So close yet so far: Pure fingers manufacturing balls-up for leaving firm $20m wide of its target Pure Storage's Q4 fiscal '19 revenues of $422m missed its own guidance by around $20m due to a manufacturing balls-up and customers preferring subscription to license deals. The sales haul represented year-on-year growth of 24 per cent for the three months ended 31 January, but this was way lower than the 34 per cent bounce …
Pure's ES2 service is immaterial. What probably happened was that Pure pushed some of these customers that route so they can recognize some revenue.
I have a hunch the problem is deeper. It is not coincidental that pretty much every vendor missed their numbers in some form or fashion. There is definitely a slowdown in storage.
https://asia.nikkei.com/Business/Companies/TSMC-chip-output-for-Nvidia-and-Huawei-hit-by-defective-chemical
https://seekingalpha.com/news/3434949-firm-downgrades-pure-storage-nvidia-deceleration
So it appears Pure hanged its hat on TSMC. Does anyone really believe that a) All Flashblade shipments are AIRI shipments and b) AIRI contributed 20mil revenue this quarter? I don't.
Lastly, there is prior history here that doesn't help the company.
https://www.storagereview.com/gartner_issues_apology_for_pure_storage_s_wildly_inaccurate_sales_data
I still don't see the TSMC connection. Although AIRI is a mixture of NVidia & Pure i'm sure they can't sell the Nvidia part - they are not a reseller. It's a joint proposal between them so they cannot claim TSMC as a distributor. As the guy said Pure use NAND and CPUs, nothing TSMC manufacture would be in a pure system.
You're trying to crate drama where it doesn't exist. The stock market shows that - they finished UP 1.5% because they missed.....but kept guidance and the street waited to hear the reason. Compare that to Nutanix who missed, lowered guidance, gave a terribly poor reason (IMHO) and dropped 32%.
As for the prior history it's the first time i've seen that. If I had a dollar for every time a broker or analyst mis-calculated expected revenue, earnings or any other form of financial info as an outsider i'd be rich. Although it does show the growth.... this article says they just did $442m in a quarter, Gartner said they thought they were doing $276m a *year* back in the day.
Maybe you research AIRI then...
As to the prior history you once again missed the point. The data calculated by Garther was sent to the company for review and they did not revise it although the real data was considerably lower (276 vs 154).
"That's exactly what happened when Pure revealed their product sales to be $154,836,000 for 2014 where Gartner had posted data showing revenues of $276,329,000 for 2014."
I can see what is in AIRI, it doesn't mean they actually sell it on their books. The same way Netapp don't actually sell the UCS servers in a Flexpod. See below - through resellers, i.e partners. People who actually....resell many vendor technologies.
"AIRI is available now through selected reseller partners, such as ePlus Technology, FusionStorm, GroupWare Technology, PNY, Trace3, World Wide Technology and Xenon"
https://www.theregister.co.uk/2018/03/27/pure_nvidia_ai_airi/
As for this "As to the prior history you once again missed the point. The data calculated by Garther was sent to the company for review and they did not revise it although the real data was considerably lower (276 vs 154)."
I think you get the point but, like the the AIRI comment, you just don't understand the way the finance or the industry works. A private company is *never* going to play higher-or-lower with its financials to anyone.
The one thing Nvidia isn't going to have a shortage of is the DGX.
If what you say is correct then IBM, Netapp and all the other Nvidia partners will be calling profit warnings. If this cost pure $20m (which you don't think possible, and I agree there), how much would it cost Netapp and IBM and other partners?
In short, simple industry knowledge points at it being any kind of other supplier in the chain. Not some random chip manufacturer
All-flash storage ages better than disk based storage. It consumes less power, generates less heat, and is not subject to mechanical "wearing out" like HDDs.
The concerns about flash wearing out due to writes was overstated based on the nature of I/O and data. I/O is more read than write, and as data ages it is updated less.
As a result, all-flash arrays have a much longer useful life than HDD based arrays. And many enterprise customers have stretched the amortization of HDD based arrays out to 4 and 5 years compared to the typical 3 years in the past. Customers expect to keep arrays for 5 years now.
In some ways this is like electric cars vs. gasoline cars. The electric car requires less maintenance. Other than tires, it has fewer parts to wear out (things like regenerative braking mean less wear and tear on brake pads and rotors). No belts and hoses, no radiator, etc. There was the concern the batter would wear out in 5 years, but the battery wear out issue, like the NAND wear out issue, is not as bad as expected.
We are now at a point where most enterprise customers have replaced their older high-performance HDD based systems (i.e., 10K RPM and 15K RPM SAS and FC HDD arrays) with all flash.
The result is there is less churn in the high performance array space. This means the all-flash array sales curve is flattening as this wave of initial sales is completed, and there will be more time between that initial wave and the replacement wave.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
