This is how Microsoft dies
Putting the 3rd largest company in the world on deathwatch may seem laughable, but I reckon there’s legitimate cause for concern. I’m worried that within this decade, Microsoft may deplete the resources on which its recent stellar growth depends: legacy enterprise cloud customers.
My ears pricked up when, on the All-in podcast last year, Chamath Palihapitiya proclaimed that 2022 will be roughly “peak Big Tech”, though he left open exactly which companies he felt fell within that category. Curiously, he went on to pick Satya Nadella as his CEO of the year.
That’s because Satya has done a remarkable job of addressing cultural issues within Microsoft from the decades of the previous leaders. He’s also kept the company out of trouble at a time when governments have had their sights set on Big Tech for their anti-competitive practices and apparent power over populations and even governments.
During his time as CEO, Microsoft’s market cap has risen a staggering 2 trillion dollars. By this measure, Satya Nadella is one of our all time greatest corporate leaders, though it’s probably fair to say that a large amount of that success is showing up at the right time.
Microsoft is an extremely impressive business: it’s perhaps the world’s greatest company. I hold its stock and will do for some years yet. So why then do I worry it’s at the beginning of its end?
Put simply, Microsoft is peak Industry 3.0 and I think its failing to collect customers in Industry 4.0. What’s more, choice of cloud vendor has proven to be really sticky. It’s still early days of this megatrend, but so far customers tend not to switch cloud once they’ve began moving workloads.
I’m concerned that young companies don’t choose Microsoft products, and that its products lack simplicity and UX obsession. Its products seem far more attractive to top-down Industry 3.0 corporate decision-makers, than to the discerning people in new companies that now make decisions for themselves. CIOs choose Teams, while teams choose Slack.
I worry that spending billions exploring and opening up new markets isn’t followed up with excellent execution, but creates opportunities for other product-led companies to see, enter and own. Up until now, this has been the other way around: start-ups innovate and Microsoft creates a me-too product that cleans up, a manoeuvre that’s been made possible by central software buying within existing Microsoft customers. But this sales-led model is changing fast.
Much of Microsoft’s record run this last decade has come from the long tail of last century’s companies moving database servers to Azure and subscribing to Office 365, and then rushing to get Teams stood up while forced to keep workers at home. How long do they have until the enterprise reservoir is drained?
Also a concern is that in Nadella, Microsoft has a great sustaining CEO but less of a great risk taker when it comes to the type, frequency and sheer ambitiousness of innovation it will need to keep a business of eye-popping dimensions at the very top. Just how do you keep a fire this big, burning this bright? Are they failing often enough?
Apple is said to be building an electric vehicle, while Microsoft has only its HoloLens product. Its only major, visible bet on the future involves hanging heavy black plastic from your face. For most people today, that’s just weird, so taking this wearable idea mainstream will require the world’s best industrial designers and the sort of product, supply-chain and marketing obsession that only Cupertino seems to have.
And that’s not all. From my experience working in Industry 3.0 companies, I’m not convinced Microsoft’s current Azure subscribers can make the shift to becoming software companies and could eventually succumb to disruption from teams wearing flip-flops who prefer to code on macOS and host on AWS, Vercel, Cloudflare Edge or even on a blockchain.
If we look at the broad trends in people and technology today we can see that Microsoft isn’t skating to the puck anywhere near fast enough.
People, work and devices
The world has already moved from 20 years of using a PC under a desk to a range of new form factors. This trend will continue with wearables and voice assistants, in car UIs, devices around the home like the Amazon Echo Show and augmented reality glasses, plus stuff far beyond my imagination.
Crucially, we’re transitioning from using a PC in an office to a laptop in the kitchen. Or a tablet in the coffee shop. Or a phone in the gym restaurant. And perhaps just a watch and a Bluetooth headset, whilst walking the dog.
People are leaving old-fashioned corporations for companies that use devices in this way to offer fluid modes of work and they choose tools that work brilliantly on these form factors.
Moreover, I suspect that the companies most comfortable with super flexible work also let their employees choose their own tools and allow them to choose their own devices. I suspect they buy hardware with an Apple on the lid and rent software that runs in a browser tab.
Furthermore, the power of mobile phones is at the point where they could become our primary work device. Innovations in low-energy but performant ARM-based chips and wireless technology could mean we’ll soon be pairing our phones to a large screen, keyboard and mouse when we want to sit and do some deep work.
There’s no Microsoft in this vision except possibly in its Teams product, but it may not be long until we move on to different (and more fun) ways to hang-out.
Microsoft know how to sell via the slow, risk-averse central purchasing gatekeepers within ossified regulated industries. But this Command & Control corporate management style is also on the way out in favour of more agile, cellular and self-organising company structures.
Displacement of last century’s companies
The companies joining Microsoft at the top of the Fortune 500 used to be banks and petrochemical companies. They’re now either outright software companies or integrated hardware and software companies like Amazon, Apple, Tesla and Nvidia.
The best product experiences are the best software experiences running on the best hardware. Elon Musk deeply understands this and so too does Rivian’s RJ Scarringe; both have invested heavily in owning the entire tech stack within their vehicles.
Leaders of incumbent Industry 3.0 manufacturers have looked on in amazement at new competitor’s stock prices and how they’ve been relatively unimpeded by a global chip shortage, a supply squeeze that apparently impacted the oldest, chunkiest chips the most.
Alas, incumbents have no experience in providing the environment needed for creative designers and software engineers to do their best work and will struggle to attract and retain talent.
For two decades, they’ve been offshoring this yucky business they call “IT” and forcing their overpaid nerds to work in suits, in cost-centres surrounded (and vastly out-numbered) by pseudotechnical bureaucrats. These out-moded companies are Microsoft’s core customers. Yikes.
Software began eating the world in earnest more than two decades ago and yet it is a trend that is still gathering pace. Open source software written by enthusiast and collectives, running on community hardware threatens to profoundly upset the entire financial services industry.
Even if Decentralized Finance is regulated into oblivion, there were already a group of FinTech behemoths looking to attract retail deposits by offering slick, integrated services and generally better experiences than our ancient, talent-starved banks can muster with the ink still wet on their latest tech outsourcing deal and their get-back-in-the-office-or-be-fired mandates.
We already pay for stuff by waving our Apple and Google devices over the counter, and wouldn’t it be convenient if we just had our salary paid into our phone and did away with the bank card and its app?
The old banks use .NET, Windows and Azure. The new banks use Apple’s Macbooks, Google’s Golang, Amazon’s AWS. And soon they’ll be using Solana and Arbitrum.
Today, new businesses are started and led by software nerds and designers who see the technology they choose for their companies as deeply tied to their identity and culture. They build from the ground up using whatever most rapidly and inexpensively gets them to product-market fit, whilst maintaining the street cred needed to hire top talent.
Meanwhile, our brightest young developers can smell a cargo cult “digital transformation” Enterprise job spec a mile off. Having been born with 1s and 0s coursing through their veins, they’ve no idea what digital transformation even means. It’s something people who play golf and wear suits are doing to help catch old men up with the last 20 years.
GitHub was a coup for Microsoft as it bagged the undoubted capital city of the world for open software development. It’s growing as a platform for private projects, too. Unlike LinkedIn, it also happens to be cool to be seen there–at least for now.
However, some influential programmers are beginning to say less than complimentary things.
Another triumph for Microsoft has been its free, lightweight code editing tool VSCode, which has quickly become the de-facto editor for web design. It’s strange to see the hipsters running (indeed loving) a Microsoft-branded app on their MacBooks–it’s like seeing kids in retro dad trainers.
But I just don’t think developers have been completing the look by rocking an Azure cloud subscription. In fact, I think Microsoft would do better to repackage and offer Azure cloud services under the well-loved GitHub brand, using GitHubs’ UX team.
Meanwhile, thousands of huge enterprises are sitting on IT that’s years passed its end of life and are paying punitive support costs to Microsoft to keep it patched and secure while they procrastinate on its replacement. There’s a good chance that they’ll replace whatever software is running on all those old servers with Silicon Valley subscription SaaS products that, behind the scenes, run on AWS and Vercel and, yep.
Or their most dawdling customers will have their lunch eaten by an Industry 4.0 start-up before they can even complete their Herculaneum upgrade projects. Many of these companies have employees in the hundreds of thousands, yet have more obsolete servers than they have obsolete employees. I know of an old European bank that has two old Windows servers for every employee! What a drag.
And when the gallant Microsoft and Amazon cloud knights ride to the rescue waving enormous enterprise contracts on the end of their lances, they discover that, for the execs of these institutions, relinquishing an IT empire built over a two decade career is um, um, can we reschedule next month?
Web apps are eating software
While software consumes our world, let’s take a moment to consider that the software itself is being rewritten for the web and will no longer be installed onto our devices. This is an important shift that’s been a long time coming.
Most of the installable software that was written this last decade has been for iPhone and Android. This is largely because the primary distribution channel is through the app store but because “native” software can access the hardware features on the target device.
But web browsers are fast closing the gap on what device features are available to app developers, mitigating the costly need to build and maintain a native app for iPhone and another for Android.
See What Is Project Fugu — Google’s Initiative To Unlock All Native Device Features For The Web
Even difficult workloads that have normally required installed software, such as the code authoring tools developers rely on, are moving to web-based editors and even no-code web design tools, e.g. Replit.
The effect is that for most consumers the only things they really need to look for in a new device is a web browser, RAM for all those open browser tabs, and a keyboard. This raises an important question, “Do I need an unfashionable Windows laptop when I’m only going to run Google’s Chrome, or will I be happier with that slinky iPad Pro?”
Another trend is the web app that replaces an Excel workbook. Many smaller entrepreneurs look to what kind of things are currently done in Excel, usually a worksheet and perhaps some macros built by an industrious corporate employee, but which can be better solved with a simple, dedicated app for free or a low cost subscription.
Excel is one of the few remaining reasons to need a large screen, and one of the few remaining reasons people need Microsoft Office. It’s a meal that needs a proper sit down with a knife and fork. But even Excel is under attack from no-code and low-code web apps like AirTable and Coda.
And education is moving to the browser tab. Kids take naturally to touch-screen devices so the tablet has become the standard form factor for education of pre-teens before they step up to more keyboard oriented work in secondary school.
Kids these days
I believe that if you want to know what the future will look like then look at what motivates kids and young adults. Maybe it’s just me but it feels as if creativity is being bashed out of children at a slower rate than prior decades.
I think this is partly down to the ease with which they can directly access markets through screens. Whether it’s music, stock art, jewellery, makeup advice or virtual clothes, the young seem to be making, sharing and selling a lot of content. The typical career path for a child born today is unknowable but I’d wager that they are less likely than their parents to spend years sat at a Windows PC in a skyscraper.
And then there are the spaces and places where they hang out. Roblox, Fortnite, Minecraft, and content sharing and messaging apps and forums of people like themselves.
A kid’s sense of belonging, or rebellious not-belonging, looks completely different to mine as a 44 year old. Online worlds and games that offer the ability to heavily customise your persona have normalized freedom of self expression from a very young age, making for an important new tribalism.
How will these kids will feel upon walking into an office for their first interview at a company with grey carpet tiles and grey people sitting at uniform rows of grey desks with old grey screens and dirty mouse mats and Internet Explorer running crappy intranet sites?
As a kid, I might have had a “crappy” Commodore 64 at home while my mum worked at the proper IBM workstation in the office. Today’s kids have a 5G supercomputer in their pocket and collaborate all day with their friends on Roblox, while their parents trek 40 miles each morning to sit down at a dust-filled Dell running Windows 7.
The nearest a kid gets to a Windows device today is an XBox. Even older kids are taught to code using non-Microsoft languages like Python and JavaScript and often on super cheap Raspberry Pi hardware that can easily interface with electronics.
The only reason a student might be interested in learning a Microsoft-branded technology is for game development using Unity, which uses Microsoft’s excellent C# programming language. However, C#.NET is now open source and cross platform, meaning that programmers no longer need to run a Windows machine to build apps, or use Azure to host them.
Since coding is on the school curriculum in many countries today, and VSCode has been so successful, it’s the most likely reason a young person will use a new Microsoft product.
Developers
Along with kids, coders are the other set of people to watch for a clue to where things are going.
I have coded almost every day for the last 20 years and over the last decade I’ve noticed that tutorials and screenshots have shifted from DOS commands and Windows GUIs to Unix shell commands and the three dots of a macOS window.
Here’s an example taken from the homepage for checkr.com, a company that runs background checks on potential recruits.
And another from AI community HuggingFace:
It’s no longer cool to use a Windowsy window.
The reason for this is simple. Before Facebook, Airbnb, Twitter and all the other tech companies, developers worked at industrial enterprises and used Windows. When they blogged about a new technique, or created a new tool, they did so from the Microsoft world.
Today, all that blogging, building of open source projects and sharing is done by programmers working at tech companies created in the 21st century. These companies did not start out by installing Windows NT onto a server and setting up a domain and an Exchange Server.
No, they just used Gmail and Google Docs and Linux servers and to remote into a Linux server was easier if you used a Mac. Much easier. And besides, off the back of the 90s-era bad Microsoft, they didn’t want to use Microsoft stuff.
So today, all the influencing is done by techies on macOS. And when the tech company alumni break off to start their own thing, they use what they know.
Ah, but what about mining rigs?
Right now, a new movement is underway: web3. These projects are interesting because the coders are often young and relatively inexperienced or old and highly academic. This is because blockchain development has been largely rejected by the incumbent programmer-influencers who make no bones about not liking it one bit.
The young tend to code on Apple Macbooks while the academics tend to dislike Microsoft anything, loudly preferring the Unix design ethos. However, when it comes to blockchain, there’s a problem: the GPU.
Most crypto projects are built on the Ethereum blockchain which uses a similar mechanism to Bitcoin to ensure the integrity of its data. This requires a juicy graphics card which requires a boxy PC with a big power supply, running Windows or Linux.
But this year, Ethereum will move from Bitcoin’s GPU-intensive proof-of-work to a proof-of-stake design which requires comparatively little compute, merrily running from an 8Gb Raspberry Pi.
I believe blockchain is likely to be a very large part of the future of apps and financial plumbing, and that the sceptical old guard developers will either concede or retire.
Moreover, blockchain could become be the cloud of the future. A distributed network of general compute running everywhere, probably on devices with a low power or open architecture CPU and a tiny Unix based OS.
It seems that Microsoft will feature in developer’s lives mostly via GitHub, VSCode and, to a growing extent, its OpenAI tie-up.
Satellite internet for the world
The new corporate IT department will be managing subscriptions and identities to the SaaS stack, and Active Directory logins for those in finance that need Excel.
The only thing getting in the way of the entire world working through subscription web apps from the beach house via an iPad Pro, is fast internet access where you live. SpaceX Starlink now boasts faster connection speeds than that of most average European internet service providers.
Getting high-speed internet access–and the digital economy that comes with it–to almost everyone on the planet has profound implications. Many millions of people will come online, initially through shared local networks and whatever older devices they have already.
But when this nascent user community upgrade, in the knowledge that learning, commerce and money flows through this critical device, they’ll be looking for either the cheapest way to run a web browser, or a status symbol. The other key thing will be low power consumption.
I predict the devices they opt for will run either Google’s or Apple’s OSes. The software will come via a web browser and will be coded by people on Macbooks in cafés, and hosted on AWS, Vercel, and yep.
