In the beginning there was ARPANet together with various other government and university networks, which enabled you to use command-line interfaces to send text-based messages over packet networks using open standards. This was a huge deal at the time. Then came the hyperlink-based World Wide Web, let’s call it Internet 1.0, and let’s peg that stock market era to the Netscape IPO in 1995. Cue early web stuff, boom and bust in the market, online declared a fad by most people and by 2002 the whole thing was moribund and depressing, according to your regular kind of investment expert. Value was where it was at, once again. Let’s hear it for dividend growth at Southwestern Bell.
Quietly though, a few people started to talk about things then called ‘web services’ and ‘orchestration’ and ‘microservices’. In 2003 this was a hobbyists’ and specialist investor topic. The notion was, islands of code spinning away in datacenter CPU cycles which could be corralled and managed to create freeform applications. These apps would not, the thinking went, be monolithic and proprietary as most enterprise apps were at the time; instead they would be built of tens, hundreds, thousands of open source islands of code. Nobody paid any attention. Investors who grew up with proprietary source code as a store of intellectual property and therefore value couldn’t deal with the notion of open source nor the business model that dare not speak its name, the API-first company. None of this was helped by the general “set the code free!” attitude emerging among developers at the time because they meant, like, really free. As in, for free. No money exchanged. Who needs money, man? I mean, how many more lentils can you buy, really?
Between say 2000-2011, some solid companies were set up and IPO’d or sold for what now look like modest sums, a few hundred million, a billion or two, to major tech corporations here and there. Mainly companies built on older principles by older folks. They still trade today; Splunk ($SPLK) and ServiceNOW ($NOW) are among the poster children of the time and only now are they experiencing their declining years
There were also some true landmark IPOs, Google among them, and some M&A deals that looked odd at the time, but turned out to be incredible. The acquisition of YouTube by Google being probably the finest of its era. And in amongst the mix, some things started to hint at what was to come. Sun Microsystems, which at the time built monolithic boxes using proprietary silicon and source code - the ultimate in Silicon Valley control freakery - declared that what in fact
they were doing was, “building that big f——in webtone switch
”. To which end they bought a little-known open source database business, MySQL, for $1bn, which was a lot for a $50m revenue business at the time. Telecom companies were forced by regulators to expand fiber networks and increase the transmission speeds available on copper lines. Cellphones started to become somewhat more useful. Consumers began to expect to be able to transact securely online. Companies started to move supply chains online in a way dreamed of by the Covisints and GF-Xs of this world back in the 1990s. By 2011? Internet 1.0 was back on its feet.
Now, Marc Andreesen has brought two epochal gifts to this world. The first was his work on browser technology for the National Center for Supercomputing Applications at the University of Illinois - courtesy of some added entrepreneurial zeal from Jim Clark and a scattering of money from Kleiner Perkins, that became the Netscape IPO that kick-started the dotcom boom. The second was a lot less work for Andreesen - a throwaway line in fact - but was no less profound. In 2011, he noted famously that “software is eating the world
”. If you invest in technology that was the mic drop moment. More than the browser. More than ARPANet. More than whatever garbage AT&T or Best Buy are trying to sell you this week. What he said was, the code has been set free. But not in the way intended by sandals-wearing goatee’d Java gurus back in 03. No. The code has been set free to deflate the world. As money flows out of traditional industries - all
industries save for software - it will flow into software. That’s what he meant. And it has. And it will. And a brief rotation into value will, in our humble opinion, be just that. Something that happens every now and then when technology stocks get a little giddy.
And so from from 2011-2020 we saw what is often called Internet 2.0. Real companies generating real revenues, cashflows, strong balance sheets, the real deal. Online becoming a regular way of doing business and conducting your life. Come 2020, Covid accelerated the shift but it was en route anyway.
Off the back of that acceleration, we at Cestrian Capital Research declared the 2020s to be the Cloud Decade (you can read Part One and Part Deux of that clarion call here
2021 has seen growth investors take cover somewhat, and plenty of self-declared growth geniuses on FinTwit blow themselves up. Happens. But right now, if you think technology is all over, if you think high octane, high p/e, no p/e, big-multiple-of-revenue “stupidly valued” stocks (“Woodstocks” being our favorite new epithet) are destined for the Dumpster, we salute you. Plenty of people make plenty of money investing in and trading old-line names like Boeing, Lockheed Martin and so on. We would never argue the wisdom of that - indeed in staff personal accounts we usually own a healthy dose of these names. But we have news. Technology is gearing up for another run. A big run, if you ask us.
Internet 3.0 will in our view see the Internet finally achieve the tectonic shift away from the academic / military-research “it just about all hangs together and emails kind of arrive on time, sometimes” paradigm which was born and grew up with no fundamental security DNA nor even guaranteed transmission DNA, the libertarian nature of which fueled the impression held by consumers that if it’s online it must be free.
Internet 2.0 started to work properly, as Mr Andreesen noted in his now-clearly-prophetic blog post linked above.
Internet 3.0 in our view is going to be the Internet, hardened, fit for purpose. Remember Fight Club? Coming into the 2020s, the internet had become a wad of cookie dough. By the end of the decade? We think it will be carved out of wood.
The players are now jostling for position. As with the emergence of any new paradigm, you can’t really say for sure at this point who the winners will be, but you can point to some early leaders. We’re scouting them out in our various research services and we own them in staff personal accounts. We intend to hold them for a very long time because we think software has plenty more of a meal to make of the rest of the world yet. We might trade them here and there as a side hustle. But core positions? We hope to sell just before the machine becomes self-aware and declares money fiat or otherwise no longer a thing. Probably the only tradable securities by then will be CPU cycles in exchange for coolant.
Here’s two early winners we own and have very high hopes for.
Cloudflare ($NET). Quietly building both the development environment and the runtime of Internet 3.0. Read that again and think about it for a minute. Think we’re joshing you? Read their earnings call transcripts. Think this looks like Akamai of yore? It doesn’t. Think it’s Fastly’s nerdy twin? It isn’t. Think it looks expensive? It isn’t, in our view, as long as you zoom out. (Remember the exhortation from mathemetician Carl Jacobi, “invert, always invert”? In the winning Internet stocks, when they’re hurting, we exhort ourselves to “zoom out, always zoom out” and we aren’t talking about WebEx 2.0 here). You can read our most recent post on Cloudflare here.
ZScaler ($ZS). Building the onramp. Less ambitious than Cloudflare and more likely to sell the business early in order to pocket cash amounts. AT&T should buy it right now, in order to return itself to a pre-1984 monopoly transport provider, but they won’t, because (1) they won’t get it and (2) their shareholders definitely won’t get it. How soon they forget. AT&T, let’s not forget, was the crucible within which Unix was forged, the datacenter within which the application server first gained a foothold. If they had simply developed their software - Tuxedo for one - rather than selling it to wiser wallets, well, they wouldn’t be out appeasing dividend investors all day long. ZS is a lesson in how to run a business. They reported blowout numbers after the close yesterday. If you’re a Cestrian Fundamentals subscriber you can read our earnings review here.
Internet 3.0 pervades our work and our personal account holdings. We do own some Boeing. Some AT&T. We own a little United Airlines. And a bunch of metals. We think they can all make us some money. We post about some of these names sometimes. But the knockout punch, the retirement-fund-defining moves? We have those bets in tech. Because tech always outperforms in the end. Don’t believe us? Just ask Marc Andreesen. It’s ten years since software started its meal, and trust us, it’s still on the amuse-bouche.