Friday, November 20, 2020

What are Actual NAND Costs Today and in 2030?


In our FMS paper on NAND costs for all vendors and all technologies, we showed how NAND bit costs are being reduced as we go from 64L to 256L. We also showed that equipment companies and Fabs are improving efficiency much faster than we expected. We also showed that QLC will grow to over 50% of bits by 2025 and that this will lead to cost reduction.

As a result, ACTUAL NAND bit costs are on track to drop ~20% per year if we continue the increase in layers at this rate.

Details are at with multiple reports

Some data:


+ Each layer increase brings significant cost reduction
+ Costs are significantly reduced each year for each existing technology (efficiency)
+ QLC is modeled at 22-23% cheaper than TLC

Overall, this leads to Gbyte cost reduction (these costs include package, assembly, test) over the next 10 years:

Another presentation at the same FMS session, re-published as well, also had costs for each company and technology. While the technical analysis of the processing and teardowns is excellent, their cost numbers are obviously incorrect. If those costs were the actual costs, every NAND company would be running negative 10% gross margin. Their margins are bad… but not that bad. Our costs line up with and are calibrated to what actual ASPs are and what actual margins are. All the gory details are available for discussion.

Mark Webb


Wednesday, November 18, 2020

Emerging Memory Revenue (MRAM, Optane, 3 XPoint, ReRAM, FE RAM)

A Picture is Worth a Thousand Words

Emerging Memory Revenue (MRAM, Optane, 3 XPoint, ReRAM, FE RAM)

What is the Market for Emerging Memories in 2030?


We have a lot of requests to predict the Emerging Memory Market in 2025 and 2030. We provide an estimate and the reason why most other forecasts are way off.

At Flash Memory Summit, we presented details on Bits and Revenue for Optane 3D Xpoint in 2025. This presentation is available at our website  along with NAND Costs and other presentations.

But that wasnt good enough! People wanted a summary of ALL emerging technologies and data through 2030. When people make predictions, I like to say "did you predict this year? or last year?  or next year correctly? if you can't predict the past... give up on 2030. 

Emerging Memory as we define it is 3D Xpoint+ MRAM + ReRAM + Fe RAM + others. It is does not include NAND, DRAM, NOR, SRAM, EEPROM, etc. 

This is for stand alone memory only. Not embedded/SOC

Summary (Call us for more details and how you can monitor our progress)

Year        Optane/3D XPoint    MRAM         All New Memories (includes low density)

2020        1.1B                            80M                1.25B

2025        4.3B                            250M               4.75B

2030        9.7B                            610M                10.7B

The emerging Memory Market will grow to $10.7B in 2030. 90% of that is PCM/Optane

Why is MKW Ventures number different from other estimates???

1) Other estimates use CAGR. It turns out memory and storage revenue is not very "compound" ... it actually grows more linear than exponential after first couple years. This is why SSD and NAND sales are half of what people predicted back in 2013. BITS grow CAGR... Revenue does not due to cost and ASP reduction.... and the fact that customer can't print money.

2) Even for existing "new" technologies, they are not growing that fast. MRAM revenue is growing 10% per year right now... during a huge breakthrough period. We predict huge revenue increase for Optane simply because Intel is investing Billions in it and is driving new bus connections for it. Without Intel ... divide the numbers by 5-10 

3) New technologies are expected by others to take off quickly. If ReRAM or FeRAM take off. it will be in 4-5 years and it will not replace other markets. Optane is 5 years after production introduction and Micron sales are less than $10M per year other than Intel. It take a long time ... AFTER we agree the technology works

No technology or product has had 35% revenue CAGR for 10-15 years.... Obviously it won't happen with Emerging Memories either. 

I have details by technology and checkpoint to monitor... and details on why PCM/Optane will grow. Call for more info

Mark Webb

Friday, November 6, 2020

Mark Webb/MKW Ventures at Flash Memory Summit

We have multiple Virtual Presentations at FMS next week. Please check them out and schedule a meeting to discuss questions and answers. We have a zoom meeting Friday 11/13/20 at 11AM Pacific time where we will add commentary and additional numbers

Our Presentations:

Session A-3: Using 3D XPoint to Overcome Storage Roadblocks 2:35 Tuesday

Mark Webb: 3D Xpoint Market Bits and Revenue

We present Rev 1 and 2* Optane/3d Xpoint technology updates and we guesstimate what some future teardowns might show. We show revenue, cost over time and why we think there will be big improvements in both SSD and Persistent memory sales in 2 years. The Market is not going to $8B but it could surpass 3B  per year in the next few years. We can discuss the information used to make those predictions after the presentation. And what are the margins on Optane looking like these days??

Session C-3: How Data Centers Can Profit from New Memory Technologies 2:35 Tuesday

Mark Webb: Annual Emerging Memory Update

We give an update on memories for the datacenters. Our focus is not on every technology. We show that using the Memory Product Lifecycle, technologies are in VERY Different positions. The lifecycle allows you to judge each technology and where it is in terms of success probability. We also make the prediction and call to action "its time to focus and spend money on EMERGED technologies and we should spend less time on the EMERGING technologies which are 5+ years from products"

Session A-4: 3D XPoint in 2025, and How We Got There 4:45 Tuesday

Live Panel Session where we field questions on what the 3D Xpoint market will look like in 2025. We will give our opinions with bit shipments and revenue and what will cause the inflection point in 2023!

Bring questions! 

Session C-9: Flash Technology Advances Lead to New Storage Capabilities 8:35 Thursday

Mark Webb: FLASH Memory Technologies and Costs through 2025

We will show the status of NAND technologies, what each company is doing, and what the roadmap looks like to 256L and beyond. We will show relative costs for each technology over time from each company.  We also show why we are seeing costs drop FASTER than originally predicted. The cost leaders will soon change and this will impact the market. We show average price compared to HDD over time and when it will crossover.... for TLC, QLC and even PLC. Spoiler alert: Nearline HDD isn't going away as fast as some NAND companies predict

There is a ton of back up data, graphs and exact numbers for these presentations. We can set up private meetings after the presentations, just call or email


We will host a public ZOOM call on Friday Nov 13th at 11AM Pacific Time. Bring questions, comments, differing opinions and disagreements. We will provide some more quantitative data on 3D XPoint technology , NAND costs, MRAM technologies. And what about Hynix???

Meeting Name: Mark Webb MKW VENTURES Post FMS Conference

Time       11AM-NOON Pacific Time 11/13/20

Meeting ID

820 7491 8873

Mark Webb