New Post: Librem 5 Update: Shipping Estimates and CPU Supply Chain

I admit my comments were reactive to state of transparency and fluency in purism comms. I did kinda suggest the USA may implode before I get my phone and I got my ill deserved attention which has been fun. I am a reasonable person, a little impatient and frustrated, but reasonable. :slight_smile:

1 Like

Close in line. Not implied “close in delivery time”.

However, note that it says “expect to rise”. Also note Kyle’s comments on that:

As we started getting word about this shortage we were proactive in sourcing and purchasing all the CPUs we can, and continue to do so, while also factoring these increased lead times into future orders.
Based on our efforts thus far there’s a good chance it will not affect your shipping time

That doesn’t sound like the next NXP shipment is due in 26 weeks from now.


Yes, over dramatic.

1 Like

What is the most time consume task in the shipping process?
If the assembling is need a lot of time, then can we talking about the DIY sets?
I think, not just me able to assemble the components. What you think, if save month(s) then would you be willing to assemble yourself? This method save time booth for you and the Purism team .

If you are wanting a refund Do not hold your breath as I requested over 2 weeks ago and twice repeated the request since. If you are waiting for a phone don’t hold it either. No response and my order status still “awaiting shipment”. No voLTE support and now CPU shortage. We may wait until next year for a working phone and it may not work for long after the carriers only allow voLTE. I wish they would keep 1 promise about shipping times. If I get a refund in the next 2 years, then I might revisit buying a phone from Purism in 2023. In the environment we are in now ATT/T-mobile might reject the phone because they can not control your speech on a open and secure phone platform .

Service toot:
The total number of backers is 3000. About 2000…2500 of them did not yet get a phone. Those will know their shipping date soon.
Until Dec 2018, about 1000 more phones have been ordered.
So, assuming they did not yet order more CPUs, a next batch of 1000 could still be done in July.
Although I’m not that pessimistic.

Explaining customers who ordered a DIY kit what they did wrong during their assembly :stuck_out_tongue_winking_eye:

1 Like

I pre-ordered 7/31/2019. I know, should’ve gotten in sooner, but realistically, it sounds like I’m staring at a 6-12 month wait? As I’ve been without a personal phone now for six months (using my corporate spIphone for personal use in the interim), would anyone have any good short-term, budget conscious alternatives while I wait for the L5? Pine might work for me; I’ve seen other options in various threads running /e/, Lineage, Sailfish, etc. Just curious what this group would recommend for a short-term usage with potential backup use down the road for L5.


I work in the silicon chip industry and wanted to offer some advice to Purism when it comes to the sourcing of the silicon chips in a market of shortages. The automotive industry is starting to lay people off right now while simultaneously experiencing a high demand for their products, for one reason. Silicon chip customers are having difficulty sourcing the needed silicon chips that go in to their products. The work from home situation caused by the pandemic has significantly increased the demand in the consumer electronics market and silicon chip manufacturers can’t keep up. All segments that use any given chip are affected.

Where I work, there is a term for the affect described below. The executives use this term routinely. I’ll avoid using that same term here because I doubt that other silicon manufacturers use the same term to describe this effect. But they must all have some term for the same affect because of how greatly the related issues affect company profitability.

It takes a minimum of ninety days or longer (often much longer), from start of a new lot of wafers until the wafers are ready to be assembled in to IC packages and eventually sold to customers. Five months or more is a reasonable time from placement of a new order until that same product could be shipped to a customer. Of course, that is too long for the customer to wait. So the silicon manufacturer has to predict in advance, what they think the future demand might be and build the chips long before the orders are placed. If they build too many, they end up scrapping un-sold product, which lowers profits. If they make too few, a large backlog occurs, causing customers to wait weeks or up to several months (maybe even years) to receive their orders. When the demand greatly exceeds the supply, customers double and triple or quadruple (or more) their order sizes. This allows them to fulfill their own back-orders and plans ahead for their anticipated future needs, based on the shortage (order more than you really need). It makes them feel better psychologically too, given the shortage. If they order too many, they can always cancel future product deliveries before those orders are shipped to them. Amidst sky-rocketing demands of the moment, the silicon manufacturer has to project how many of these orders are real and how many of them are “phantom orders” (that will be canceled later), as world markets fluctuate and will be different in six months from now than they are today. The manufacturer has to see in to the future to make only enough chips to meet the real demand as it will occur six months from now. If they get that wrong, it costs them money. The pipeline through silicon fabs is slow and narrow. The manufacturer can’t just let their fab employees go for six months while while waiting for their stockpiles to diminish to normal levels and expect these highly skilled fab employees to return to work six months later when they may need to ramp-up production again. They can only add fab employees as needed, and let attrition lower the number of fab employees to decrease staffing over time. Anything less shows disloyalty by the company to the employees and causes a myriad of other problems that result. Whether a layoff is big or small, the company pays a significant price too. So in the midst of an extreme backlog, a silicon manufacturer might even slow production, depending on other factors (near prophetic predictions of future world events and conditions). A game of rapidly escalating and collapsing demand from customers vs near-prophetic prediction of future world events by the silicon manufacturer occurs, as the silicon manufacturer only builds and stores the correct number of chips to balance customer needs against company profits as will occur six months in the future. If they build too many chips, profits suffer. If they build too few chips, customers have to wait several weeks or months (maybe even years) to get the product they ordered.

So if I were Purism, and I knew that my ramp-up and future production of Librem 5 shipments was imminent and not dependant on new demand, I would plan accordingly when ordering the silicon chips. Ask the sales person if the pre-payment of the entire order under terms of no possible cancelation will give your order a priority in order of shipments they send out. If the answer is “yes”, this will let the silicon manufacturer know that none of Purism’s orders might be “phantom”, as created to buffer against other conditions in world markets that might occur in the future. If they let you, pay in advance for all anticipated needs, place the order that way to give your order a higher priority. If the answer is “no” and you need 10K chips for the foreseeable future or to cover the combined needs for all back-ordered Librem 5’s, order 100K chips now and pay the minimum amount down. Be prepared to cancel 90% of the order before it is delivered to you. Set the delivery schedule up to deliver 10K chips now and 10K chips each month afterward. But be prepared to act quickly on any cancelations. Basically, you want to ‘out-phantom’ their prediction of phantom orders to assure that you get your product sooner than others in the market will, under a shortage situation. If the shortage is really bad, they may even ship you partial orders each month after you pay in full. This game involving phantom orders has a name. I am sure that each different manufacturer has their own name for it. You order what you want them to think you need in hopes of receiving more product sooner than you would have received otherwise. They try to predict how big your lie might be. Ultimately, they try to distribute as much product as they can, as fairly as they can until their backlog is caught up. If you only order what you need without pre-paying for it, you might end up being last in line after they consider a possible phantom order element of their over all customer demand.


A straw man argument and out of line here.

But if you really believe it, then this is a good reason why the U.S Constitution must be defended and not attacked.

I’m in about the same position in the queue. I think Q2 of 2021 is more likely. The first Evergreen batch was limited to a very small number (probably because there are still many software issues so Purism can’t overwhelm its support staff with a large number of customers, and developer models that don’t yet have FCC/CE certification have to be limited in number). Linear projection based on the first batch is inevitably going to wrong.

Evergreen is ready for early adopters like me, which is why I’m glad that Purism is finally shipping, but the software still needs a lot of work and Purism may still need to change the hardware design if the FCC/EC certification finds problems or they find some issue with a component such as the cameras, smartcard reader, etc.

Your expectations need to be calibrated to the fact that Purism only has 10 developers working on the Librem 5’s software. It is unrealistic to expect a finished product at this point in time, given the amount of technical challenges Purism has to overcome.

If you want to criticize Purism’s poor communication with the community and the marketing of the Librem 5, then I think that you have a point, but criticizing the company for not yet having a phone ready for mass production is not fair in my opinion, given the company’s resources.

You are comparing apples to oranges:

Purism Librem 5 PINE64 PinePhone F(x)tec Pro1
SoC NXP i.MX 8M Quad Allwinner A64 Qualcomm Snapdragon 835
SoC mass production 2018-01 2015-06 2017-Q1
Kernel The Librem 5 has been an early adopter for many new components (i.MX 8M, LM3692 x , MAX17055, LSM9DS1, BQ25890, MLAF057WE51-X, etc.) and Purism has made over 100 commits to mainline Linux to support the new hardware. PINE64 selected older components that already had good mainline Linux support and only had to make a couple commits to get the PinePhone supported in mainline Linux. The Pro1 uses Qualcomm’s kernel and Android drivers through libhybris.
HDMI or DP video out 2020-07 2018-12 (Linux 4.20) 2017-Q1
camera interface still adding to mainline Linux? 2019-05 (Linux 5.1) 2017-Q1
Suspend-to-RAM not yet Crust developed since 2017-10 2017-Q1
How designed Designed from scratch and early adopter problems with a new i.MX 8M SoC and no i.MX 8M phone reference design was available. Required 5 prototypes (pre-Aspen, Aspen, Birch, Chestnut, Dogwood) before mass production. Designed from scratch since no A64 phone reference design was available, but PINE64 has been designing boards with the A64 since 2015. Required 3 prototypes (pre-developer, v1.0 Developer, v1.1 Braveheart) before mass production. Used a Snapdragon phone reference design from Qualcomm or an ODM.
Electronic components 1304 + roughly 100 more in two M.2 cards 704 400-600
Software development Developed the new Phosh mobile environment. Made over 100 commits to mainline Linux. Paid Silicon Labs to alter the RS9116 firmware. Outsources all software development to communities. 70% of Pinephone users report using Phosh developed by Purism. Some porting work of Ubuntu Touch, but very little software development.

Purism outsourced the Librem 5 design to a Chinese firm, but if I were to guesstimate the amount of labor by hardware engineers to design these phones, I would say that the Librem 5 required twice as many hours as the PinePhone and four times more than the Pro1. From what I can gather the Pro1 was mostly designed by 1 guy, which is doable when working off a standard reference design.

PINE64 does no software development and relies heavily on the software development being done by Purism, so your comparison makes no sense.


Omg this forum is all fanboys.

My complaint has nothing to do with software. It’s with purists inability to deliver and their INSISTENCE of hiding their process as if it’s normal. I’ve backed 2 other phones, but designed in-house, amd both are somehow able to do what purism is struggling to do, and fanbois are choosing not to acknowledge. Product or transparency. Stop holding us hostage to our deposit. Especially after they raised money last year.

1 Like

FanMEN. And fanwomen.

1 Like

I’m definitely a fan. :slightly_smiling_face:


If you need a reliable smartphone for daily usage, you probably don’t want the PinePhone, since you are probably going to install Phosh, and going to be waiting for Purism’s software development, just like you are waiting for the Librem 5. The Ubuntu Touch port for the PinePhone still has some problems and the Plasma Mobile requires even more dev work than Phosh.

What I did was buy a Xiaomi Redmi Note 7 for $175 and installed LineageOS. It also works with Ubuntu Touch. The downside is that you need a computer with Windows (not a virtual machine) to unlock the bootloader for Xiaomi phones and the Note 7 doesn’t support many of the LTE bands in the Americas, but there are many other phone models that support LineageOS. If you live in Europe, the Fairphone 3 with /e/ preinstalled is a good option, if you don’t want to mess with the hassle of reading xda-developers forums to figure out how to install LineageOS on your phone.

I think this is an issue of finances at Purism. If Purism has the money on hand to pay for a big manufacturing lot and can hire more people at its Fullfillment Center in Carlsbad, California, then it should be able to produce 10k phones very quickly. Another issue is that Purism may not want to overwhelm its tech support staff by shipping too many phones at once when there are still a lot of software issues to resolve. If Purism gets 10 new customers per day complaining that WiFi doesn’t work, Purism can tell them how to set the datetime to make it work, but Purism can’t do that when a couple hundred new customers are getting their Librem 5’s every day.

1 Like

Thanks for the feedback @amosbatto!

If you have an excuse for a lie it doesn’t make you honest, it just makes you a liar with an excuse.
As much as I understand what you are saying and still want them to succeed with L5, it doesn’t allow me to become their “customer” instead of a “backer”


Not even close.

There is worst in this forum : haters who subscribe only to criticize Purism, again and again with the same broken arguments
Which one, fanboy or hater, gives some fuel to the motivation of Purism’s team ?


Which two other phones?

It took 5 years of software development (2003-10 to 2008-09) before Android shipped in its first device, the HTC Dream. It took Canonical 3.5 years (2011-10 to 2015-02) to develop Ubuntu Touch before it shipped in its first device, the BQ Aquaris E4.5. KDE Plasma Mobile has been in development for 5.5 years (since 2015-07), and it is currently behind Phosh. Apple managed to get iOS to market in 2 years, and Mozilla developed Firefox OS for two years (2011-07 to 2013-07) before the release of the ZTE Open. Given that it has taken Purism 3.3 years to get the Librem 5 to market, it isn’t doing too badly compared to other new mobile environments, especially when considering that Purism has only had between 8 and 13 people working on the Librem 5’s software.

As I said above, you have a point if you want to criticize Purism’s communication and its marketing of the phone, but it is downright wrongheaded to compare the Librem 5 to the Pro1, because they are vastly different in terms of the dev work to bring them to market. I have worked in companies that developed new hardware and new software, so I have some idea of the work involved. I can tell you that developing the F(x)tec Pro1 from a Snapdragon reference design with a kernel and drivers supplied by Qualcomm and porting Ubuntu Touch with libhybris is nothing like starting with a new SoC and no reference design and having to do kernel work to support the hardware and having to create a new mobile environment.


What about Jan 2019? How many until then…asking for a friend.

1 Like

It goes beyond that. There are good reasons to criticize Purism (there are good reasons to criticize anyone, goes with being human). The mindless anti-purism garbage makes it much harder to have a clear-headed discussion about their strengths and weaknesses. It also makes anyone criticizing them look like a jerk. Seriously, you might not agree with the explanation of something, or not be satisfied with the reasoning, but seeing the same tired old doom-and-gloom brought to every thread does no one any favors.