, but I can also see how Apple's side is self-serving. I think FRAND (fair, reasonable, and non-discriminatory) should be FRAND. Everyone should pay the same price and, in fact, hidden royalty prices should be enough evidence alone that a company is violating FRAND. Apple making better software or a better display doesn't mean Qualcomm should get to leach off that. Of course, that's kinda self-serving for Apple: a non-percentage royalty rate would mean royalties would have to rise on cheaper, non-Apple devices. It would also stifle innovation in areas we don't even know LTE might be useful yet. Maybe there's some hot new use for the technology just waiting to be invented, but if it requires a fixed fee, it won't get off the ground. Should a farm sensor that uploads maybe 1MB/day really require the same royalty as an iPhone where I slog through a couple hundred MBs/day? Do we want to say, "sorry, IoT is cancelled due to royalties"? Maybe we just need categories. Phones cost $15/device, sensors cost $0.50/device, etc. based on some notion of value add.

dman 9y ago
Wonder if Apple will take the same stance on the cut it takes off apps. Ie if you look at the value add that Apple provides by running the app store it has a fixed cost, whether the app costs $2 or $200.
kemiller 9y ago
One angle missing is accounting. Percentage makes it really easy — Tell me revenue earned on licensed devices in aggregate, without even revealing breakdown, and we know the royalty.
kalefranz 9y ago
It seems that FRAND in this case is a farse. Would be interested in informed individuals’ opinions on how the regulatory framework is falling over in this case.
dis-sys 9y ago
Apple doesn't pay Qualcomm directly, thus the retail price you see in apple stores means nothing here. Companies like Foxconn pay Qualcomm to make those phones, because they are the M (Manufacturer in MSRP) here. Open data such as custom export data indicates most iphones shipped by companies like Foxconn are valued at around half of the retail price you see in apple stores.
- icebraining 9y ago
  Yeap. In fact, Apple pays Foxconn which pays Qualcomm which pays Apple! https://appleinsider.com/articles/17/09/21/qualcomm-loses-two-key-rulings-in-its-patent-royalty-fight-with-apple https://appleinsider.com/articles/17/09/21/qualcomm-loses-tw...
tooltalk 9y ago
1) Apple does not have a license with Qualcomm. 2) Apples' MSPR is irrelevant here since it's Foxconn who pays Qualcomm with their existing licensing agreement from pre-iPhone era. Apple outsourced their manufacturing and "licensing" to cheaper manufacturers in Asia from the getgo. Foxconn's royalty in turn is based on their manufacturing cost, not retail cost of, Apple iPhones. So the prices for the current lineup would be more like $200, $230, $250, $300, $350 and Qualcomm's 5% (without any volume discount or other "rebates" agreed upon) is probably no more than $17 and change per device at max, or less than 1% or 1.5% of Apple's MSRP, which you grossly overestimated to make your lopsided argument. The way I see it, Apple's problem here isn't that Qualcomm's licensing rates are unreasonable or something out of ordinary, but that Apple believe that while their patents are worth ~$30 per device for a handful of frivolous utility and design patents (see Apple vs Samsung 2012), everyone else's, especially those that are essential and committed to SSO, are worth $0.0000000X per device. That is why they challenged pretty much every single wireless patent holders (eg, Nokia, Ericsson, Samsung, Moto, etc) whenever their licensing contract was up for renewal (but lost or settled every single one of them). 3) >why should it cost more? I think most of us with smartphones would agree that Qualcomm's underlying tech is the single most important driver for market demand. We know from the significantly lower sales of Apple iPod, which is a perfect substitute good without Qualcomm's IP's, that Qualcomm's IP's adds significantly more value and drive market demand for Apple's iPhones. 4) >Before Apple, most device manufacturers created a variety of devices at a variety of price points so it didn't really matter... No, it always mattered and regulators in other countries have challenged Qualcomm's licensing practices to lower their rates in the past. This time around we have a domestic company with significant political connections and influence. 5) I think you are compounding a few issues that aren't quite related. IMO, it's fair to charge per device at whatever rate Qualcomm is charge if Qualcomm's IP's indeed add significant value and drive market demand for end-products. For decades, while there were always some chatter that Qualcomm's gain was unfair -- just in the same way that some see Apple's profit as excessive and unfair today -- it helped catapult the mobile communication market from the 90's to what it is today. I expect there would be a different pricing tier, with perhaps different licensing models, for IOT's and other products (ie, cars or airplanes) that aren't communication devices and whose value and market demand are not driven significantly by Qualcomm's IP's. This is common sensical stuff. 6) As I replied to your other misguided comment about FRAND, the problem here isn't that Qualcomm's licensing is discriminatory -- many would probably argue that it's equally too high. Standard Settings Organization or FRAND says nothing about what precise royalty rates or basis ought to be* -- it's to be determined by market participants by trial and error, and through negotiation -- Apple has been the least willing participant in this, having developed or contributed nothing to offer in return, other than what Apple sees as "exorbitant" licensing fees. FRAND isn't certainly about making Apple or Apple fanbois happy and changing their rates as they see it fair, reasonable, fit to improve Apple's profit margin. * note that IEEE made it an official policy to use the smallest salable patent practicing unit that embodies IP's as the royalty basis a few years back, but it is the only SSO to announce such policy.