On Hardware Stocks

Investment Analysis Clubs Saul’s Investing Discussions
1

On Hardware Stocks

I placed a short post on this on a Nvidia thread, but it thought I should give it a thread of its own for those who weren’t following Nvidia.

Nvidia’s CEO, who is a very bright, innovative and charismatic CEO, explained Nvidia’s current problems as a problem of pricing and an overstock of inventories.

I’m sure that he knows what he’s talking about, but overstock, inventories, pricing, order postponement, and all the rest, are a whole vocabulary of potential problems that you don’t ever think about, you don’t even ever hear, if you are not in a hardware stock.

Hardware companies have to wait for orders! It’s as simple as that. For example your best customer says, “Sorry, we may have inadvertently misled you, but it looks like we over-ordered (or business was a little slow this quarter), but we have to work off some inventory. We won’t be ordering this quarter, but we’ll look at it again next quarter.” Or “Sorry, but this competitor of yours has this product which is almost as good, but 40% off your price, so we ordered from them.” Or “I don’t think we need to order the upgraded new chips, the ones we have already are good enough.” Or “Our business has slowed down because of the economy (or a change in technology trends), so we only need half as many this quarter.” You get the point. The SaaS customer, by contrast, just pays his monthly lease because that software is essential for running his business. And he may even order another bell or whistle because he sees it will make his business run more efficiently and save him much more than the small monthly fee.

Then there’s the question of increasing your revenue. To put it simply: if you sell $100 million of refrigerators this year, to increase revenue by 50% next year, you have to sell the same number of refrigerators, and then half again more. And to increase 50% the year after, you have to sell as many as you did the first year plus 125% more. It’s clearly impossible to do that even for 4 or 5 years in hardware, unless you invent something that changes the planet, like the iPhone.

Our SaaS companies don’t have to go out and sell that first $100 million again next year. It just comes in automatically, on a lease, and maybe even there will be more bells and whistles attached next year. No customer company is going to “not order” the software that is running their company for a quarter or a year.

I, personally, decided “No more hardware companies!” after 2015/2016 when I had had major positions in Skyworks and Infinera, two tech stocks that were much beloved by members of the board, including myself. They had great stories, and great management. We thought they were taking over their worlds. Their CEO’s told us about all the orders they were expecting, and in retrospect they weren’t lying, they really thought they would get them. But then “Hardware companies have to wait for orders” happened! … Infinera, which was as high as $23 at one point while I had it is roughly $5 now (I just looked it up), and Skyworks (when Apple sneezed, it got pneumonia) which had hit $109 in 2015, and had hit $114 in 2017, is $73 right now. It’s way below where it was in 2015. Some of you may still be holding Skyworks or Infinera, hoping. Most of our stocks are up at least 500%-1000% (in as much as you could value them pre-IPO) since 2015, by comparison.

I hope this helps,

Saul

74 Likes
2

No customer company is going to “not order” the software that is running their company for a quarter or a year.

Except when they go another direction or build their own, as in the case of Uber and Twilio.

Long TWLO,

11 Likes
3

Saul,
I get your focus on SAAS. But I am somewhat worried about Zscaler. The stock seems to be valued for 60%+ growth. Anything else I feel will be bad for the stock. I am wondering if I should sell but then I don’t want to be a trader.

2 Likes
4

I now hear a number of people on the Nvidia thread questioning: How did Jensen Huang get it so wrong? Was he inattentive? Why didn’t he see it coming? Was he lying at the last call, or just wrong?

Sound familiar? That’s just the way we felt back here:

2015/2016 when I had had major positions in Skyworks and Infinera, two tech stocks that were much beloved by members of the board, including myself. They had great stories, and great management. We thought they were taking over their worlds. Their CEO’s told us about all the orders they were expecting, and in retrospect they weren’t lying, they really thought they would get them. But then “Hardware companies have to wait for orders” happened!

Saul

8 Likes
5

But then “Hardware companies have to wait for orders” happened!
I’ve been trying for some time to wrap my head around the disdain for HW companies but I’m starting to really understand it. And the key is not that SW is better, it’s that SaaS is better. And by extension, note that IaaS and PaaS would be in the same favorable camp as SaaS. The key here is the “aaS” part. Tell me if I’m wrong.

This reminds me when Adobe switched to the Creative Cloud offerings. A lot of professional photographers I knew were pissed they had to rent the SW rather than buy it. Why? Simple – many were running old versions of Photoshop because it did what they needed it to. So while the CC will save you money long-term if you like to keep stuff up to date, it costs you more if you tend to be happy using old SW.

So why does Saul keep saying “they have to keep paying that rent check” to the SaaS companies to keep the business running? Simple – they don’t own the software, they have only rented it for a month at a time. They don’t have the luxury to use the old version for a few more months before forking over more money. With traditional software sales, and traditional hardware sales, you can do that. You can delay that next purchase if you need to. Not so in an SaaS world. Skip your payment and your license expires and the stuff won’t run.

This is also why I said that Infrastructure as a Service and Platform as a Service fall into the same league as SaaS. Skip a payment and you no longer have HW to run that SW on.

Now, why do customers put up with this? Mostly it seems a CapEx thing. They spend money small chunks at a time, rather than big chunks up front. There can also be flexibility in scaling - they can rent extra resources for short periods of time, and revert back to normal loads more easily.

It’s starting to sink in.

Bill Jurasz

12 Likes
6

Now, why do customers put up with this? Mostly it seems a CapEx thing.

Bill, it’s also that they get their updates in the Cloud… no fuss, no having to install it in 1000 workstations, no IT salaries for installing the updates, and in many cases the new software update is free, is installed instantly, and you could get two updates a week (or a day) instead of one every 6 months, while struggling on with all the security or other flaws of the old version. Who wouldn’t choose the SaaS?

Saul

14 Likes
7

Who wouldn’t choose the SaaS?

Possibly someone who wanted to heavily customize the software to suit their business.

3 Likes
8

But I am somewhat worried about Zscaler. The stock seems to be valued for 60%+ growth. Anything else I feel will be bad for the stock. I am wondering if I should sell but then I don’t want to be a trader.

TexMex,

I see you started posting in October so you may not have followed the boards in July-August when we had a similar discussion. Here are the links I recommend:

Diablito’s Financial Summary
http://discussion.fool.com/zscaler-a-look-at-the-numbers-3312932…

My answer to GauchoChris who expressed similar sentiments
https://discussion.fool.com/zscaler-qs-as-33995027.aspx?sort=who…

As stated before, you’ll have to thoroughly study the company and its prospects and then make your own decision. Nobody can make it for you. If I feel very uncomfortable with a stock, its valuation, outlook, etc., rather than loosing sleep over it, I sell.

Hope the above helps a little.

im

2 Likes
9

im,
thanks. I think I have seen those links. Will look at it again. I think Zscaler is a great company. Just that market seems to be in an unforgiving mood now. Any miss or lower guidance is punished lot more. And ZS happens to be our most expensive stock. But looks like a lot of people have already started bailing out on Zscaler since yesterday. Will be an interesting couple of weeks till E but before that we have Nutanix.

10

Hardware stocks typically have been known to have very little moat. Yet, most will agree that Apple and Nvidia have excellent moats. I was researching Apple sometime back. If I recall, between 2013-17 the stock returned a CAGR of 15-20%. Yet the revenues were up only about 7% CAGR. The way Apple does it is that they buy back boat loads of their shares, like 4-5%/y. Plus the dividend. Their high margin and cash flow makes that possible. All together even if they grow rev. at 5%/y over the next 5-10y the stock could still return 10-12%/y which is market beating. I realize this is very low for these boards but some may find it acceptable. Now Apple may not grow rev 5% next year but there is a very strong likelihood that they can grow at that CAGR over the next 5-10y. Nvidia has similar advs. 35% net margins etc and likely to grow faster. But there are some specific issues surrounding it, some of it peculiar to hardware stocks as you have pointed out.

1 Like
11

As a business man who completely switched everything to the cloud and hammered the server into. the ground as if it were a Nazi invader, I can attest to the incredible productivity gains it enables, and the flexibility it enables, and the simple simplicity of it all.

Further, think of the simplicity, relatively speaking from the vendor perspective of running a SaaS business vs. running say even a dominant chip business like Nvidia. Just imagine how complex running Nvidia is. You have to be brilliant in so many things vs. what is required to run a SaaS. That is not to say Nvidia will not be back.

https://finance.yahoo.com/news/nvidia-apos-earnings-were-sol…

A bit more of an upbeat perspective, and it seems possibly correct. Nvidia may give back what it took over time as we get into May/June of next year. Nvidia still has the only ray tracing chips in the business, is still the only AI chips (except for some other niches like FPGAs that are a nice niche, but nothing more, as is Google GPUs) and with the now disaster scenario valuation and the sentiment this dark, even a hardware company may make a comeback and give back what it took, if you just forget what it was, and take it for what it is right now. A company that missed by $700 million in guidance, an issue that is likely to clear in 2 quarters, and perhaps come May/June the dominance that is Nvidia will start with at least some upside surprises.

But even at that, again, think of the incredible complexity of running Nvidia vs. running Mongo. Mongo has the software done, it has the market position, it just needs to keep iterating competently, it just needs to deploy its salesforce competently, and run general business things competently. Mongo no longer needs to be brilliant. Nvidia cannot stop innovating and being brilliant and running all these verticals, supply chains, staying ahead of everyone else (which it is still doing a great job of), prcing properly et al.

Compare it to Abiomed who faces possible device malfunctions and liability lawsuits. Albeit, except for these, Abiomed seems to face a much simpler market opportunity itself as it to just has to remain competent. It no longer needs to be great. And its business model is nearly SaaS in that reorder rate is nearly 100%, meaning it gets as much recurring revenue as any subscription business.

Another thing to look at when comparing investment opportunities. Buffett loved Coke. Carbonated sugar water. How simple and repeatable! Perhaps, all things being equal, go with the simpler business model if you have to choose.

But just more things to think about. Yes, Nvidia may be of interest again next year, but don’t toss out your capital losses to buy back in now, and make a decision in regard to how complex you want the businesses in your portfolio to be. If you need brilliant, Nvidia is about as brilliant as they get.

If you need absolute impentarable moat, that is simpler than a Nvidai Abiomed.

If you want something that simple, with repeateable business like that, which is growing faster (but less profitable as most SaaS are) but without the liabiltiy risk, there are few businesses simpler than MDB or Twilio or even Zscaler. Simple does not mean easy, it simply means once you get them going and in their market position, they are much more simple to successfully run with less things that can go wrong. Hackers is the one thing that can hit any of them, even Abiomed (just ask Abbott who had 6 figures of product recalled because their heart product could be hacked while it was working the hearts of patients - yikes!). Mongo has no such catastrophic concerns absent leaving some real serious unpatchable security opening.

Tinker

13 Likes
12

and hammered the server into. the ground as if it were a Nazi invader,

A most inappropriate comment.

7 Likes
13

A different hardware area is with the companies that make equipment to make chips. Lam Research (LRCX), KLA-Tencor (KLAC), and Applied Materials (AMAT) are three large ones. There are others, but I’m not sure if any are of similar size. There was a failed merger between KLA-Tencor and Lam over a year ago.

These companies have taken a hit lately, just like the other hardware stocks, but they seem to have lower risk than the semiconductor makers, since for one thing, they would generally have a broader range of customers, and likely not to be as dependent on one of them as many of the chipmakers are.

14

“Who wouldn’t choose the SaaS?”

I don’t want to get off on a tangent to much but here are a few simple reasons not to choose SaaS and/or “the cloud” offerings:

  1. More expensive in the long run.
  2. One blip (power outage, virus, bug in software) from SaaS provider and large swaths of clients crash all at once (even with multiple data centers). Solving these issues points to why #1 holds true.
  3. Rural Internet speeds stink.

These are a few reasons at the moment that cause many clients, even my own company, not to go “The Cloud” or use SaaS services.

Don’t get me wrong, I like what our SaaS companies are doing and the many that are to come in the future. As mentioned in posts above, CapEx has become king (to me at least) and falls in line with what seems to be a change in society’s philosophy from ownership to monthly payments.

Leasing everything is great as long as the business/individual has money coming in to make the payment. What happens when the economy tanks and the company/individual has a bad quarter or no paycheck?

From and investment perspective, maybe we need to make sure that we analyze who the actual customers of each SaaS company are. Maybe we may want to stay away from SaaS companies that have a certain % of customers that would not be able to weather an economic downturn?

21 Likes
15

Semiconductor capital equipment manufacturers are extremely cyclical and not ideal long term investments. All this talk about how hardware customers “have to find new customers each month”, these guys have it much worse.

Having said that there is money to be made in investing in them, if you time it right. If during a lull more and more pieces of good news comes out indicating things are getting better, you buy. The share price will still be low. Then sell when the opposite occurs (if you wait that long). You do not want to get sucked into “the price dropped and it’s not really that bad news and it’s cheap” thinking with these guys. It will be dead money for years.

And besides that, they generally do not have a broad range of customers. There are only so many semiconductor manufacturers out there.

As for the semiconductor industry, there are indications that the market is softening. But it’s coming off a period of huge constraints however. It’s still taking a year to get a reel of chip resistors.

Another industry that can be difficult is retail or clothing. I can’t even begin to count the “fads” that came along where a growing clothing company suddenly became unpopular. We can start with crocs. Canada goose may be different. But it’s another one of those companies I would not want to buy on the way down.

1 Like
16

1. More expensive in the long run.
2. One blip (power outage, virus, bug in software) from SaaS provider and large swaths of clients crash all at once (even with multiple data centers). Solving these issues points to why #1 holds true.
3. Rural Internet speeds stink.

Show me the evidence of each. More expensive? More expensive from owning and operating and updating and distributing and powering and debugging and having to upgrade every year plus having to pay people who do nothing but work on this for you (and not your business) is cheaper?

One blip…when. I have had far less down time since going cloud than I ever did running my own servers. It is not even close. Azure has multiple facilities around the country, the world, one data center in a rare even goes down, it all moves quickly to be covered elsewhere.

Name a time Azure has gone down leaving its customrs in a lerch.

Rural internet speeds…who cares. If you are a businessman, it is the cost of business to make sure you are connected with proper bandwidth. So you pay a premium to get it done (whether or not you were in the cloud you would do this if you are a legitimate business of any real size and not just making a few hundred dollars a day). Or you move the business to where there is bandwidth.

None of these criticisms apply to SaaS vendors. There may be a cost/benefit tradeoff that each business needs to make on cost. It is not given that the SaaS will be more expensive or less expensive. That just the depends. The rest is just red herrings.

Tinker

11 Likes
17

More expensive from owning and operating and updating and distributing and powering and debugging and having to upgrade every year plus having to pay people who do nothing but work on this for you (and not your business) is cheaper?

I used to develop and sell ERP software, mostly in a pre-SaaS world. There were some dramatic differences in pattern of use. One client company, a publisher, discovered that I could make modifications to the software much more cheaply (and better!) than their prior vendor and so they brought me a constant stream of customization requests, ending spending more money on the customizations than they did for the initial system. Another, an importer/distributor, spent next to nothing beyond a few initial things like customizing the invoice form. All of my customers stayed on maintenance, but I know that there are other vendors where the customer had made the modifications and drifted so far from the original base product that they knew they could never upgrade and consequently stopped paying maintenance.

With SaaS, though, while there is less outlay up front, one keeps paying and paying and paying. Case in point as has been mentioned here recently, I am one of those not pro photographers who used to upgrade my Photoshop every three versions or so when a really compelling new feature came out, but who otherwise was perfectly satisfied with an older version for some years at a time. For me, the subscription model is vastly more expensive.

And, I think that you miss on the potential limitations on speed and reliability. I am sensitive to it because I have been doing a lot of on-line things since the late 80s and so have experienced the improvement, but also the unevenness of the improvement. Some businesses are located in places where they may have the ability to run reliably with local systems, but there is just not any option for a high quality, high bandwidth internet connection.

3 Likes
18

Oh and SaaS most often forces the customers to accept any changes to functionality and adjust the old, very well established workflows whenever the marketing of the SaaS provider decides, something is cooler or better and kills the previous versions.

While generic availability of data and service is technically good, the TOC usually are so completely inacceptable that really sensibly thinking companies slowly start to think twice.

Cloud and SaaS sounds less expensive, until you really need to make sure data and service is available on your terms, not the TOC of the SaaS provider.

As an investor, I bet on the victims of the marketing offices and believe, the SaaS will grow fast in the coming years. But, e.g. by following what Tim Berners-Lee is working on (Solid Pod), I think there are signs that at some point in the not too distant future, owning hardware infrastructure that keeps your data and basic management of that data alive and on premise, will come back. Probably sooner than it now seems.

3 Likes
19

1. More expensive in the long run.
2. One blip (power outage, virus, bug in software) from SaaS provider and large swaths of clients crash all at once (even with multiple data centers). Solving these issues points to why #1 holds true.
3. Rural Internet speeds stink.

Show me the evidence of each. More expensive? More expensive from owning and operating and updating and distributing and powering and debugging and having to upgrade every year plus having to pay people who do nothing but work on this for you (and not your business) is cheaper?

To me - cost of ownership can be higher compared to actual ownership. Rental/subscriptions are usually more expensive than purchase although total cost of ownership needs to be considered.

Another area is control. Whether it be security or any other aspect of control, sometimes companies want to have direct control or need to exercise direct control to satisfy clients etc. Another example is companies might not want to be feathering the nest of a competitor - particularly in the Amazon case. Also existing SAAS providers may not offer sufficient differentiation in the service requiring in house or custom made solutions.

These reasons are why there is still a mix of cloud and on premise or a mix of SAAS based plus owned.

I would think 5G should resolve the bandwidth & speed issue.

A

1 Like
20

Tinker

Show me the evidence of each. More expensive? More expensive from owning and operating and updating and distributing and powering and debugging and having to upgrade every year plus having to pay people who do nothing but work on this for you (and not your business) is cheaper?

SaaS is more expensive in some situations depending on the size, type and product of the business. I have this debate on tech forums all the time so I know it cannot be hashed out here. SQL on Azure is the perfect example of how it can be much more expensive in the long run for a small business.

One blip…when. I have had far less down time since going cloud than I ever did running my own servers. It is not even close. Azure has multiple facilities around the country, the world, one data center in a rare even goes down, it all moves quickly to be covered elsewhere.

Yes, AWS and Azure have become better but each one has had an outage in the past 5 years. Don’t forget, if you want the ability to use those multiple facilities, you have to pay. You want backup, you have to pay.

Name a time Azure has gone down leaving its customers in a lerch.

Azure outage Sept 4 2018 effects a south region. Don’t forget, Internet gets cut off at customer premise due to many reasons.

Also, don’t forget the dyndns outage of 2016. Please, oh please, don’t be one of those that says, “That will never happen again.” At this time DNS is at the core of the Internet but they are working on changing that too.

Rural internet speeds…who cares.

People/businesses in rural areas.

If you are a businessman, it is the cost of business to make sure you are connected with proper bandwidth.

Ummm… You do understand how telecommunication landscape works in the US correct? We should all know, at this time, T, VZ and the like control how much bandwidth you get and when you get it. The customer does not. Hopefully 5G solves this issue in the next 5-10 years.

So you pay a premium to get it done…Or you move the business to where there is bandwidth.

If it were only so easy. T and VZ and the like get to choose what they want to give whether you like it or not. As stated above, hopefully 5G and the future will change this ALL.

It is not given that the SaaS will be more expensive or less expensive. That just the depends.

So you do agree that it can be more expensive? Thank for confirming my #1 statement.

Don’t misunderstand. I believe TWLO, OKTA and even MS Office 365 are good deals and cheaper for different size businesses that need them. Yes, there are cost/benefit details in each service offered and there will be times when SaaS is more expensive.

6 Likes