4 ms·

From my understanding almost every phone with CDMA uses a Qualcomm chip. So why was CDMA adopted if there was only one supplier of the technology? Telco's and

From my understanding almost every phone with CDMA uses a Qualcomm chip. So why was CDMA adopted if there was only one supplier of the technology? Telco's and manufacturers hate to have one supplier of anything, because it means the supplier can haul them over a barrel. Sounds like Apple's complaint is that they overpaid Qualcomm for CDMA components (among other things) after they introduced a CDMA iPhone. So why was CDMA adopted at all if Qualcomm controls the patents and (seemingly) is the sole/dominant supplier of CDMA modems? CDMA doesn't seem to be better than GSM in any technical respect.

uxp 10y ago
The article points out that Apple did switch to an Intel manufactured modem at one point, but they allege Qualcomm retaliated by withholding a rebate payment to Apple. I'm assuming that Qualcomm has a list price for their chips, but negotiates a rebate on that price if a contract is fulfilled, or some other generic business promise is arranged. It also goes on to suggest that the portion of payment Qualcomm received from each phone sold is a percentage of the retail or wholesale price of the phone, pointing out that increasing the storage capacity of a phone from 128GB to 256GB, which I assume does not alter the bill of materials supplied by Qualcomm in any way, results in a larger payment to Qualcomm. Your point that Apple should not have developed a CDMA phone makes sense, but it also means that millions of people would be stuck on a network without an iPhone. Apple got so much shit from the public for having an exclusivity agreement with AT&T here in the states, which in large part drove early jailbreaking progress to open the carrier lock, that it makes good business sense to try and reach those potential customers. I would imagine that Apple reluctantly agreed to a contract for CDMA modems that was less than favorable to Apple in hopes of renegotiating at a later date when Apple became more of a dominant customer, and has since failed in all of their tactics.
openmosix 10y ago
There's a part of me that "enjoys" this. When patents hit small fishes - like X-Plane - (and for no reason), nobody seems to care. When a $630B company pays too much in patent royalties, it's big news. Perhaps Apple could deploy parts of its capital towards changing the system, but for everyone - not just for their fees.
anthonybsd 10y ago
>CDMA doesn't seem to be better than GSM in any technical respect. CDMA is vastly superior to TDMA in spectrum utilization, although OFDMA (LTE) fairs better than both of them. For this reason all of the 3G standards including GSMs own UMTS were CDMA-based.
gok 10y ago
It's a way more complicated story, and it's important to disambiguate between "CDMA" the standard and "CDMA" the channel access method. The core issue was that in the 2G era, "CDMA" (the Qualcomm name for IS-95/IS-2000) was built on CDMA (code division multiple access), while GSM was built on TSMA (time division multiple access). CDMA (the channel access method) is more spectrum efficient than TDMA, and has some other nice properties. When the time came for carriers in many countries to deploy their 2G networks, it seemed like a win to use CDMA (the standard) because they could support more subscribers with the same amount of spectrum. Companies other than Qualcomm did make CDMA chips in this era. In the 3G era, the GSM standard also moved onto CDMA ("UMTS"). But this was a pretty serious migration effort, so carriers that had already deployed IS-95/CDMA typically moved along to the 3G variant of CDMA (IS-2000), which required buying less new equipment. As the 4G networks starting showing up, Qualcomm saw the writing on the wall so they (a) got on the GSM/LTE bandwagon and (b) started investing heavily in their SoCs, which could be bundled with their basebands for smartphones. Since they happened to also be the only remaining baseband manufacturer that was making CDMA support, they also had a tremendous selling point: use our chips in your smartphones and you can still support 2G/3G CDMA networks, which at this point were widely deployed. Plus Qualcomm has so many patents on LTE that you're probably going to end up paying them quite a bit of money even if you don't buy their chips. So today even companies like Samsung which make their own SoCs make variants of their phones with Qualcomm chips to support CDMA carriers. Qualcomm fans/employees will often mention that this really all goes back to the 80s/90s, when Europe forced everyone to use GSM (which just so happened to be hugely reliant on Europe-owned patents). CDMA never really got a fair shake at world domination. Perhaps if the FCC had just picked one cell phone standard (perhaps CDMA) like they did for TV and radio things would have worked out differently. China briefly tried coming up with their own cell standard (TD-SCDMA) for similarly protectionist reasons and it didn't really work out. As it is...well, the world finally ended up settling on LTE.
- kogepathic 10y ago
  > The core issue was that in the 2G era, "CDMA" (the Qualcomm name for IS-95/IS-2000) was built on CDMA (code division multiple access), while GSM was built on TSMA (time division multiple access). CDMA (the channel access method) is more spectrum efficient than TDMA, and has some other nice properties. When the time came for carriers in many countries to deploy their 2G networks, it seemed like a win to use CDMA (the standard) because they could support more subscribers with the same amount of spectrum. Okay, so I have a stupid follow up question. The company I work for does a lot of work in developing countries (sub-saharan Africa) and outside of major cities 2G is the only thing you're gonna find. But to my knowledge, there is no CDMA to be found anywhere in these countries, despite what you've said about spectrum efficiency. So, why is it that CDMA (meaning IS-95/IS-2000) was only ever successful in North America? Or am I being dense? Could it be that carriers in sub-saharan Africa are actually installing 3G/LTE base stations but limiting the connectivity to handsets to 2G because of tower-tower throughput? > use our chips in your smartphones and you can still support 2G/3G CDMA networks, which at this point were widely deployed. So I've heard that a number of US providers (AT&T at least) have already shut off their 2G networks. Does this mean that Apple no longer requires CDMA modems in iPhones because the number of CDMA networks is declining? You mention that Sprint and Verizon have significant CDMA networks. What is their migration strategy for 4G? To my knowledge the only 4G standard in use is from GSM.
pjc50 10y ago
> So why was CDMA adopted if there was only one supplier of the technology? Other way round: Qualcomm lobbied for the adoption of CDMA precisely to create this situaion. GSM is somewhat uncommon in being a standard not dominated by a particular manufacturer. GSM has a patent pool under FRAND terms; see the history section of https://pure.tue.nl/ws/files/2477048/375588270387968.pdf https://pure.tue.nl/ws/files/2477048/375588270387968.pdf on how European IP policy avoided these problems. American public policy is to allow one company to win in the market and then use intellectual property to capture overwhelming market share and share of the value generated. This is called a "free market" for some reason.
mdasen 10y ago
It's a bit complicated, but I'll try to address the issues. First, let's say that you want to make an isosceles triangle on a sheet of paper with the base parallel with the bottom of the page (as part of a triangle standard). Nokia owns a patent on lines being parallel to a side of the page. Ericsson owns a patent on figures with two sides of equal length. Qualcomm owns a patent on angles less than 90-degrees. In order to make your triangle, you need all three patents. You can't just go to Nokia and say, "wonderful, you own one of the patents and I want to make a triangle according to this standard and I'm going to compare your bid against the bids of Qualcomm and Ericsson." Nope, you need a license from all three. Now, they've agreed to license the technology under FRAND (fair, reasonable, and non-discriminatory) terms. But what does that mean? The idea is that the patents of these three companies are all needed by the standard. If I want to implement that standard, I shouldn't get a license from Nokia and Ericsson for their patents and then during negotiations with Qualcomm be told that they want 99% of all my profits. That would mean the standard would go nowhere because people would be holding it hostage. The idea is that none of the patent holders should be able to exercise market power due to its patents being used as part of the standard. But given that some patents are more valuable to a standard than others, how do we know which ones should cost what? We don't. Therein lies the problem. Most of the time, companies like Apple negotiate bilateral agreements with companies like Nokia, Ericsson, and Qualcomm. The only thing keeping those three in check is the specter of court action - court action that might look very unfavorably upon them. But we still don't have a great definition of how exactly FRAND licensing should work. Ultimately, I think this is a new challenge for the world that we're trying to figure out. -- Now, on to CDMA. CDMA is a standard and while Qualcomm is often thought of as the owner, the case is actually that it owns a high share of standard-essential patents on 2G CDMA. They own a smaller share of 3G CDMA patents and an even smaller share of LTE patents (to the point that they aren't the #1 holder of LTE IP). Qualcomm has committed to FRAND licensing on its standard-essential patents for CDMA. As such, CDMA isn't really single source - Qualcomm doesn't even own all the IP for CDMA. But the FTC is alleging that despite Qualcomm's FRAND commitment, it isn't licensing its standard-essential patents to competing baseband developers. The FTC is also alleging that Qualcomm is getting paid a lot more in royalties than OEMs have to pay other holders of equivalent standards-essential patents. -- Let's say that consumers need a BLT sandwich to use mobile networks. I own a patent on bacon, you own a patent on lettuce, and someone else owns a patent on tomatoes. I say that bacon is really the hero of the sandwich and demand 3x more for my patent. Is that reasonable? Maybe. I decide that I won't license bacon to anyone unless they're willing to let me manufacture the sandwich on my bread (bread having no patent). I'm capturing a lot of money for my bread business, but I'd committed not to exercise market power like that when the standard agreed to use my patent. But I am. You want a BLT, you're going to need to buy it from me as I manufacture it using my bread. More, let's say that it's 1901 and I'm the only company that makes non-stale bread. Your customers don't want stale bread - at least not the rich ones where you earn your money. So, I tell you that if you buy bread from other companies for your peasant BLTs, you'll have to pay me extra money for all those peasant BLTs or I'll stop giving you my bread for your rich customers. Maybe the other companies could make non-stale bread with capital investment, but if I raise the cost of their bread enough, you're not going to give them bread orders. Which means my bread will keep getting better while their bread doesn't improve. Ultimately, the FTC is alleging that Qualcomm agreed that CDMA and other standards would be standards and wouldn't be single-sourced, but now Qualcomm is wielding its patents to try and make it single-sourced. -- But this really isn't about CDMA even. The FTC complaint goes into how Qualcomm basically supplies all the premium LTE basebands. Part of this is because of Qualcomm's practice of "no license-no chips". If you want Qualcomm chips (which you're going to need for any premium LTE device), you need to pay Qualcomm a license fee for non-premium baseband chipsets you get from other companies. That differs substantially from the agreements that all the other holders of standards-essential patents are enforcing. Normally, if I create anti-Qualcomm to create baseband chips and license the patents from Qualcomm, the people I sell the baseband chips to don't need a license from Qualcomm (just as you don't pay Qualcomm a license on top of paying Apple for your iPhone). The FTC is noting that Qualcomm's "no license-no chips" policy has meant that OEMs can't challenge Qualcomm in court which means that Qualcomm can charge higher royalties than its IP should command since there isn't the judicial threat. Losing access to Qualcomm's baseband processors during the fight would basically doom a mobile company. -- Qualcomm has been using this as leverage. They make the best basebands and a company like Apple or Samsung or HTC need access to those basebands for their premium, profit-making products. If they can make competitive basebands less price competitive by demanding royalties on those chips from Apple/Samsung/HTC, then their baseband is "Total Patent Cost + Chip Manufacture + Markup" while competitors cost is "Total Patent Cost + Chip Manufacture + Markup + Cost of Qualcomm's Patents Again". So, in theory: Qualcomm has chips people need. If you want a Qualcomm chip, you must agree to pay Qualcomm royalties on any competing chips you buy making the competing chips less economically viable. This means that a company like Apple can't give a big contract to a competitor that gives that competitor the chance to invest in their chips and potentially make them as good as Qualcomm's. Which means that a company like Apple continues to be reliant on Qualcomm. Which means they can put the onerous "no license-no chips" term in their contract that requires royalties on purchases of competing chips. . .which means the competing chips don't get the investment to become as good as Qualcomm chips. . .which means Apple continues to be reliant on Qualcomm chips. -- Hopefully that cleared up some of the issue. Qualcomm found itself with the best basebands and to make sure that competitors couldn't invest and become as good, they made sure that OEMs couldn't buy competing chips under threat of losing access to Qualcomm's chips (unless double royalties were paid). This also meant that Qualcomm could punch up the amount that royalties cost for their patents because no one wanted to lose access to their chips while their chips are the best. But their chips will always be the best if no one can give contracts to a competitor without having to double-pay royalties. It's a vicious cycle. -- As an aside, CDMA definitely had advantages over GSM. 2G CDMA had much better voice capacity. 1xRTT was a lot faster than GPRS. EV-DO was way ahead of EDGE and faster to market than UMTS. It was also a lot easier to deploy than UMTS. UMTS meant clearing 10MHz of spectrum (and carriers back then often only had 20-50MHz so it meant clearing a lot of space on an already congested network). UMTS, as a voice and data protocol, meant you had to have enough reliability to support voice without too many faults or customers would get annoyed so you had to wait longer before you could launch it. EV-DO only required 2.5MHz of spectrum which was a lot easier to clear out. As a data-only layer, you could launch it without worrying as much about it being unreliable since it wouldn't drop calls. CDMA definitely had advantages - heck, there's a reason why UMTS was based on CDMA technology. But CDMA specifically is a bit of a red herring here, in my opinion. Qualcomm found itself in a position where it had standard-essential patents for LTE and the best basebands. If you wanted their basebands, you had to agree that you'd pay double royalties on competing basebands that you purchased. You had to agree to that because you needed the best basebands for your flagship products. Given the inflated prices and lackluster performance of competitors, why not source everything from Qualcomm down the product line? Of course, that's a vicious cycle since no one will get good if no one buys their stuff to given them the money to put into R&D. "Oh, I see you need our best-of-class basebands. . .It would be a shame if there were a supply disruption because you didn't agree to our terms. You don't think it's fair that we ask for double royalties when you buy competing basebands? Well, don't use ours! Oh, right, you can't make that decision. So pay up and don't use our competition, you hear?" That sounds ridiculous, but the FTC complaint literally says "On some occasions, Qualcomm has induced certain OEMs to accept its preferred license terms using both the “stick” of threatened supply disruption and [redacted]". -- Anyway, excellent questions and I hope this clarified some.

uxp 10y ago

The article points out that Apple did switch to an Intel manufactured modem at one point, but they allege Qualcomm retaliated by withholding a rebate payment to Apple. I'm assuming that Qualcomm has a list price for their chips, but negotiates a rebate on that price if a contract is fulfilled, or some other generic business promise is arranged. It also goes on to suggest that the portion of payment Qualcomm received from each phone sold is a percentage of the retail or wholesale price of the phone, pointing out that increasing the storage capacity of a phone from 128GB to 256GB, which I assume does not alter the bill of materials supplied by Qualcomm in any way, results in a larger payment to Qualcomm. Your point that Apple should not have developed a CDMA phone makes sense, but it also means that millions of people would be stuck on a network without an iPhone. Apple got so much shit from the public for having an exclusivity agreement with AT&T here in the states, which in large part drove early jailbreaking progress to open the carrier lock, that it makes good business sense to try and reach those potential customers. I would imagine that Apple reluctantly agreed to a contract for CDMA modems that was less than favorable to Apple in hopes of renegotiating at a later date when Apple became more of a dominant customer, and has since failed in all of their tactics.

openmosix 10y ago

There's a part of me that "enjoys" this. When patents hit small fishes - like X-Plane - (and for no reason), nobody seems to care. When a $630B company pays too much in patent royalties, it's big news. Perhaps Apple could deploy parts of its capital towards changing the system, but for everyone - not just for their fees.

anthonybsd 10y ago

>CDMA doesn't seem to be better than GSM in any technical respect. CDMA is vastly superior to TDMA in spectrum utilization, although OFDMA (LTE) fairs better than both of them. For this reason all of the 3G standards including GSMs own UMTS were CDMA-based.

gok 10y ago

It's a way more complicated story, and it's important to disambiguate between "CDMA" the standard and "CDMA" the channel access method. The core issue was that in the 2G era, "CDMA" (the Qualcomm name for IS-95/IS-2000) was built on CDMA (code division multiple access), while GSM was built on TSMA (time division multiple access). CDMA (the channel access method) is more spectrum efficient than TDMA, and has some other nice properties. When the time came for carriers in many countries to deploy their 2G networks, it seemed like a win to use CDMA (the standard) because they could support more subscribers with the same amount of spectrum. Companies other than Qualcomm did make CDMA chips in this era. In the 3G era, the GSM standard also moved onto CDMA ("UMTS"). But this was a pretty serious migration effort, so carriers that had already deployed IS-95/CDMA typically moved along to the 3G variant of CDMA (IS-2000), which required buying less new equipment. As the 4G networks starting showing up, Qualcomm saw the writing on the wall so they (a) got on the GSM/LTE bandwagon and (b) started investing heavily in their SoCs, which could be bundled with their basebands for smartphones. Since they happened to also be the only remaining baseband manufacturer that was making CDMA support, they also had a tremendous selling point: use our chips in your smartphones and you can still support 2G/3G CDMA networks, which at this point were widely deployed. Plus Qualcomm has so many patents on LTE that you're probably going to end up paying them quite a bit of money even if you don't buy their chips. So today even companies like Samsung which make their own SoCs make variants of their phones with Qualcomm chips to support CDMA carriers. Qualcomm fans/employees will often mention that this really all goes back to the 80s/90s, when Europe forced everyone to use GSM (which just so happened to be hugely reliant on Europe-owned patents). CDMA never really got a fair shake at world domination. Perhaps if the FCC had just picked one cell phone standard (perhaps CDMA) like they did for TV and radio things would have worked out differently. China briefly tried coming up with their own cell standard (TD-SCDMA) for similarly protectionist reasons and it didn't really work out. As it is...well, the world finally ended up settling on LTE.

kogepathic 10y ago
> The core issue was that in the 2G era, "CDMA" (the Qualcomm name for IS-95/IS-2000) was built on CDMA (code division multiple access), while GSM was built on TSMA (time division multiple access). CDMA (the channel access method) is more spectrum efficient than TDMA, and has some other nice properties. When the time came for carriers in many countries to deploy their 2G networks, it seemed like a win to use CDMA (the standard) because they could support more subscribers with the same amount of spectrum. Okay, so I have a stupid follow up question. The company I work for does a lot of work in developing countries (sub-saharan Africa) and outside of major cities 2G is the only thing you're gonna find. But to my knowledge, there is no CDMA to be found anywhere in these countries, despite what you've said about spectrum efficiency. So, why is it that CDMA (meaning IS-95/IS-2000) was only ever successful in North America? Or am I being dense? Could it be that carriers in sub-saharan Africa are actually installing 3G/LTE base stations but limiting the connectivity to handsets to 2G because of tower-tower throughput? > use our chips in your smartphones and you can still support 2G/3G CDMA networks, which at this point were widely deployed. So I've heard that a number of US providers (AT&T at least) have already shut off their 2G networks. Does this mean that Apple no longer requires CDMA modems in iPhones because the number of CDMA networks is declining? You mention that Sprint and Verizon have significant CDMA networks. What is their migration strategy for 4G? To my knowledge the only 4G standard in use is from GSM.

pjc50 10y ago

> So why was CDMA adopted if there was only one supplier of the technology? Other way round: Qualcomm lobbied for the adoption of CDMA precisely to create this situaion. GSM is somewhat uncommon in being a standard not dominated by a particular manufacturer. GSM has a patent pool under FRAND terms; see the history section of https://pure.tue.nl/ws/files/2477048/375588270387968.pdf https://pure.tue.nl/ws/files/2477048/375588270387968.pdf on how European IP policy avoided these problems. American public policy is to allow one company to win in the market and then use intellectual property to capture overwhelming market share and share of the value generated. This is called a "free market" for some reason.

mdasen 10y ago

It's a bit complicated, but I'll try to address the issues. First, let's say that you want to make an isosceles triangle on a sheet of paper with the base parallel with the bottom of the page (as part of a triangle standard). Nokia owns a patent on lines being parallel to a side of the page. Ericsson owns a patent on figures with two sides of equal length. Qualcomm owns a patent on angles less than 90-degrees. In order to make your triangle, you need all three patents. You can't just go to Nokia and say, "wonderful, you own one of the patents and I want to make a triangle according to this standard and I'm going to compare your bid against the bids of Qualcomm and Ericsson." Nope, you need a license from all three. Now, they've agreed to license the technology under FRAND (fair, reasonable, and non-discriminatory) terms. But what does that mean? The idea is that the patents of these three companies are all needed by the standard. If I want to implement that standard, I shouldn't get a license from Nokia and Ericsson for their patents and then during negotiations with Qualcomm be told that they want 99% of all my profits. That would mean the standard would go nowhere because people would be holding it hostage. The idea is that none of the patent holders should be able to exercise market power due to its patents being used as part of the standard. But given that some patents are more valuable to a standard than others, how do we know which ones should cost what? We don't. Therein lies the problem. Most of the time, companies like Apple negotiate bilateral agreements with companies like Nokia, Ericsson, and Qualcomm. The only thing keeping those three in check is the specter of court action - court action that might look very unfavorably upon them. But we still don't have a great definition of how exactly FRAND licensing should work. Ultimately, I think this is a new challenge for the world that we're trying to figure out. -- Now, on to CDMA. CDMA is a standard and while Qualcomm is often thought of as the owner, the case is actually that it owns a high share of standard-essential patents on 2G CDMA. They own a smaller share of 3G CDMA patents and an even smaller share of LTE patents (to the point that they aren't the #1 holder of LTE IP). Qualcomm has committed to FRAND licensing on its standard-essential patents for CDMA. As such, CDMA isn't really single source - Qualcomm doesn't even own all the IP for CDMA. But the FTC is alleging that despite Qualcomm's FRAND commitment, it isn't licensing its standard-essential patents to competing baseband developers. The FTC is also alleging that Qualcomm is getting paid a lot more in royalties than OEMs have to pay other holders of equivalent standards-essential patents. -- Let's say that consumers need a BLT sandwich to use mobile networks. I own a patent on bacon, you own a patent on lettuce, and someone else owns a patent on tomatoes. I say that bacon is really the hero of the sandwich and demand 3x more for my patent. Is that reasonable? Maybe. I decide that I won't license bacon to anyone unless they're willing to let me manufacture the sandwich on my bread (bread having no patent). I'm capturing a lot of money for my bread business, but I'd committed not to exercise market power like that when the standard agreed to use my patent. But I am. You want a BLT, you're going to need to buy it from me as I manufacture it using my bread. More, let's say that it's 1901 and I'm the only company that makes non-stale bread. Your customers don't want stale bread - at least not the rich ones where you earn your money. So, I tell you that if you buy bread from other companies for your peasant BLTs, you'll have to pay me extra money for all those peasant BLTs or I'll stop giving you my bread for your rich customers. Maybe the other companies could make non-stale bread with capital investment, but if I raise the cost of their bread enough, you're not going to give them bread orders. Which means my bread will keep getting better while their bread doesn't improve. Ultimately, the FTC is alleging that Qualcomm agreed that CDMA and other standards would be standards and wouldn't be single-sourced, but now Qualcomm is wielding its patents to try and make it single-sourced. -- But this really isn't about CDMA even. The FTC complaint goes into how Qualcomm basically supplies all the premium LTE basebands. Part of this is because of Qualcomm's practice of "no license-no chips". If you want Qualcomm chips (which you're going to need for any premium LTE device), you need to pay Qualcomm a license fee for non-premium baseband chipsets you get from other companies. That differs substantially from the agreements that all the other holders of standards-essential patents are enforcing. Normally, if I create anti-Qualcomm to create baseband chips and license the patents from Qualcomm, the people I sell the baseband chips to don't need a license from Qualcomm (just as you don't pay Qualcomm a license on top of paying Apple for your iPhone). The FTC is noting that Qualcomm's "no license-no chips" policy has meant that OEMs can't challenge Qualcomm in court which means that Qualcomm can charge higher royalties than its IP should command since there isn't the judicial threat. Losing access to Qualcomm's baseband processors during the fight would basically doom a mobile company. -- Qualcomm has been using this as leverage. They make the best basebands and a company like Apple or Samsung or HTC need access to those basebands for their premium, profit-making products. If they can make competitive basebands less price competitive by demanding royalties on those chips from Apple/Samsung/HTC, then their baseband is "Total Patent Cost + Chip Manufacture + Markup" while competitors cost is "Total Patent Cost + Chip Manufacture + Markup + Cost of Qualcomm's Patents Again". So, in theory: Qualcomm has chips people need. If you want a Qualcomm chip, you must agree to pay Qualcomm royalties on any competing chips you buy making the competing chips less economically viable. This means that a company like Apple can't give a big contract to a competitor that gives that competitor the chance to invest in their chips and potentially make them as good as Qualcomm's. Which means that a company like Apple continues to be reliant on Qualcomm. Which means they can put the onerous "no license-no chips" term in their contract that requires royalties on purchases of competing chips. . .which means the competing chips don't get the investment to become as good as Qualcomm chips. . .which means Apple continues to be reliant on Qualcomm chips. -- Hopefully that cleared up some of the issue. Qualcomm found itself with the best basebands and to make sure that competitors couldn't invest and become as good, they made sure that OEMs couldn't buy competing chips under threat of losing access to Qualcomm's chips (unless double royalties were paid). This also meant that Qualcomm could punch up the amount that royalties cost for their patents because no one wanted to lose access to their chips while their chips are the best. But their chips will always be the best if no one can give contracts to a competitor without having to double-pay royalties. It's a vicious cycle. -- As an aside, CDMA definitely had advantages over GSM. 2G CDMA had much better voice capacity. 1xRTT was a lot faster than GPRS. EV-DO was way ahead of EDGE and faster to market than UMTS. It was also a lot easier to deploy than UMTS. UMTS meant clearing 10MHz of spectrum (and carriers back then often only had 20-50MHz so it meant clearing a lot of space on an already congested network). UMTS, as a voice and data protocol, meant you had to have enough reliability to support voice without too many faults or customers would get annoyed so you had to wait longer before you could launch it. EV-DO only required 2.5MHz of spectrum which was a lot easier to clear out. As a data-only layer, you could launch it without worrying as much about it being unreliable since it wouldn't drop calls. CDMA definitely had advantages - heck, there's a reason why UMTS was based on CDMA technology. But CDMA specifically is a bit of a red herring here, in my opinion. Qualcomm found itself in a position where it had standard-essential patents for LTE and the best basebands. If you wanted their basebands, you had to agree that you'd pay double royalties on competing basebands that you purchased. You had to agree to that because you needed the best basebands for your flagship products. Given the inflated prices and lackluster performance of competitors, why not source everything from Qualcomm down the product line? Of course, that's a vicious cycle since no one will get good if no one buys their stuff to given them the money to put into R&D. "Oh, I see you need our best-of-class basebands. . .It would be a shame if there were a supply disruption because you didn't agree to our terms. You don't think it's fair that we ask for double royalties when you buy competing basebands? Well, don't use ours! Oh, right, you can't make that decision. So pay up and don't use our competition, you hear?" That sounds ridiculous, but the FTC complaint literally says "On some occasions, Qualcomm has induced certain OEMs to accept its preferred license terms using both the “stick” of threatened supply disruption and [redacted]". -- Anyway, excellent questions and I hope this clarified some.