More on PSTG Cloud Block Store

Investment Analysis Clubs Saul’s Investing Discussions
1

I beat the drum and PSTG a lot and fully realize I could be wrong and the market may not actually be missing a hidden growth story. But the more I learn about the CEO and this company, the more I confidence I have that they’re using this period of time as almost a pause, prepare, and launch into the next stage of growth.

They have a large, happy customer base (86.6 NPS and lots of repeat purchases)built up and are now truly moving to an almost 100% software and subscription model with their Evergreen Storage Service and Hybrid Cloud Storage Solutions.

I know that stocks are often “cheap” for a reason, and that could be the case here, but from everything I’m seeing, there’s no way that during 2019 and beyond (barring any major market melt-downs) this company will be valued like a commodity hardware/storage business or, similar to the slower growing NTAP who no longer shares new customer adds because they only appear able to sell into their existing customer base.

Here’s a good article and 10-minute video (interview and product demo with a PSTG Technical Director at AWS re:invent) from an independent reviewer.

From the video interview: estimated 10% of companies all public, 10% all private, 80% combo
My Opinion: good for PSTG, NTNX, TWLO, etc

Cloud Block Store Info:
Enables customers to store data wherever they run their applications. On-Prem, or in the cloud. Same seamless experience with Pure’s products regardless of where the application is running.

Very easy to replicate data from on-prem to cloud or from cloud to on-prem.

Does not require on-prem. Can be cloud only

5 Likes
2

Yup…link might be helpful.

https://gestaltit.com/tech-talks/pure-storage/stephen/pure-s…

another article from the same site that shares more info about the rest of the cloud products:

https://gestaltit.com/exclusive/ken/pure-storages-entry-into…

3 Likes
3

One thing to take into consideration about enterprise storage in the cloud… network latency is a b*tch.

One of my clients has their ERP hosted in AWS. They have a fairly fat “lifeline” connection between their campus and the AWS farm. They also have a “mirror” of the ERP on-premises for reporting purposes. Things that run against the on-prem system run SIGNIFICANTLY faster than those run against the cloud-based system. The culprit is the unavoidable network latency when utilizing the internet. Lest this be blamed only on the ERP, large file transfers are also painful. They work, but the time is obviously much longer than it should be. The same is true even of transactional processing across the internet – latency makes for significant delays.

This connection link has been investigated by AWS, by level3, and by the client. There are still discussions going on, but it seems that running ‘enterprise’ systems over the internet has a hidden cost in transmission time. Thus, all the extra speed afforded by PSTG’s high-performance flash storage may be nullified by issues beyond their control. Not sure how to factor that into your investment thesis…

Tiptree, Fool One guide and IT geek, no position PSTG or L3, long AMZN

9 Likes
4

This connection link has been investigated by AWS, by level3, and by the client. There are still discussions going on, but it seems that running ‘enterprise’ systems over the internet has a hidden cost in transmission time. Thus, all the extra speed afforded by PSTG’s high-performance flash storage may be nullified by issues beyond their control. Not sure how to factor that into your investment thesis…

Not all storage is created equal and, yes, network latency is a fact of physics and it gets worse with distributed web apps. The news that many enterprise applications have been moved back from the cloud to the enterprise datacenter is no longer news and security is an additional factor in this move. Flash storage won’t – can’t – solve these problems, different domain.

I see fast storage very useful where you have to crunch mountains of data where the speed of access of the CPU or GPU to the data is a significant contributor to throughput. If you have followed the evolution of the technology to access data you’ll realize the number of hoops that developers have jumped through including guessing and pre-fetching into fast caches what the CPU might request next. Incredibly complex and convoluted stuff to try to overcome the latency of spinning disks.

So, yes, Pure has its uses but it will not take over the universe. At this time I see flash storage useful in time critical applications such as machine and deep learning, and inferencing. Things like autonomous vehicles where latency can be a matter of crash or safety.

One has to be careful about using anecdote as data. Anecdote is certainly useful as the initial clue but it needs to be evaluated for significance.

I’m having my own issues with PSTG and that is the slowing down of “reported” [GAAP] revenue growth from 300% five years ago to 22% in 2018. Is it a problem or is it the result of the change of business model from sales to SaaS? Anyone know the answer?

Denny Schlesinger

3 Likes
5

I’m having my own issues with PSTG and that is the slowing down of “reported” [GAAP] revenue growth from 300% five years ago to 22% in 2018.

Denny,

I’m not sure where you are getting the “22% in 2018” number. They are forecasting full year of $1.38B, which would represent 36% yoy from 2017. In calendar year not fiscal. They are in fiscal Q4 2019.

Darth

1 Like
6

Thanks, Tiptree, I have been trying to make this point for months … in two contexts. One is claims for extremely low latency for Pure, which makes no sense for any storage accessed over a network, even a local one. The other is the 2ms latency for 5G which only makes sense relative to a very nearby resource.

2 Likes
7

I’m not sure where you are getting the “22% in 2018” number.

From Pure’s SEC filings (GAAP) and estimating the last quarter of fiscal 2019 from the first nine months.

http://secfilings.nasdaq.com/efxapi/EFX_dll/EDGARpro.dll?Fet…

http://secfilings.nasdaq.com/efxapi/EFX_dll/EDGARpro.dll?Fet…


**Year    Revenue     Change** 
2014     42,733	
2015    174,451    308.23%
2016    440,333    152.41%
2017    727,977     65.32%
2018  1,023,019     40.53%
2019  1,250,144     22.20%

They are forecasting full year of $1.38B, which would represent 36% yoy from 2017.

Forecast vs. extrapolation. Still, at 36% it’s still the lowest ever.

They are in fiscal Q4 2019.

Sorry, my typo.

Denny Schlesinger

1 Like
8

Pure sales growth is slowing down - period. I doubt very much that it materially accelerates again. The reason being is that they have a slower growth dominant product, a faster growth, but much smaller FlashBlade product, and they are trying to dominate and old (but yet now new again) niche of a new architecture, DAS - Direct Attached Storage, but this time with all the benefits and yet networked so as to replace NAS and SAN.

The higher growth FlashBlade is of course mitigated by the lower growth (but still growth) legacy products. The DAS is a niche architecture in the making. It will roll out, it will grow. But it is not going to spread across the world here and everywhere as something new and viral.

Further problem is that almost everything PURE does so does NTAP. NTAP some say NTAP does it even better. Head to head that is not likely I think, but the difference in performance is often not worthwhile to consider as material to a purchase decision. Thus PURE is not necessarily singular in the primary aspects of its markets.

Thus, conclusion, where is the place for the Tornado to form? FlashBlade is not going to grow any faster than it is now with Facebook buying tons of the stuff (secretly). Yet, despite this, sales of FlashBlade are still small in comparison to overall sales.

Good company, good product, but in general nothing they do that NTAP does not do at least similarly well in the mainstream. The rest of what PURE does, its distinct products, are all niche products presently (not small niches, very good and profitable niches with real potential) but still minority markets with no sign of an expansive Tornado forming at present.

Tinker

3 Likes
9 
Year    Revenue     Change 
2014     42,733	
2015    174,451    308.23%
2016    440,333    152.41%
2017    727,977     65.32%
2018  1,023,019     40.53%
**2019  1,250,144     22.20%**

Denny, where is 1,250,144 coming from? They’ve got more than that in the TTM.

Bear

10

Denny, where is 1,250,144 coming from?

937,608 * 4 / 3 = 1,250,144

937,608: http://secfilings.nasdaq.com/efxapi/EFX_dll/EDGARpro.dll?Fet…

They’ve got more than that in the TTM.

TTM: 308,884 + 937,608 = 1,246,492

1,246,492 < 1,250,144

308,884: http://secfilings.nasdaq.com/efxapi/EFX_dll/EDGARpro.dll?Fet…

Denny Schlesinger

11

https://s21.q4cdn.com/687136699/files/doc_financials/2019/q3…

Slide 5 quarterly revenue. Q4 is seasonally their largest quarter by a significant amount.

That explains why the 4/3 method will be way off of guidance when using nine months ended sept data.

1 Like
12

TTM: 308,884 + 937,608 = 1,246,492

1,246,492 < 1,250,144

No Denny, 308,884 was the Q ending July 2018. The Q ending January 2018 was 338,253.

1,275,861 > 1,250,144

But even with correct math, your method isn’t meaningful, because it still ignores the fact that this coming quarter will be greater than the one from last year. As you said, They are forecasting full year of $1.38B, which would represent 36% yoy from 2017. And as you pointed out, that 36% is only down from 40% last year.

Not much of a slow down this year. I wouldn’t sweat it.

Bear

3 Likes
13

As you said, They are forecasting full year of $1.38B, which would represent 36% yoy from 2017. And as you pointed out, that 36% is only down from 40% last year.

I didn’t say that, I repeated what Darthtaco said to point out that even at that level growth is still slowing down.


Year    Revenue     Change 
2014     42,733	
2015    174,451    308.23%
2016    440,333    152.41%
2017    727,977     65.32%
2018  1,023,019     40.53%
2019  1,380,000     34.89%  <-- Darthtaco's number

Not much of a slow down this year. I wouldn’t sweat it.

I’ll make a deal with you: you sweat as little as you want and I’ll sweat as much as I want. Deal?


**Year     Revenue   Change     Rate   |   Year     Revenue   Change     Rate** 
2014      42,733                     |   2014      42,733      
2015     174,451  308.23%            |   2015     174,451  308.23%   
2016     440,333  152.41%  -50.55%   |   2016     440,333  152.41%  -50.55%
2017     727,977   65.32%  -57.14%   |   2017     727,977   65.32%  -57.14%
2018   1,023,019   40.53%  -37.95%   |   2018   1,023,019   40.53%  -37.95%
2019   1,250,144   22.20%  -45.23%   |   2019   1,380,000   34.89%  -13.90%

I’ve added a column to show the rate at which growth is slowing. Even using the company’s estimate of $1.38B the growth rate is down by 14% y/y. This rate is the symptom. Tinker explained quite well the probable causes. I’ve been backing the idea of a disruptive paradigm shift in storage but if NetApp can latch onto it then it’s enabling, not disruptive innovation and PSTG ceases to be an interesting investment. I’m not sweating, panicking, and selling just yet but PSTG is the slowest revenue grower of my high tech stocks, not something to ignore.

Denny Schlesinger

BTW, my times 4/3 estimate is higher than the TTM number. Where did you get the higher TTM number?

3 Likes
14

Where does the upcoming 3D-Xpoint technology come into this? I suppose Pure could make products using it…

15

Where does the upcoming 3D-Xpoint technology come into this? I suppose Pure could make products using it…

I don’t think hardware matters as long as data can be accessed in parallel. Pure is more about architecture and software than about hardware.

Denny Schlesinger

16

Better than 4/3 would be to figure out what percent of the year the 4th quarter was last year and apply it to this year.

1 Like
17

Better than 4/3 would be to figure out what percent of the year the 4th quarter was last year and apply it to this year.

Could be but precision is not the issue in this matter, rough ball park will do, IMO.

Every alternative presented comes to the same conclusion, revenue growth rate is slowing.

Denny Schlesinger

2 Likes