This is presumably my last post of 2021 and the audience is primarily myself. I wrote 16 posts this year (not including this one). Three posts were Cloud City Meetup event updates (we’re waiting to be able to return in person, so it could be any year now), with 13 more substantive posts, which is one more than 2020. Word count, however, plummeted by 20% to a mere 10,177 (my apologies). Traffic was down slightly from 2020. The top three sources are search, Twitter and then LinkedIn. Hacker News traffic has plummeted but also didn’t have any posts trend there this year.
I mined my perennial favorite topics with three posts on CAPEX, one on the evolution of cloud computing, (only) one making fun of IBM for pretending it is relevant in the cloud era, and posts on automating the Federal Reserve, Microsoft history, and how SaaS might play out. New topics include two posts making fun of the New York Times’ banal and hypocritical tech coverage (with more, albeit poorly threaded reactions on Twitter), metaverse real estate prices, and Starlink performance/why telcos suck. The top posts in terms of traffic were the annual cloud CAPEX recap and the Cloudless Cloud Company.
As usual, I have big blogging ambitions for the new year and still have some half-written early pandemic posts I might finish. Topics from last year’s list still to explore include edge networks, the scope of AWS, and the perfidious financialization of software. Newer topics include geopolitics through the technology lens, institutional revitalization, and how an ebb tide due to DCF recalculations might play out in the startup ecosystem.
“The three hypercloud companies – Amazon, Google, and Microsoft – collectively spent almost $97 billion on CAPEX in 2020, up 32% from 2019’s $73.5 billion.”
It is bonkers how much of Amazon’s pandemic-swollen revenue got poured back into CAPEX in 2019, when they spent more on CAPEX than any other company on the planet. Looking forward to seeing how 2020 compared.
Looking at Digital Ocean’s numbers underscores just how vast the chasm is between the hyperclouds and anyone else. The only way to compete with the hyperclouds is asymmetrically, chipping away at their edges and targeting very specific use cases and audiences. Full frontal assaults seem inadvisable.
Starlink is great. Traditional telcos suck (and none suck more than serial re-brander with a recurring customer satisfaction problem Lumen née CenturyLink née CenturyTel née Qwest née US West), and who, in lieu of shaking their fists at the sky, unleash their lobbyists.
Oracle is still a clown, not a cloud, when it comes to CAPEX, but they have promised to spend $4B this year on CAPEX. That isn’t nearly enough to put them in the hypercloud leagues, where they’re on the order of a $100B behind, but is at least directionally correct (and they’ve stopped claiming they don’t need to spend CAPEX, because, something, something, underpants, profits, etc.). We’ll be watching to see if they followed through, because Oracle statements always go in that “trust but verify” bucket.
I offer some more self-aware column ideas for On Tech, the world’s worst tech column.
People seem to assume nothing can go wrong with the SaaS model. But what happens if a price war breaks out?
Delighted to see the SaaS sprawl term getting some traction.
I got triggered by one particularly insipid On Tech column and show how almost any of their columns, with minimal search and replace, also apply to the New York Times itself.
UPDATE: also now sweeping the world of fashion.
A quadrennial tradition since 2009. Which is more transitory, inflation or Fed Chairs?
If only IBM put as much energy into their product portfolio as they do into managing the press, industry and financial analysts. Mean reversion seems to be the best story they can come up, which Barron’s duly regurgitated. I probably need to write a post on how the cloud is finally coming for mainframe, which strikes at the heart of IBM’s latest stock tout (they’re using a mainframe cycle to pretend they’re a growth company, while changing the comps by divesting their worst businesses).
“Unimaginative, rent-seeking skeuomorphism.”
Interesting things are afoot further up the stack in the cloud.
The “supercloud” is not a thing.
The Back Catalog
Posts from previous years account for over a third of views, and sometimes specific posts have retro moments. My most popular post of all time remains A Very Cold Take on IBM, Red Hat and Their Hybrid Cloud Hyperbole, which holds up remarkably well, accounting for almost 5% of views this year. A look at some early 21st century Microsoft history written in 2013 also had a moment.
Onward to 2022!