When Andy Grove Saw The Future Of Networked Computing

Some things never change: 1.The telecoms industry will always lag the computer industry; 2. People will always prefer personal, local control over their information to relying on remote sources.

One of the most astute players in the tech world these past 40 years, Andy Grove, the great former CEO and co-founder of Intel, remarked on these characteristics a decade and a half ago.

Back in 1995 he observed: “”The computer industry does its stuff delivering increased performance and reducing prices because it’s a competitive and open industry with hundreds of companies in it, but the telecommunications industry has a government monopoly legacy which is not used to delivering its services in a competitive economy. We need improving bandwidth and declining prices and neither trend is being served sufficiently.”

How true that statement remains today. Where is LTE? Where is Wimax? Where, in most places, is 3G?

In 1998, another prescient observation from Grove was his reaction to the idea of the ‘Network Computer’ (a thin client device which got its programmes and information from the Web – what we now call Cloud Computing) which was, in the mid-1990s, being promoted by Larry Ellison of Oracle.

“What’s made this industry very exciting”, observed Grove in 1998, “is that people speculate on new platforms, and new applications, but the world’s population has shown – generation after generation – is that they prefer the attributes of local control to those of a dumb terminal.”

When asked if Ellison’s idea was driving the adoption of networked computing, Grove responded: “The sun doesn’t rise because the rooster crows.”

Networked computing was happening, he explained, not because of anything Ellison said, but because a couple of hundred million people around the world had computers technically capable of accessing the communications networks.

Nonetheless, Grove enthusiastically promoted the idea of the networked computer and, asked if he was promoting networked computing just because it helped sell more microprocessors, he replied:

“I get paid to serve Intel’s interest but I don’t get paid enough to lie. It’s clearly in our enlightened business interest that consumers’ eyeballs move over to the PC from the TV but we actually believe this stuff. It may only be a skip and a jump from you saying ‘there’s Grove giving a marketing message’ but I believe the message to be true. We walk the walk as well as talk the talk.”

Grove has turned out to be right. Everyone talks about The Cloud today but the most successful devices, like the iPhone, have masses of space to store apps, information and programmes. Everyone, it seems, still prefers local control to remote access.

And why? Because the other part of Grove’s observation is also still true – the capabilities delivered by the computer industry (which now includes smartphones) still vastly out-perform the capabilities provided by the telecoms industry.

Although the telecoms industry has LTE and Wimax in its locker, it is typically slow and reluctant to deploy them, while a decade after the 3G licenses were sold, coverage remains spotty at best.

Which is why people don’t want to rely on such links for their basic computing needs.



  1. ” the network operators have to be dragged. kicking and screaming to deploy it. Look at BT with ADSL – thingss almost got to the stage of a public scandal before they started offering it.”
    Sorry David but this simply isn’t true. BT and the telecomms equipment manufacturers such as the one I worked for at the time were began planing the rollout in 1996 and were ready to go bigtime a couple of years later but the SC guys who promised working silicon in 1996 failed to deliver ‘to spec’ product for almost another 3 years.

  2. Well, El Rupester, I look at it like this. When a new generation of CPU or new density of DRAM, NAND or hard disc appears – the PC makers put them straight into their boxes.
    When a new telecoms generation comes along – DSL, Fibre, 3G, 4G – the network operators have to be dragged. kicking and screaming to deploy it. Look at BT with ADSL – thingss almost got to the stage of a public scandal before they started offering it.

  3. I think you are being too harsh on telecoms industry, and I disagree with your conclusion on network computers.
    “We need improving bandwidth and declining prices and neither trend is being served sufficiently”
    In 1995 you were lucky to get 14Kbps on a dial-up wired link. You’d pay an ISP £X per month – and in UK you’d pay BT £Y per hour opf call charges too.
    Nowadays most people in developed countries get >1Mbps — and 10MBps is common for cable, with 100Mbps if you live in Asia.
    And the prices have fallen: I pay less for my 20Mbps cable than I did fifteen years ago for dial up. If you want to you can get DSL for effectively nothing.
    Compared to computers that’s pretty good. Although the CPU may have followed Moore’s Law, the PC as a whole thing (including software) hasn’t, and the performance or utility I get from it certainly hasn’t: Word & Excel cost the same as 15 years ago and seem to deliver about the same…

  4. To take cheese’s rephrase: “many of us” choose not to sacrifice local control, in particular for what you first mentioned for “photos or emails or travel plans”. No way will I use a remote IMAP or Webmail server for my email, personal or work; my photos are firmly rooted in my home PC’s RAID mirror with occasional backups when necessary; travel plans get PDF’d and printed when firmed up.
    We all segregate our local/remote storage according to personal preferences based on experience, expectations and a even a bit of prejudice… “optimization” if you like. Of course I still do useful work with a computer so maybe that’s just my prejudice but I do not understand all the people who put their trust in remote email services – gimme a POP Server or go away.:-)
    Right now, I’m watching the accumulated snow pulling a tree branch down so that it’s damned close to the telephone cable which carries my ADSL and I’m not too worried – I can still work; if it gets to the power cable then I lose more than just access to my data.
    No doubt things are going to get interesting in the not too distant future: how far will the wireless phone device converge with the personal computer; does the “zero client” (see Pano Cube) have a corporate future? I lived through the corporate data center horror which the IT guys are currently dying to re-establish and the analogy I have is this: do you want to have a car or do you want to get on the bus? From my POV I still see things more from the Grove perspective – Ellison was, and still is, SQL*Nuts.

  5. I see the reasoning on both sides of the debate and I have a couple of remarks.
    Firstly: We often do not clearly distinguish storage, content and content consumption / experience. This leads to a lot of confusion.
    Storage is ubiquitous, cheap and fast. Content needs storage, but content is not necessarily ubiquitous, cheap, or fast. Consumption is a different animal altogether. It is becoming ubiquitous, but not necessarily cheap or fast.
    So consumers are left with an optimization problem – find the optimal choice for storage, content choice and content consumption experience.
    This is a non-linear and subjective equation. Most of us will find a solution to this, but will never be happy with it. A bit of Zen in our lives will do us no harm. 🙂
    Secondly, for the control freaks, here’s my 2 cents. Reality TV happened sometime back. And many of us could barely tolerate it, let alone relate to it. What is happening now is the internet version of the same. You can ignore it, wish it away, or join it. That’s your choice. Reach, ego massage, self promotion – this is the new black. And btw, the computer is not a “computing device” anymore, so spare us the references to good old FORTRAN or IBM 360 and PDP11. We are talking of devices that help common people share and discuss their otherwise boring lives. Let me rephrase a bit the sentence from my earlier post: Yes, to be clear, many of us sacrifice (knowingly or not) local control for reach and remote access. The world would be a very different place if not for this.

  6. Martin, I think you and Andy Grove are right. Somehow giving up control goes against the grain of human nature.

  7. Greetings:
    I started to use computers some 45years ago: punched cards and (if VERY lucky) occasional acces to a terminal. All software, access and storage under the control of the Computer Center (yes,I was in the USA and the term IT was some distance of…). Then came the time-share:same central control though and eventually – and at some expense – the PDP8. Joy! Local control of everything . Next came the really low-cost microcomputer (Z80/8080) with tapes. Cheap! Controllable! Always available! Then the IBM PC and all it’s clones – Cheap(ish), good selection of programs, adaptable, expandable. MINE.
    Now we are supoosed to give back control, storage and access to some nebulous entity and let them ‘look after’ my requirements (and data). No thanks. To ‘cliche’ (or saw) ‘Been There, Done That (etc)). Cloud computing is just simply a bad idea (and this argument has nothing to do with the Telecoms System – which is another story (quoth he ruefully from a remote rural location….)

  8. Well Peter it’s an interesting point you raise about the relative scale of investment. I think it’s about 10x. But whereas the IC industry trend capex figure is 20% of sales – so it spends about $20 billion a year on sales of about $200 biliion, the telecoms industry only spends $225 billion a year on capex when it has revenues around $4 trillion a year. If the telecoms industry had the same sales/capex ratio as the IC industry, it would be spending $800 billion a year on capex – in which case, maybe we’d have ubiquitous LTE, Wimax and femotcells, and we’d all feel happy about saving stuff in the Cloud because access would be instant and reliable.

  9. I don’t think the IT/telecoms industry comparison is really valid. It’s like saying that consumers refit their bathrooms every fifteen years but water utilities take 150 years to replace water mains. There’s a huge difference in the scale of investment.
    Concerning where people store data and run applications, I don’t think users really care. What they care about is speed of access and availability.

  10. Yes indeed I remember that too Steve. I wonder if that’s why Schmidt was so firm in saying No when asked in Barcelona if he was getting into the network operating business. If he’s actually thinking of getting into the cable business, then there’s why. I do hope you’re right. It’s what the world needs.

  11. Several years ago that there were stories about how Google was buying “dark fiber” in large quantities, like this one: http://www.voip-news.com/feature/google-dark-fiber-050707/. IIRC, the stories all ended with, “… and no one knows why.”
    I’m really hoping that this pilot is how they start making use of it, for all the reasons you’ve already enumerated. Couldn’t agree more: here’s to greater inclination!

  12. Well, Steve, I suppose 50,000 people is a start. And I’m very grateful to you for telling me that. I had no idea. So I’m wrong to say ‘disinclined’. I should have said ‘the pity of it is is that Google is not more inclined to invest in the links’! Hopefully the trials will go well and they will be more inclined.

  13. The pity of its is Google seem disnclined to get into providing the links.

    Does this not count? http://www.google.com/appserve/fiberrfi

  14. Thanks, lexdabear, I’m sure you’re absolutely right. The pity of its is Google seem disnclined to get into providing the links. Maybe they’ll bag up Clearwire when it runs out of money!

  15. I follow your argumentation. The MNO/telecoms is the bottleneck. They have all the power to make the communication lines fast or, as it is now, hold onto their business model and do a short-term business, extracting most possible revenue at this point in time. If one looks from this point of view, it makes sense why Apple and Google are doing what they’re doing. With a snip Apple turned the complete market upside down and is now #1 in smart phones. Google I consider even more visionary. Own phone was just the first step. More amazing was the reach into frequency band, and now high speed internet plus energy. Watch and learn or be an ignorant and disappear.

  16. Well Cheese, I’d come at it from a different direction – if Cloud computing is the big thing, why do PCs come with 320Gbytes of storage (and getting bigger)? Why do iPhones have 32GB? Why do we download apps and music and store them locally instead of accessing them on-line? Why do we keep OSes and application programmes on our hard-discs? I would venture to suggest it’s for the same reason now as it was back in Ellison’s NC days – i.e. the links are so flaky no one wants to rely on them.

  17. A couple of remarks:
    What do you really mean by local control-vs remote access?
    All our data is stored on the internet – be it work (inc. blogs such as this), or photos or emails or travel plans …infact every single aspect of our lives is on the net, including all our savings and insurance. This data can be accessed from anywhere, and that qualifies to be called “remote access”.
    There is a semblance (and growing awareness for the need) of local control of remotely hosted data, but we all know that facebook (for example), doesn’t delete your data as soon as you want it to be gone. We have no clue what other cloud based services do with our data. In fact, apart from trust, there is not much else guaranteeing local control. And still, people do not mind loosing local control. Yes, to be clear, we sacrifice (knowingly or not) local control for reach and remote access. So that’s my first disagreement with Grove.
    Secondly, Grove is a visionary, no doubt. But his remark on the sun rise vs rooster crowing would need a bit of dwelling upon. With Intel pushing Atoms as heavily as it is now doing, the successors of Grove have proven him wrong – Intel (which Grove probably relates to the Sun) is making Atom because the roosters (Acer, Asus, you name it) are asking for these chips. And they will gladly move to Arm if Intel focuses on heavy duty chips instead of the Atom. Does Intel have a choice?
    Finally, the high ground of technology that Intel stayed put in the last several decades, is thankfully over. The paranoia of Grove et al resisted the commercial application of Intel’s own Moore’s law by keeping prices a constant (with clear implication on profits) in a near-monopoly market. There will not be many tears shed if this monopoly Sun sets. Ironic as it may seem, it is the telecom (or should I say mobile) industry that is at the forefront of this.

Leave a Reply

Your email address will not be published. Required fields are marked *