The Other Bubble
wheresyoured.at
wheresyoured.at
I agree that a lot of the AI stuff will fizzle out and be wasteful in retrospect, but that doesn't mean that companies are foolishly paying for SaaS they don't need or that the whole ecosystem is a bubble. I think most companies are run a lot better than the author believes, and if they're not then there's a huge opportunity for a more efficient competitor to come in and steal the market.
It isn't. The majority of tech companies are speculative and funded by venture capital, and what's worse, many of the largest companies, which actually do generate revenue, are monopolies
What this means is that efficiently competing is only a viable strategy if you can convince a small pool of potential investors - many of whom think "competition is for losers" - that it's a good play, while also avoiding being bought out and crushed by one of a few trillion-dollar companies
The dynamics you're describing are those of an open market, and with wealth disparity at an all-time high (so your customers have less influence on average) and financial scams fairly rampant due to decades of lax enforcement, most companies right now are not effectively operating in an open market
Buying SaaS is trendy conventional wisdom. It has benefits and downsides, but in my experience the biggest benefit is that VCs think it's smart to do and also often benefit from a company buying the products of other companies. Whether it helps with things like making products faster or better overall is close to immaterial in this environment
The game at companies that have achieved high valuations prior to IPO is to get to IPO or be acquired. To grow toward market dominance in their sector or to threaten an existing player while being cheaper to purchase than to duplicate rapidly.
Having worked at one such rapidly growing 'unicorn' myself, every decision made by senior management was made in order to please investors. Competition, our tech stack, even the product itself were basically irrelevant. What mattered was how the investors perceived the company - i.e.: how much we resembled the kind of company that they saw as likely to recoup 1000X.
This resulted in very strange cargo cult behaviour. Fo example my whole department existed to replicate a similar department in our major competitor. Incidentally that department has since been liquidated and replaced with a variety of off the shelf AI solutions. Our customers had little interest in the 'value' we created in any case.
We also purchased just about every SAAS solution imaginable. Because that's what shiny highly capitalised startups do. Our onboarding included two full weeks of 'interactive' videos, with tests.
It wasn't just SAAS either - everything we did was performative. Including custom outfitting a carbunkle office in a landmark building on the most expensive street of our country's capital.
The founders weren't eccentric nineteen year olds and our product wasn't some crypto scam or half baked AI offering, although it was itself a service. The founders' behaviour made complete rational sense. They were optimising to appear successful to the investors whose valuations made them appear successful to other investors. They used these investments to 'build a runway' for the tech industry contraction that was obviously coming, but also to gobble up every other similar company in their sector. I believe the industry we operated in is now essentially a duopoly, with EU level data protection regulations being the moat that prevents it from becoming a monopoly.
It's not that SaaS products don't offer some value, it's that they offer a nebulous, hard-to-quantify value, and when the CEO says "Can we ditch Salesforce?" the immediate response is "how will we track client relations then?" And while you can imagine a much more lean, custom-fitting CRM costing a lot less for the company, it's hard to imagine them finding it, either as a different SaaS offering or an internally developed project.
A big part of SaaS stickiness is organizational intertia.
Hard choice.
Unfortunately, this isn't how large org CEO compensation works. CEOs of large publicly traded companies are mostly compensated by the stock price. Either directly as it increases the value of their significant stock grants, or indirectly as their cash bonuses are tied to stock performance. Stock performance is mostly a result of quarterly financial performance. Therefore, directives that take a long period of investment are much less attractive.
Say you were the CEO of a multi-thousand person organisation. Your SaaS bill is through the roof, and so you want to in-house everything. Not only do you need to disrupt nearly every employee by changing the way they work (depending on what you're replacing), you need to enter into co-location contracts, buy hardware, and hire staff to mange that hardware. It's a very large up-front cost (likely in the tens of millions on hardware alone) that in theory will pay off in the long term, but investors and the market will see a spike in your assets on your P&L, and will question why you're tying up capital in hardware instead of building the stuff that you sell. That's assuming you have the cashflow to buy all that hardware in the first place.
I am sufficiently familiar with Zitron's work to recognize the irony in that it should be one of his pieces that prompts the thought.
But Zitron conflates this with a lack of potential or capability of A.I. itself. This is absolutely not true. Any developer who has spent any length of time pair programming with A.I. or using it to analyse/debug code will understand immediately what I’m talking about. 5,000 LOC files debugged, discussed or refactored in seconds. Bugs tracked down in an instant instead of an hour. When AI fits a task well, it adds immense, jaw dropping value and it’s clear to me that, with LLMs harnessing transformers we have discovered something new and revolutionary.
Where's the equivalent in customer support? or document creation? or any of these other mythical AI use cases? genuinely asking.
The article makes a good case that the SaaS bubble is deflating and needs a new hype cycle to keep investment up. AI makes sense for that, so at least that's one good use case :-)
The former could but wouldn’t necessarily follow from the latter. And why bet against the latter yet? Sure we’re probably on a sigmoid, maybe even near the top, but are you ready to put on your big short today?
Another possibility is capability & cost DO improve, somewhere between substantially & seismically, and this cooks a lot of SaaS worse than if it didn’t.
Plus he gets in one incredible groan-worthy pun:
> As a result, it's just cheaper to use Microsoft's faux-Tableau no matter how hard it Tablows
This rant against SaaS completely forgets what things were like before SaaS and where upgrades were only done every few years.
The idea behind SaaS is to align the buyers and the sellers on long term support for the software. Prior to SaaS, the software seller only had to earn the sale once every few years, and the seller also had a serious revenue problem in having to fund development for 3-5 years based on ever decreasing sales of the prior version.
The team doing bug fixes was understaffed, if they were staffed at all, since everything was focused on "the next big release". Software kept rotting away year after year until it became less and less usable, and you could only hope that your employer eventually upgraded to a newer version.
> Microsoft Teams is universally-loathed and regularly threatens to crash every time you load it.
He obviously never used the predecessor to MS teams, Office Communicator. It was... bare bones and failed at a lot of tasks very frequently, such as making calls. (It did have some cool features that, when they worked, were nice).
> entire companies that exist to sell other companies' (like Microsoft)'s software
Partner resellers have been a thing since literally forever, this is not new to the SaaS world.
> developers that cost six figures that develop just for Salesforce
Developers specialized in one platform have also been a thing forever, same as Oracle Developers or even Windows developers back when that was a thing.
> They also make money — though not as much — by selling a variety of managed services.
Every large tech company has been selling enterprise support plans since, quite literally, mainframes were invented.
When I worked there, Microsoft's enterprise support was absurdly good. If you had a high enough support plan, bugs could get escalated to the dev team who made the product and they'd roll a fix just for your particular issue. Sadly that type of support is mostly non-existent now across the entire industry.
> The goal of these programs, as I've suggested, is to anchor you to an ecosystem. The more money you spend — hiring developers to code specifically for one platform, paying for specialized support from one platform, adding users and add-ons and storing data and doing stuff on a system — the more costly it becomes to leave, and said cost becomes only more burdensome the larger your organization becomes.
This is not new to SaaS. Platform lock-in has been a play for decades. Microsoft did it in the 90s, IBM did it prior.
Everyone wants to achieve lock-in because large software projects are just not that profitable without a surrounding ecosystem.
> Well, have you ever used a piece of software at a company you work for that sucks? Was it sold by Microsoft, Salesforce, Google, Atlassian or another big SaaS company? Well, it was probably bought by somebody who doesn't use the software, and it'll cost far more to remove than your annoyance matters.
Again, not a SaaS thing. Arguably SaaS has made this better by funding continual ongoing development of software. Prior to SaaS, it was "here is the software, enjoy what you have because nothing is getting fixed!" (Outside of maybe 1 or 2 updates shipped on CDs across 3 or 4 years!)
> Microsoft in particular has monetized this chaos by selling you an entire suite of apps.
Microsoft has always been selling a suite of apps. The suite has expanded over time, but Office traditionally included an entire database! Products come and go from the Office Suite, but again, this is nothing new.
All this said, is there a bubble? Probably. Imagine if everyone in America is working for a company that is buying into MS's SaaS portfolio. At that point, if MS wants to grow revenue 10%, the entire economy needs to grow at least 10%, or some other portion of the economy has to shrink. Now if the software makes everyone 15% more productive (woohoo!) and if MS can shave off 10 points of that for themselves, so be it.
But at some point we'll have reached maximum productivity, maximum output, and software won't help us do our jobs any better.
> You'll notice, of course, the very obvious problem: that it isn't obvious what any of these AI-powered products do, and when you finally work it out, they don't seem to do that much.
The author, in this post and the last, craps on AI, but does so from the perspective of someone who has only read press releases from mega-corps.
Right now companies are adopting AI and saving tons of money. From voice AI agents (that you cannot tell are voice AI agents!) at call centers, to dramatically speeding up how fast unstructured data can be processed. Real money is being made, real savings are being had, and real productivity is happening. It Office co-pilot amazing? Meh. Is Google probably losing money hand over foot with its AI features? Likely so. Are a lot of low-key industries adopting specialized AI tools to save on money? Yup.
Of course it doesn't work as advertised, but that's not the intent. The intent is to market the software. It's an investment.
Where I completely agree, though is that AI is eventually going to be a black hole that in a decade will be looked back on as what put Microsoft, Amazon and especially Meta into years long holes financially given the size of the bets they're making. I'm not as bearish on SaaS companies due to this though because their investments are very small and they can quickly pivot to the next hot thing that no doubt emerges in a few years.
I do think the hyperscalers are united by a common thread in their pursuit of AI which is not SaaS companies but rather the CEO's like Zuckerberg and Dell having an axe to grind with Steve Jobs and Apple as well as Google and having this unending desire to beat them and create a new platform. Mark Zuckerberg is a smart dude. His instincts for running a social media company are second to none. But he has never successfully launched a consumer hardware product because doing so is extraordinarily difficult and I don't think is overcome simply by money and a lot of smart people.
If you follow the arc of the smartphone it goes all the way back to the telegraph and it all boils down to one thing- enhancing communication. Steve Jobs and Apple understood how to move forward the phone in a revolutionary fashion and the device is more or less now perfected. Interestingly, Apple hasn't made huge investments in AI. Nothing prevented them from spending tens of billions on data centers and Nvidia GPU's yet they didn't. Instead they outsourced the off-device LLM piece to OpenAI. I think that speaks volumes. I mention the arc because Apple's unveiling of the iPhone was a natural progression from what people had already demonstrated a preference for and were used to using. Steve Jobs himself said that the phone became popular not because it was inherently superior but because telegraphs required specialized knowledge to use so nobody owned them. The great leap forward came when phones were made easy enough to use that anyone could operate them. I don't see a natural progression from what a smartphone does to a headset.
Now Mark Zuckerberg seems hell-bent on creating this new platform but it is a near certainty he will fail simply because the iPhone is an extraordinarily high bar and people's preferences are well established. They love their iPhones and they love Instagram.
Hyperscalers are under enormous pressure from Wall Street to perform. If your largest competitor is investing tens of billions on this thing that everyone thinks will dwarf the internet, you'd better either follow suit or show that you're beating them without doing it.
So the whole thing becomes this positive feedback loop where Microsoft in an attempt to overtake Google and Apple goes all in on OpenAI. In response, Amazon, Oracle and Google respond by doing the same. Meta is kind of a special case because I truly believe Zuckerberg is still chasing the dream of creating a platform to destroy the iPhone. Then SaaS companies follow by showing that they're on the bleeding edge of this AI stuff and quickly incorporate it into their products and talk up its benefits ad nauseum.
To me its been self-evident for a long while that these tools just aren't that useful at scale. What will be the thing that pokes a hole in it is anybody's guess, it probably won't be one thing. It seems that cracks are already forming- just today SMCI got a Justice Department probe and the stock is down 10%. Nvida, Micron and Dell are already in large drawdowns. It seems likely this will be death by a thousand cuts as new details crack open this idea that AI is taking over. The end result will be hyperscaler stocks crashing along with semiconductors and there being a data center overhang for many years.