Enough people pinged me on this – and I’ve seen so much baloney analysis – I guess I might as well say it publicly, openly, and for posterity sake. I also wanted to let a little time go by so people could let things sink in, and they could reflect.
I put out a tweet with my $0.03 cents, but while Twitter is awesome for certain things – let’s just say the form doesn’t enable rich complex thought… or for that matter a good method for talking about complex geo-political issues or policies :-)
Here was my tweet:
That’s it in a nutshell. You can stop here – and save yourself a TON of time…
…But, you all know me well enough dear reader, that why say something in 140 characters when 4500 words will do?
So then I thought twice and said to myself: “Dude, do NOT comment. Haters will hate, and say you’re being negative. It’s too easy to caricature you and Dell EMC as the ‘Goliath’. Just keep your mouth SHUT”.
Then, I realized: “when have you EVER kept your mouth shut?”.
Time elapsed. Then I had an epiphany. Today is inauguration day – so… “Park your ego buddy. No one is going to read your silly blog, Chad!”
Sidebar: As a proud and public Canuck, like my average fellow Canadian patriots - I tend to be center-left. Canadians, on average think that “good government” is not an oxymoron, but something that can exist and is worth striving for – it’s a Royal Flush (rare – but possible and worth striving for) but not a leprechaun (imaginary).
Yes, I’m an exec sponsor of the LGTBTA community in the company. Yes, I’m not a fan of the incoming administration, and I’m not shy about it. It’s not that the new administration are right or left (good people can have different points of view – and there have been great Presidents and administrations from both sides of the aisle) – rather I believe great leaders bring people together, not apart. I believe great leaders have great empathy, great perspective, great courage, great passion – but also great humility. I believe that great leaders #BuildBridgesNotWalls. I believe great leaders need to lift people up, not bully or belittle them – especially when those same people are challenging them (it takes courage to challenge power, and it takes courage for power to listen to challengers). I believe great leaders look for facts, look for data – and accept that data – especially if it means adapting their thinking.
I don’t see any of that in the incoming President. Furthermore, I don’t think what is happening or the new President reflects the best of the America that I know.
But, I have other things I also believe that for me, supercede some of the above . I believe in the US. I believe in the people. I believe in democracy. I believe in their institutions that check power, and check each other. I believe in the free press that needs to continuously, perpetually challenging authority.
Even more importantly – I believe in diversity of thought.
That always means embracing people who disagree with me, listening, learning, and yes, furiously debating. I have great friends that are excited about the new President. I don’t get Trump, but I DO get them. I don’t respect the Presidential choice, but I do respect these friends of mine. If I believe in my principles, I need to speak out, but also to keep an open mind – to listen… so I’m choosing to be hopeful and optimistic.
The clear, and obvious “right” thing for me to do is simply say nothing (as a senior leader in a big org, a big team, a broad diverse family) on both the HCI topic (a nit) and the new Administration (important) – being milquetoast, being boring is the “safest” route. That isn’t really right. I need to be what what I am, right (on both the HCI topic and the Presidential topic, right?)
..Then I thought a little more and came to the obvious way to make a critical decision with great import – Twitter poll! Survey says…..
So if you want the whole story, all the bad judgement, far more information that anyone could want (and bringing together two totally unrelated topics) – at least one man’s perspective (remember, I have my own biases – so use your critical faculties and judge for yourself!) …read on!
So – let me breakdown and elaborate on the 3 thoughts in the tweet.
FIRST PART = “Congratulations.”
Some thought this was a back-handed compliment. It’s not. I mean this with complete and total sincerity.
Look – if we’re honest (and why not be?) there is simply no way to cast the Simplivity exit as a successful one. After a quarter billion in VC funding, and a valuation in the last round that was north of $1B, it was a necessary exit.
A negative way to characterize it would be a “fire sale”. A positive – and IMO more correct – way to characterize it would be a “wise business decision in the best interests of the employees, customers and investors”
As I said in the tweet – I have participated in 3 startups.
The first was a really cool company with amazing technology called Metricom. They made a product called Ricochet which was absolutely face-meltingly awesome and ahead of its time.
Imagine if someone came out with a “6G” cellphone radio tomorrow that was literally 10x faster than everyone else, and more importantly a dense mesh network made of low-cost components that could support 100x the subscriber density of competitors upcoming 5G counterparts. That was Ricochet. It was 1999, and state of the art at the time was a 56Kbps Novatel Wireless or Sierra Wireless PCMCIA card (remember those!) or a 9.6Kbps pager-style Blackberry. Along comes this startup with a dream of a national mesh microcell network that could (for real) do 256Kbps to almost 512Kbps to a small modem – the speed of the best ADSL services. You could stream a video, jump on a VPN while on a train or in a car. It was a revolutionary product. I fell in love.
That story ended with a post IPO crash – nothing to do with the product or the technology.
We decided to take the funds from the IPO (almost $2B!) and go for bust (racing the 3G network rollouts that we knew were a few years out). We knew that if we proved the tech worked at scale (check!) and the customer acquisition rates were good (check!) we could easily go back to the public markets for more funding to operate the network (network builds are VERY expensive businesses and require massive capital expenditures and have terrifying operating expense burn rates).
That decision made TOTAL strategic sense in 1999 – surely, we can just raise more money! Of course, it was a disaster two years later in 2001 when we needed the capital injection. Capital markets closed after the dot com implosion and subsequent recession. I had to lay off the SE team I just built – hundreds of good friends. My colleagues running sales, marketing, engineering – they all did the same, and then we fired ourselves and closed the doors behind us. It sucked.
Lesson learnt? Great technology alone does not make a good business – and the name of the game is a good, sustainable business.
I keep an old Ricochet modem on my desk as a constant personal reminder of that lesson.
One of the other of the startups that makes for a good story was Allocity.
Like Metricom, Allocity had an amazing dream that starts with a simple idea: storage sucks. Why not have all storage be software-defined, completely distributed and run on commodity hardware? And ultimately shouldn’t you just use apps and have the storage be invisible? No LUN management, no zoning, heck, no storage management console at all. Storage would become invisible and would just integrate into the application stack. I started talking about “invisible infrastructure” at the time – I’ve noticed some have picked up the idea :-)
It was so much fun – I was SE #1, worked with a great group of people, many of whom are still friends today, wrote some code, did a lot of QA, heck did customer support on Friday and Sundy.
I would trundle around a suitcase with 3 laptops and a 1GbE switch and plop it on customer tables and show them how I could transmogrify it via software-only into a small iSCSI SAN, and how storage behind Exchange 5.5 would just magically get created, snapshots magically occur, IOps and capacity adjusting automatically as mailboxes got created – no storage management at all. It was a hyper-converged application appliance.
Allocity’s founding idea was genius – and so right, but SO ahead of its time. For perspective – it was 2002. Funny sidebar story – it’s back then where I met Dheeraj Pandey (Nutanix CEO and founder) for the first time – but that’s a story for another day.
What happened with Allocity – where was the strategic error? There were two errors:
The first error is that we were too early. Customers (even those that loved the product) were skeptical, our implementation surfaced as an iSCSI target and Microsoft still wasn’t shipping the 1.0 iSCSI initiator. Cisco iSCSI initiator was GA, but kinda sucked. We had to make our own initiator – and while it was fine, all the uncertainty made it all uphill.
The second error was an important lesson for me – displacing storage incumbents is really, REALLY hard. Storage is high risk (storage failures are #EPIC #FAIL because data persists), and data is hard to move – so people are reluctant to move unless your value prop is really, really good. Though there are success stories storage startup failure rates are high because of these things. Customers really want someone that they know will stick with them through thick and thin.
Allocity was forced to do a hard strategic reset, which is a move startups do this all the time.
We jettisoned the idea of building our own storage target. Interestingly, many of the people on the storage target team lived in Boulder Colorado and joined a new startup called Lefthand Networks, and many ultimately found themselves at another similar startup called Solidfire.
I’m not claiming (at all!) that they used Allocity IP – but rather that the IDEA (distributed storage stack on commodity hardware, iSCSI targets) needed more BAKE time and more MARKET ACCEPTANCE time.
Heck, it helped that their implementations were better than Allocity’s too – building on years of experience, wisdom, failures, and frankly maturing technology ingredients. In the same way that Lefthand was better than Allocity, Solidfire was even better still – and then Solidfire expanded their storage stack principles to many other ideas that were smart and core to their design and business (API-driven, focus on the SP vs. taking on the giants directly in the Enterprise).
Hats off to those Solidfire folks – smart engineering, but also strategically smart to stay out of the crosshairs of the giants. They don’t just make good socks and Cards against Humanity decks ;-)
In our “reset” Allocity pivoted to the 2nd half of the founding idea – and focused on making exsisting storage “suck less” by building application-integrated automated policy on the big players in the mid-range storage market at the time (HP EVA, EMC CLARiiON, NetApp FAS).
Business picked up – and we made the fatal startup error of narrowing down our engineering, our pipeline, our co-selling to the company who was partnering with us the most – which was EMC.
When it became clear we had no path to “IPO exit”, and we needed to have “sale exit” because the VCs were done – we had zero leverage. Net? EMC got some great IP and a small number of great people, and one runt of the litter (little ol’ me).
Lesson learnt? Three linked lessons – 1) startups need to often make several strategic pivots, so get too fanatical about your idea; 2) don’t take on giants and incumbents frontally – that’s just stupid; 3) great startups are ALWAYS thinking strategically of their exit strategy. If they aren’t – they are stupid.
I keep one of the old “gold master CD-ROMs” of the Allocity LiveEx 1.0 bits on my desk as a reminder of the lesson.
Look, doing a startup is HARD. It requires passion. It’s deal to deal, funding round to funding round.
Startup life is constantly changing gears, with each funding round having totally different “character”.
- Round A is all about proving the idea.
- Round B is all about proving the tech getting your first 10-50 customers which is the crucible where you harden your pitch and value prop.
- Round C is all about getting to 100-500 customers and prove the business and get to profitability.
- Round D is in vogue now because infrastructure startups almost are never able to self-sustain after 3 rounds – but this starts to create huge pressure because at this point, only massive exits are good (because there has been so much funding), and dilution starts to become really bad.
- Anything other rounds after that (bridge loans, strange private rounds, etc) = oh oh, something is not working to plan.
BTW – ProTip: if you’re joining a startup in their C-round (heck in their B-round in some cases), it’s almost always too late. It’s not a startup – it’s a small business. My “rule of thumb” I try to impart to people around me thinking about this startup path and wondering if they would be smart to join a startup goes like this:
If you find out about them at a VMUG, they have a product, and good presentation that names a reference customer… then it’s too late to be called a “startup” and to make any real coin as a new employee in any of the exit scenarios. You’ve missed the most fun parts.
If you love the idea of a startup, and have a passion for an idea – START ONE. Do not join a startup (particularly a late stage startup) because you think you’re going to make a lot of money – you won’t. If you ARE going to start a startup, my recommendation is NOT in infrastructure.
Now, startups only have 4 exits and I’ll list them from worst to best:
- Close the doors (Metricom case).
- A “fire sale” where they don’t control the sale terms, have low leverage, and usually have only one suitor (Allocity case, Simplivity case).
- A “good sale” where they control the sale terms and usually have multiple suitors (3PAR case, Data Domain case, Isilon case Solidfire case, Nicira case).
- IPO (Pure case, Violin case, FusionIO case, Nutanix case).
My “rule of thumb” is that for 25 startups, 18 will end up with exit type one; 4 will end with exit type two, 2 will end with exit type three, and 1 MIGHT end with exit type four. In infrastructure startups, it’s even more grim.
Heck, even the “good” outcomes (3 and 4) are themselves just the start of yet another gear!
- If it is a “good sale” outcome (3) – the next chapter is defined by “can you survive/thrive inside the acquiring entity” (a new kind of hard!!!).
- If it is an IPO outcome (4) – the next chapter is defined by “can you actually make a sustainable business out of it”. Clearly and sadly this was a “no” for Violin, FusionIO. The jury is still out on Pure and Nutanix. Personally, I think they will survive. But, the post IPO euphoria dies down and gets replaces with a new challenge, a new kind of hard… All of a sudden profitability and shareholders matter, not just the fun part of customer acquisition and high “sizzle”. All of a sudden, you can’t hand-hold every PoC with your best engineers. All of a sudden, you can’t saturate the partner ecosystem with hardware. All of sudden it’s the boring part of being a successful business.
I’m not trying to be negative, I’m just trying to paint a more complete picture.
Q: With all that – why the heck do people do startups?
A: Because: It is a TON of fun, you are burning with passion for an idea that you just have to get out, and you’re surrounded with people that share your fanaticism.
“But Chad,” you may also ask, “If startups and that ecosystem are so gruesome, why does it consume so much press, and get so much attention?”
Because: frankly there’s a disturbingly obvious self-serving “pump and dump” ecosystem of VCs, banks with sell side businesses with a wafer thin Chinese wall between them and the analysts and parts of same said banks who are backing the startups. After all – when people put in a quarter billion into a startup like Simplivity, or almost half a billion into a startup like Nutanix in funding rounds – dammit, there BETTER be a good exit.
Furthermore, you have people who are press or analysts who are paid on eyeballs who want sensational headlines and a sexy story. Some do great research and do fantastic work – but sadly, many don’t A lot of digital ink is spilt with perspectives that fold like a cheap card table with the slightest scrutiny.
Sidebar – think of what that means about the role of the press, the role of all the people in the entering new Administration and era. We know that we are surrounded by fake news. We know that propaganda from foreign states was a real thing and a real factor in the election. We know that one of the core principles of a democratic (small “d”) systems – one of it’s foundations is under attack from foreign powers. Americans (and democratic people of all nations) of all political stripes should be infuriated. With whom? Ourselves. If you believe in freedom, the answer is real, critical thought, and bad ideas need to be beaten by good ideas. Fake news cannot be unchallenged. Wrong opinions need to wither under attack from FACTS. This FAR more important in the geo-political sphere than in the world of startup hype, but the same idea nevertheless.
Hmmm… As I think about this - this sounds an awful lot like I’m crapping all over startups and the VC/feeder ecosystems that create them.
To the contrary! I LOVE startups!
Startups are an engine of innovation. Startups disrupt markets. Startups cause chaos and make life so so hard (but fun!) for the giant incumbents. Startups screw with established business models. Startups make for fun press for a basic human impulse – who doesn’t like a David vs. Goliath story (whether it’s true or not :-)
It’s been really interesting to learn from the other side over the last decade – EMC Ventures is FASCINATING (and has only strengthened as they became the center of the new Dell Technologies Ventures team). Listening to episode 8 of the “In the Hot Aisle” podcast is a peek behind the curtains of an amazing world, an amazing world, and something I’ve suggested as near-mandatory listening for folks in EMC and I now extend to new heritage Dell readers.
The day the deal is announced, someone created a phony twitter account with Doron Kempel’s (Simplivity CEO and founder) name and posted a series a tweets that while kinda funny superficially, but as I thought about it kind of pissed me off.
Q: Why did it piss me off?
A: Whatever you think (or I think) of Doron is irrelevant. YOU try to do a startup, and see if you can do better. If you’re slinging poop from the sidelines (and anonymously), that’s just cowardly.
Perhaps the most ironic thing – while it’s not like I did forensics, it seems like it might have been created by someone at another startup. Show some humility. Startups are really hard, and if you’re arrogant, well, pride cometh before the fall, hubris and all that jazz.
Funny sidebar story – I happened to sit beside Doron on the flight back form VMworld Barcelona (think it was 2010). He had no idea who I was, but I knew him so played a little game with him – because I can sometimes be “that guy”. I pretended to be ignorant about he space and asked him what the “Simplivity” sticker on his laptop was about. He explained. I asked a deeper question. He answered. I asked a deeper question. This went on for about 15 minutes at which point we were deep into the IO path and how their VMkernel integration worked, and how their integrated data protection approach functioned. At which point I felt guilty and told him I was busting his chops… Moral of the story – don’t trust people on airplanes, particularly short blonde Canadians :-)
Look, I don’t know the guy – maybe he was good, maybe he was bad – but you know what, he’s created multiple startups, and there’s been multiple good exits – what evs. And MORE IMPORTANTLY - the story isn’t about him – it’s about the whole Simplivity team.
Hallelujah for startups and all the crazy passionate nutbars who fight, who create, who build – no matter how the story ends.
This is why the first part of the tweet is sincere: CONGRATULATIONS to the Simplivity team. It may not have been the exit you wanted, but there’s incredible courage in what you did, and there is nobility in the fight itself. Be proud of that.
Now, on to the next part of the tweet:
SECOND – “Few will be left standing.”
I’ve been calling it for a while. This is the year for HCI startup Armageddon. Why? Ummm – have you SEEN how many HCI startups there are? The HCI market cannot, will not sustain 20+ HCI startups, I’m sorry. Hey, if you can find a crazy VC to fund you creating YAHS (yet another HCI startup) – go to town… I’m just saying I’m skeptical.
While the HCI market is growing like mad (the IDC says 105% CAGR Q3’15 to Q3’16)… but it’s still only a baby market of about $2.5B. I saw some silly analyst state that HyperFlex is a $2B business for Cisco… Argh. This is just intellectually lazy, stupid and wrong. The whole HCI MARKET is ~$2.5B right now. I saw another analyst put Pivot 3 in the “leaders” section of their analysis…are you HIGH? Pivot 3 I’m sure has great technology, and happy customers that analyst talked to, I’m not slagging Pivot 3… but an HCI leader?
Look on this I need to be a little careful. I have data that is not public I need to tiptoe around… But let’s find some real, public data we can chew on. Where shall we turn?
- In my opinion, IDC are the analysts that are the most grounded in data. They specialize in quantitative data analysis – and they are the big dog at that. I’ll caricature the IDC folks as the scientist community you want to have telling you about global climate change – no opinions, just data, constant measurement against scientific principles. Not perfect by any means, but stark, harsh facts.
- Gartner is of course the big dog when it comes to qualitative analysis. I’ll caricature the Gartner folks as the people at a party you want to hang out with – they have great stories, perspectives based on talking to all their friends. Gartner uses data real analysis and a ton of inbound client inquiries to flesh out those perspectives.
- There are many other great analyst firms that do great work (even the one that called Pivot 3 a leader in HCI – we all get stuff wrong periodically).
- It’s a value to have multiple different analysts’ point of view – not just on this, but almost anything.
- The folks at IDC are also a machine – I don’t how they do it, but their market share data is AMAZINGLY precise. They seem to know (for better or worse) nearly exactly our bookings/revenues, customer counts, and more.
IDC tracks the HCI market – so no one needs to GUESS about what is happening. Their latest study is Q3 2016, and it is here. Q3 was a hard quarter for us (weird first weird “Dell/EMC are merging, but not merged” step) but we still did ok.
All I have done in the chart below is taken the hard public numbers, and put it into a form that you could easily replicate yourself with public data.
If you look at the data in that Q3 report (public!) and over the last year (public!) it’s pretty darn clear. Nutanix has the majority of the revenue share (still mostly NX, but a LOT of Dell EMC XC - how much is in their S1 filing which is public, though of course now a little dated). But woah – what is going on with Dell EMC?!?1
The answer is pretty simple.
Dell EMC XC, built in partnership with Nutanix is doing great.
The other part of the story? VxRail is the little engine that could VxRail was launched in Q1, started to roll in Q3, and in Q4 was cooking. When Dell EMC and VMware decided to lean in and lean in hard into HCI, we created an incredible (and profitable – more than what can be said of the startups) HCI appliance – and it has been rocketship. VxRail is a 10 month old baby that is running well against a lot of competition. It VxRail perfect – HECK NO. And the Dell EMC and VMware team is working on making it better every single day. So, Q3 was OK… and Q4 was OUTSTANDING.
Now, IDC hasn’t done their Q4 tally, but I can’t wait. I know exactly what the VxRail revenue was. I know exactly what the VxRack revenue was, and the Dell EMC XC revenue was.
Conversely, I don’t know what Simplivity was, or what Pivot 3 is, or Maxta is, or Hedvig is. I don’t know what Nutanix was (but we all will when they report earnings).
If Dell EMC isn’t #1 with Dell EMC VxRail, VxRack, XC in Q4, we will be DAMN close – and the yardstick we have set for ourselves is #1 in HCI by the end of 2017. I’ll go on record – we will get there much earlier.
… Heck we’re going to be close in Q4 even if you exclude Dell EMC XC (which you shouldn’t, its a real and important part of our HCI portfolio).
…. Oh, and that’s not even including customers who are building their own HCI using HCI “ingredients”. In fact, in the “Other” category (a whopping +50% of the HCI market!) customers building their own HCI using VSAN is by FAR the largest part of “other” and VSAN is surpassing a number of customers that dwarfs the Nutanix (the largest single player noted in the IDC data) customer count – by almost 2x.
Why am I making a big fuss about this? It’s because there are FACTS, and fact-free society can lead to fake news, which can lead to a lot of really, REALLY wrong ideas winning – not because they are right, but because they “sound right”.
- FACT: HCI is a very important part of the infrastructure market (over time, I expect that it will form the majority of the infrastructure on premises – though that’s my opinion!!) and is the foundation for private clouds of all types. It’s a far more important trend than the NAND transition. If All-Flash is the “IQ test” for this year, HCI and Hybrid Clouds build on HCI/CI are the “are you breathing?” test.
Beyond VxRail, we will partner insanely with VMware on VxRail’s bigger brother VxRack SDDC – and together, VxRail/VxRack SDDC will form the foundation for massive simplification of our on-premises half of the VMware-powered Enterprise Hybrid Cloud. VMware agrees – its right in this post:
“You’re looking for a turnkey system powered by VMware Cloud Foundation and you don’t want to worry about building your own infrastructure. You don’t want the headache of maintaining your infrastructure. You want one throat to choke for all your hardware and software support needs. VxRack SDDC is for you.”
And of course, in the same way that VMware will operate in an open ecosystem as noted in that post (VMware Cloud Foundation software-only for customers who want to “build it themselves” & VMware on AWS/Softlayer as clear ecosystem examples) – the same is true in the other direction. Dell EMC XC is an example. Going further, Dell EMC will use VxRack FLEX as the basis of our Azure Stack offer through the year. We will use VxRack FLEX as the engine behind other non-VMware private cloud parts of hybrid cloud stacks.
- FACT: there is a real measuring stick – which is market share, which can (and is) measured by objective 3rd parties.
It’s not about how you FEEL. It’s not about what you POSTION, or POSTURE in a press release. It’s not about your PERCEPTION of what’s going on, or a vendor you you do/don’t like. It’s not what you HEAR from one customer or another.
Marketshare don’t lie. And when it comes to marketshare data, no one does this better than IDC. Use it.
- FACT: While HCI is incredibly important, having a CI and HCI portfolio is going to matter a lot – while the majority of x86 workloads (by volume) can be happy as a clam on HCI and SDS models – there are important workloads (smaller in number, but they represent a lot of value, a lot of importance to customers) that need CI, need traditional external storage arrays.
For perspective, here’s the CI market (same IDC raw data) in the 4 quarter period of Q3 2015-Q4 2016.
This is why people who thinks it’s a “HCI vs CI” world are also just, well, wrong, and are spouting fact-free baloney. The winning position clearly is to have a CI/HCI portfolio. CLEARLY over time HCI will be a bigger market that CI (a few years of a 105% CAGR tends to do that), but CI being a $4-5B market (it really is a little bigger than IDC measures) with a 9% CAGR where the VxBlock business has a 22% CAGR. Wow. Furthermore, as Dell EMC, we enjoy 65% market share in this category. That is pretty darn good. It says we must be doing something right. Always a TON to improve, and you would be amazed how seriously we take this.
- FACT – the market for HCI startups is drying up FAST, and suitors are getting scarce.
IMO, you will NOT see another HCI startup IPO, and I’m sorry if you’re someone rooting for that outcome, or you think that I’m being mean or arrogant. That’s not where I’m coming from.
Imagine a startup trying to enter the market where you have giants like VMware, Dell EMC, Microsoft who are fighting furiously, passionately for their life. We are not “giants asleep at the wheel”, but people filled with hard working people who follow some form of the Andy Grove “only the paranoid survive” maxim. Those startups also face and a very strong newcomer (not a startup any more guys, sorry!) in Nutanix.
This is why Simplivity HAD to exit. It’s also why HPE HAD to do something. Marketshare don’t lie.
Someone noted that this was the “next domino to fall” post the Nutanix IPO. I think that’s right, but with each subsequence domino to fall, the outcome will become progressively less awesome for said domino.
There’s another aspect, another FACT that results in the conclusion that HCI market will consolidate down to the server vendors and likely Nutanix.
Yes, in an HCI, all the magic is in the software. You can try to create hardware-oriented differentiation in HCI, but it better have a MASSIVE advantage. Otherwise it’s a binding that will bite you. So, why then is it natural that it will consolidate down to the server vendors? Let’s continue!
- FACT – a large portion of the PRICE of an HCI appliance/rack scale system is the commodity server hardware. The MARGIN is all in the software (because that’s where the magic is). But, think it through – that statement of fact leads you to a conclusion: you need to have a killer x86 server supply chain to make a killer HCI appliance or rack-scale system from an ECONOMIC OFFER standpoint.
- FACT – while customers want HCI models that don’t BIND to a hardware stack, the OVERWHELMING customer preference is for appliance and rack-scale system offers, and software-only variations are used for learning, play, very rarely as a deployment vehicle. Q: Why is this? A: Customer are picking HCI because they want infrastructure to “go away”, become invisible. Customers are picking HCI because they want lowered risk, simpler lifecycles. Customers are picking HCI because they want clear and simplified support models, they want someone that they know will be with them when hardware inevitably fails, they want someone that can do it around the globe. All this leads to “HCI systems will tend to optimize around hardware/software integrated packages”.
So then – who is left after HPE as server vendors? Of course Cisco. So… What is my two cents about the Cisco “pick up Nutanix” speculation that started to circulate?
Who knows :-) And of course, Nutanix is are a public company with shareholders now - if someone makes an offer that their shareholders approve, they would take it.
However, even basic math doesn’t hold water (again, frustrating to see people speculate when even a cursory analysis of the FACTS leads to some pretty clear conclusions).
At the current Nutanix share price – and assuming a multiple that their shareholders would support – it would be a $6B+ acquisition, which is an insane price given their revenue and profitability.
Since Nutanix are public – if they reach cash-flow positive state, they don’t have to sell to anyone. For this possibility to play out, Nutanix would have to crash (not a little, but a lot), or Cisco would need to make a MASSIVE strategic pivot, not a small one.
The Cisco folks seem to be pretty confident in HyperFlex. Now I don’t know how Cisco’s HyperFlex business is doing – but IDC does. Marketshare don’t lie.
I just don’t see Cisco acquiring Nutanix as logical move unless Cisco does a giant strategic pivot, and does it fast.
So where does all that lead you? To my 2nd comment in the tweet. There won’t be many left standing - that there is inevitable consolidation. That the server vendors will be the “acquirees”. That to win in HCI you will need to have a killer global x86 server supply chain as this is the basic building block for HCI, for the future of storage, heck – perhaps over time even for networks. And, of course, you will need incredible HCI software stacks – likely one that you control completely, and several that you partner with (because customers will still want choice).
Now – to the third part of the tweet…
THIRD – “Bring it. Competition is good.”
I ended the tweet with a 3rd observation – that we welcome competition, from HPE, from startups, from everyone.
It makes us better – forcing us to be the best us we can be. It makes things fun. We like to fight fair, and we like to win. We love it when we make customers and our partners happy – and they vote with their dollars. I don’t take our success for granted. I know it depends on having the right strategy, the best people, and fighting hard – as hard as the most passionate startups, as hard as the largest giants. In the end, it’s about execution.
I will be very transparent (as always) - we intend for Dell Technologies to be that CLEAR market leader in HCI. BTW - we think we already are, and the marketshare data is the FACT-based measure, so watch IDC closely. I certainly do. I aim for us to enjoy the same rate of success that we enjoy on CI.
How will we do it? Via the strategy we already have – because it’s working:
- We will do it with HCI appliances.
- We will do it with HCI rack-scale systems.
- We will do it with HCI that customers build themselves using HCI ingredients like VSAN.
- We will do it with HCI appliances/rack-scale systems that are incredibly tightly integrated with VMware and Pivotal – they are family, and customers want strong opinions.
- We will do it with open ecosystem of partners like Microsoft, SAP, RedHat, Google, Nutanix and many, many others – and we have multiple partially overlapping ecosystems as Dell EMC, VMware, Pivotal, etc – these ecosystems are friends, and customers want choices.
- So long as people want to be friends with us, we want to be friends with them. If they don’t – that’s OK too. We’ll compete, and customers will choose.
We’re pretty maniacally focused over here, we like to win, and I’m pretty stubborn :-)
So – there you have it, that’s all the brain goop and thinking that went into my little silly 140 word tweet.
On this Inauguration Day, I am hopeful that others have similar depth behind some of their tweets – on matters that have far, FAR greater import than HCI. #BuildBridgesNotWalls.