Avalon pursuing 7nm technology

[fanatic26]

I didnt catch the link to this PDF before but there is a ton of interesting information about how they handle customers.

Its a great read.

As a result of the current favorable supply-demand dynamics, we enjoy the flexibility to select our customers and allocate our products in advance of our global expansion strategies. For example, we tend to prioritize our allocation to customers who can show an established mining plan or is able to secure low cost power supply, as we believe these customers have stronger potential for large or recurring purchase demands.

[1714140333]

1714140333

[1714140333]

1714140333

[1714140333]

1714140333

[philipma1957]

My message brings an update a year later to this thread... but as part of Canaan's IPO announce, 7nm is confirmed for the "second half of 2018", see page 94 of http://www.hkexnews.hk/APP/SEHK/2018/2018051401/Documents/SEHK201805150005.pdf

Specifically they taped out 7nm in April 2018 and final products are arriving in the 3rd quarter (p. 118)

They had revenues of 1296 million RMB (203 million USD) in 2017 (page I-32). I agree they have zero intentions to abandon the Bitcoin mining industry. Working on AI accelerator chips is just a natural way to expand their business into other areas.

Well I think mining farms with 1000 841s. Doing 1000 x 14000 gh = 14,000,000gh  or 14,000ph.

Or 1,000,000,000 TV sets with 130gh chip clocked to 100gh = 100,000,000,000 gh or 100,000,000ph.

If you are Avalon what market would you target.

Since Samsung has made decent chips for Halong . I do think for  that Samsung wants a chip in every tv they make.

So if I see this as a direct threat to a mining farm you can be certain that Avalon/canaan sees this as a problem.  I would think they are talking to TV set companies about chipping TV sets

I would think Sony sees Samsung chipping its screens so Sony would want a chip in their TVs as would Lg.

Tvs are everywhere.  Big screens are now cheap and putting in a good 10nm or better yet a good 7nm in larger screens is going to happen.

[sidehack]

Good luck selling a billion TVs. Are there even a billion households in the world with electricity, internet and TV signal available?

I would not pay a dime extra for a fancy new TV just because it can mine me a nickel a year or whatever. The parts cost of that integration would probably outweigh the revenue generated, by a long shot, even without power draw, but you know the TV manufacturer would put the cost onto the customer. Profitable for the chip manufacturer, probably profitable for the TV manufacturer, not profitable for the TV purchaser.

Been saying it for years, and nothing's changed my mind yet, that dropping single mining chips into whatever the heck electronics is a bad idea.

[philipma1957]

Good luck selling a billion TVs. Are there even a billion households in the world with electricity, internet and TV signal available?

I would not pay a dime extra for a fancy new TV just because it can mine me a nickel a year or whatever. The parts cost of that integration would probably outweigh the revenue generated, by a long shot, even without power draw, but you know the TV manufacturer would put the cost onto the customer. Profitable for the chip manufacturer, probably profitable for the TV manufacturer, not profitable for the TV purchaser.

Been saying it for years, and nothing's changed my mind yet, that dropping single mining chips into whatever the heck electronics is a bad idea.

Because you are not Samsung.

Let’s see they made a working chip for Halong the t1.

As I understand it they used their own foundry.

Most of the TVs they make are smart TVs with an internet connection.

I simply ask are the questions above all yes for the answer .

Yes .   So how on earth are they not going to do what I suggest?

Btw it is why asics are not so secure since Sony Samsung or someone will do the above.

I just looked at a Samsung 55 inch smart tv 4K price 759. Uses 65watts

So drop in  7nm from Avalon that can do 150 gh at 9 watts. Price the tv at 950 .  You may not want it but a lot will.  Plus every tv will have it.

Avalon cuts a deal with Lg
Bitmain cuts a deal with some Chinese tv
Innosilicon - Halong cuts a deal with Samsung
GMO cuts a deal with Sony

Most likely canaan has a deal lined up with a tv company.  Which is why they are going public.


I found. Stats that show from 2011 to 2015 more then 1 billion sales of TVs happened

See screen shot

https://i.imgur.com/8u3VpYQ.png



seems to me  that Canaan will be looking to get  piece of this

[vapourminer]

mining with an asic like the avalon 841 is not what I envision for the future of mining.  But I am typing on a mac mini that feeds a 49 inch tv.  If that tv set had about 2 or 3 chips it would earn some coin more or less as an exotic rebate .  Why do I push this idea right now there are more then 1billion tv sets.

This allows for decentralized miners across the world.

Avalon would be foolish to not at least research this. How many chips are in the 841 what does it hash?

Would you call the tv set mining gear ?

this is what I meant. by abandonment or shifting away from mining gear.

answer to the avalon 841 is 104 chips doing 130gh net of 13520 gh

so 1 chip in a 50 inch tv does 130 gh using 13 watts.  this is close to 2 dollars a month.

x 1 billion tvs = 2 billion a month.

No way this does not happen sooner or later.  Samsung is making the Halong chip they must be looking into this idea.

This is why both bitmain and canaan talk about spreading out their line of products.

I have to think it is the wave of the future.

i dunno about mining being incorporated into consumer gear like tvs and such. it would mean it needs network access, something i deny my tvs (and receivers/set top dish box/etc) as there is no need for items like that to need internet in my view.

1st it would lose any "eco friendly" or "green" rating.  (but maybe not if it were switchable, or go on a time schedule..) and who wants a tv with (soon to be obsolete) power sucking heat generating chips in it?  what and where would it mine to and how would the consumer get payouts with this.

im no tree hugger but my house has all led lighting, geothermal hvac/hotwater, we drive high(ish) mpg 4x4 suvs (high for 4x4s anyway), solar panels going in this summer etc. i just prefer to save power and resources when i can.. not because im some super green guy, its because it allows me to use that saved power and resource exactly when and where i want. i prefer well targeted usage of power and resources.  i would rather choose where my watts go, and where my heat sources are located.

if people want decentralized mining there need to be cheap low power miners (well under 100 watts) that run via wifi, are easy to relocate, almost disposable, dead easy to use with preconfigured local pools (say pools located in each major city, choose the one nearest you), then link payouts to a prepaid debit card or something.

[sidehack]

Wait so people will pay $200 for a 150GH miner that cost about $10 to integrate? Sign me up.

Like I said, the chip maker will profit and the TV maker will profit and the TV buyer will lose out.

The closer a miner costs to actual hardware cost, the more viable it's going to be. So miners with one power system, one controls sytem and two hundred chips will always be more cost-effective than miners with one power system, one controls sytem and one chip.

If you want to decentralize with household miners as a service, you need to utilize the waste heat of the miners as a service. Otherwise people are approximately paying for nothing. Stuffing mining chips in arbitrary electronics becomes a waste, but using mining chips for hot air or hot water - things people will be paying to generate heat for anyway - might be viable.

if people want decentralized mining there need to be cheap low power miners (well under 100 watts) that run via wifi, are easy to relocate, almost disposable, dead easy to use with preconfigured local pools (say pools located in each major city, choose the one nearest you), then link payouts to a prepaid debit card or something.

Working on it.

[QuintLeo]

21.com had a similar idea a while ago, but seem to have dropped off the map.

I forget who it was that had the idea of a "smart lightbulb" with a miner chip in it, but haven't seen that come to market.

I believe the whole "consumer device with a mining chip in it" concept is a non-starter, due to power usage (forget Green certification) and requirement for Internet access (lots of folks that don't have WiFi of any form or any interest in it) and the additional cost of such an item.

[ccgllc]

The problem with mining as part of The Internet of Things isn't the chip (or controller or PSU, althought  that was a good point on leveraging those), its on the "paperwork" to register all those devices.  Will the owner of a Mining TV be able to select the pool they want to mine on?   Maybe.  Will the owner of a Mining Lightbulb be able to do so?  Less likely.  Will the manufacturer hard-code a mining pool into the devices that they run, and that by default profits them?  All S9s come that way...

Solo mining with each device (e.g. a giant Mining Lottery) comes with the wallet problem.  Will the manufacturer generate a wallet for every lightbulb sold and notify the owner when they win a block?  Sound expensive for them, since a million light bulbs might have 1 winner a year (unless somehow all governments shut down all mining farms and only allow lightbulb mining - and that won't happen).

So scratch Solo mining and go back to pool mining.  Again, the devices can be preset to mine for the manufacturer, stealing just a little bit of power from millions of customers, and give them the option to change the pool, set a payout address, etc.  How many people in the general population are going to go through that for a few pennies (or even dollars) a year in income?  Of course, the manufacturer has every incentive to make that as hard as possible, and require revealing as much personal information as they can legally ask (which doesn't seem very constrained today, given all the demograhic information requested out of most feedback surveys (think "win a sandwich if you answer a satisifcation survey" scams). 

[sidehack]

Actually no, the biggest problem with stuffing mining into non-internet-connected devices (like lightbulbs) is that the cost of the controller and power for the chip will exceed the cost of the chip, and both together (plus the power cost of operating it) will exceed the value of mined currency. Especially if it's, say, an LED bulb good for 20 years but your mining chip's efficiency would be obsoleted within about 2.

Connectivity becomes less of an issue with "internet of things" devices because they are already networked. The issue then becomes power dissipation. The most efficient mining chip on the market right now would still kill a decent battery inside a few hours at stock settings, so now you've lost portability of anything with mining installed.

Which means you're limited to internet-connected always-on devices anyway. Integrating mining into your TV, DVD player, smart washing machine and voice-controlled toaster won't be free, and you know there'll be a gimmick markup. But let's look at it this way. Say you have 10W chips built into 10 different devices. Now you have 100W of mining going on in your house at all times. However, that also comes with 10 main regulators and IO circuits. A single 100W miner has one regulator and one IO circuit. It basically becomes buying 10 stickminers (each with associated overhead) versus one consolidated miner. The up-front cost efficiency of the overhead required to run a single chip, plus all the markup you'll get from the chip developer, the integrator, the manufacturer, the distributor and the Best Buy across town will add a few orders of magnitude to the actual cost of the "upgrade" you're buying.

After that comes the issue of pools. Which becomes an even larger issue when you consider every device is probably made by a different manufacturer, with his own software, and most chumps foolish enough to buy into the system would have to make ten different customer service calls to figure out how to use 'em properly.

The same argument has been brought up in the Samsung is Making Chips thread and CK basically had the same thing to say about it.

[philipma1957]

Fuck miners my Samsung tv can run the entire block chain.

Not quite true.
But 1 billion TV sets were sold from 2011 to 2015.

Take the usb stick sidehack makes the 1 chip one does not need a fan.

So heat is zero issue. Slap a passive heatsink inside a 50 inch tv and boom heat is solved .

Smart TVs have internet hook up solves that issue.  A rasp pi is more then enough to control the chip.

Smart TVs have built in CPUs .

So internet cost is already in a smart tv
Controller is already in a smart tv
I only need a chip added.
Avalon chip right now can do 130 gh at 13 watts.

If the 7nm does 150 gh at 9 watts my idea is viable.

Cost for a tv company with a deal for Avalon to integrate a chip is going to be under 100 usd.
Maybe under 50. So cost passes to customer. Maybe 75.

Yeah I know 150gh needs a long time to make 75 bucks. but this is going to happen .

Or not I have been wrong many times.

[sidehack]

Yeah I know 150gh needs a long time to make 75 bucks. but this is going to happen .

OR - or - you could make a $200 miner with TEN chips in it, and get ten times the hashrate for three times the money and it'd be a lot more likely to pay itself off in less than the entire service lifetime of your television.

Also how many of those 1 Billion TVs have internet connection? How many were cut-rate bargain models from Wal-Mart that can't afford decent components for essential innards, let alone the expense of add-on gimmicks? And how many people could be convinced to pay up the butt for add-on gimmicks with approximately zero value added, and zero fun interactive features to use as the excuse?

I won't say that nobody will do it, because someone will probably do it. But I will keep saying, like I've been saying for years, that the economics make zero sense. I may not know a lot about smart TVs and the whims of the general populace (who are, statistically, idiots) but you gotta admit I know a bit about the economics of small-scale miners. And it doesn't make sense.

[philipma1957]

Yeah I know 150gh needs a long time to make 75 bucks. but this is going to happen .

OR - or - you could make a $200 miner with TEN chips in it, and get ten times the hashrate for three times the money and it'd be a lot more likely to pay itself off in less than the entire service lifetime of your television.

Also how many of those 1 Billion TVs have internet connection? How many were cut-rate bargain models from Wal-Mart that can't afford decent components for essential innards, let alone the expense of add-on gimmicks? And how many people could be convinced to pay up the butt for add-on gimmicks with approximately zero value added, and zero fun interactive features to use as the excuse?

I won't say that nobody will do it, because someone will probably do it. But I will keep saying, like I've been saying for years, that the economics make zero sense. I may not know a lot about smart TVs and the whims of the general populace (who are, statistically, idiots) but you gotta admit I know a bit about the economics of small-scale miners. And it doesn't make sense.

I never said it made sense for the buyer.

I said it makes sense for samsung

[sidehack]

It stops making sense when the buyer refuses to buy it.

Also, why the hell would you root for a megacorp happily ripping off a billion people?

[philipma1957]

It stops making sense when the buyer refuses to buy it.

Also, why the hell would you root for a megacorp happily ripping off a billion people?

Why do you think I am rooting for this?

Avalon says  they are shifting and expanding from traditional sha 256 mining in their prospectus  they want a bigger base for asic chips.

I figure since 200,000,000 million tv sets were sold for the last 7 years  they may want to go in that direction.

It is their prospectus that said they want more then traditional sha 256 miners not me.

I spent a long time reading it and I thought they may try chips in a tv.
I also know samsung has now made a fairly decent chip and might want to put it in their tvs.
I am looking at this from the viewpoint of the companies not from the viewpoint of  a miner.

  For me as a miner I am stuck until buysolar expands on his solar fields. No new gear will really mean a thing until I get more power at a good price.  But Avalon said they are looking to expanding their field of gear.
I could think they will make a new chip for say LTC or SIA but I don't think so.  So what is something in every country  hmmm tvs' are all over the place. So I thought this could happen. 

 While I did  see the GPU reboot back in 2016 and made some money I did not see the gpu collapse this year. Now that I have expanded my view of mining a bit and Avalon has stated they want to  expand beyond asics for mining btc I thought they may want to do this.
Just a guess on my part. I don't want to underestimate them or other asic companies as they did more then I thought was possible. With asics in the last 10 months.

[sidehack]

I'm still operating on the opinion that a small-scale miner, somewhere between single-chip and 1.4KW 85dB industrial gear, is the optimal solution for household-distributed mining. Something whose cost/hash ratio is not painfully unreasonable, that doesn't need to be air-conditioned to stay cool, and that runs quietly and unobtrusively. The only better household miner would be built into something that is supposed to generate heat anyway, but is built such that the hashcard heat elements are easily interchangeable.

Seems like adding overly expensive tiny space heaters to other manufacturers' household devices would be a heck of a lot more of a pain than just building 50-500W quiet miners.

But hey, if they're not going to that means more business for me.

[philipma1957]

I'm still operating on the opinion that a small-scale miner, somewhere between single-chip and 1.4KW 85dB industrial gear, is the optimal solution for household-distributed mining. Something whose cost/hash ratio is not painfully unreasonable, that doesn't need to be air-conditioned to stay cool, and that runs quietly and unobtrusively. The only better household miner would be built into something that is supposed to generate heat anyway, but is built such that the hashcard heat elements are easily interchangeable.

Seems like adding overly expensive tiny space heaters to other manufacturers' household devices would be a heck of a lot more of a pain than just building 50-500W quiet miners.

But hey, if they're not going to that means more business for me.

look  the ideal thing for me is   a quiet 600 watt miner.  I have space for 100 avalon 841's and I could keep them cool  but I don't have that much cheap power..

On may 31 my rates go up until Oct 1.  16.7 cents and zero benefit for the heat.

Since all the solar power is used. 20kwatts anymore is on the buy expensive power  method. No gear works well at 16.7 + 1.3 to cool or 18 cents a k-watt. I will be turning off all my L3+ on May 31.

      I will be using my avalon 841 and halong t1 in the solar array.  I will mine with a few gpus.
  I will hope to get  buysolar's new idea off the ground blended with the avalon 841's or with 2 L3+  or about 1300 watts.

[sidehack]

I'm still working toward S1 refit boards. Should have a much-improved pod available in a few months, and hopefully S1 refit boards (and possibly whole miners built around them) with a new cutting-edge chip in the first part of next year if the guys I'm working with can pull off what they're trying to do.

Maybe it'll be something you like.

[Thetaj]

I'm still operating on the opinion that a small-scale miner, somewhere between single-chip and 1.4KW 85dB industrial gear, is the optimal solution for household-distributed mining. Something whose cost/hash ratio is not painfully unreasonable, that doesn't need to be air-conditioned to stay cool, and that runs quietly and unobtrusively. The only better household miner would be built into something that is supposed to generate heat anyway, but is built such that the hashcard heat elements are easily interchangeable.

Seems like adding overly expensive tiny space heaters to other manufacturers' household devices would be a heck of a lot more of a pain than just building 50-500W quiet miners.

But hey, if they're not going to that means more business for me.

Never gonna happen bro, and I think deep down you guys all know the reason already so no point for me beating a dead horse.

If something can make money on a small scale, it can make money on a very large scale.

The other option is to make the chips so commoditised and widely available that it no longer makes sense to invest in scale.

Like say: If every single water heater/boiler mines Bitcoin........ Constituting something like 15% of the Network.

Its still very far out though, and thats why I think we're all sitting pretty at the moment

[HagssFIN]

I'm still working toward S1 refit boards. Should have a much-improved pod available in a few months, and hopefully S1 refit boards (and possibly whole miners built around them) with a new cutting-edge chip in the first part of next year if the guys I'm working with can pull off what they're trying to do.

Maybe it'll be something you like.

Oooohh, nice. 
Is it powered by Avalon, Bitfury or something else..?

[philipma1957]

I'm still operating on the opinion that a small-scale miner, somewhere between single-chip and 1.4KW 85dB industrial gear, is the optimal solution for household-distributed mining. Something whose cost/hash ratio is not painfully unreasonable, that doesn't need to be air-conditioned to stay cool, and that runs quietly and unobtrusively. The only better household miner would be built into something that is supposed to generate heat anyway, but is built such that the hashcard heat elements are easily interchangeable.

Seems like adding overly expensive tiny space heaters to other manufacturers' household devices would be a heck of a lot more of a pain than just building 50-500W quiet miners.

But hey, if they're not going to that means more business for me.

Never gonna happen bro, and I think deep down you guys all know the reason already so no point for me beating a dead horse.

If something can make money on a small scale, it can make money on a very large scale.

The other option is to make the chips so commoditised and widely available that it no longer makes sense to invest in scale.

Like say: If every single water heater/boiler mines Bitcoin........ Constituting something like 15% of the Network.

Its still very far out though, and thats why I think we're all sitting pretty at the moment

I was going that direction with TV sets. Since 1.4 billion have been sold since 2011.

Space heaters boilers all lack internet.  They do use power a lot. The delongi space heat is 300 600 1500 watts and is free of sound it has oil based radiator but it is fully dumb.

The tv idea will use an extra 5 to 10 watts but many parts are in a smart tv.

Of course the tv will be 1 or 2 chips.  So maybe neither idea works.