Been noticing something pretty paradoxical happening in the markets lately, and honestly it's got me scratching my head. AI stocks are getting absolutely hammered at the start of 2026 because everyone's worried about whether all those massive capex investments are actually going to pay off. Meanwhile, software stocks are getting crushed too, but for the exact opposite reason - people think AI is going to destroy the entire SaaS industry. How can both things be true at the same time? That's the paradoxical doomsday we're looking at right now.

Let me break down what's actually happening here. The big tech companies have been throwing hundreds of billions at AI infrastructure and development. But investors are starting to ask the tough questions: are we actually getting better returns on this spending? The latest ChatGPT models have been facing criticism, which has people wondering if we're just burning cash at this point. Throw in the fact that valuations were already stretched, and you've got a recipe for a sell-off.

But here's where it gets paradoxical. At the same time everyone's worried AI won't deliver, Anthropic just released Claude Cowork - an AI agent tool that can literally connect to your files and handle all kinds of tasks on your computer without coding. When people saw that, the market immediately thought about how many software tools and SaaS products could become obsolete. So now software stocks are collapsing because traders think AI is about to disrupt the entire sector.

The resource question is real too. Data centers powering AI consume massive amounts of electricity and water. A report from Lawrence Berkeley National Laboratory last year estimated that by 2028, over half of all data center power consumption will be dedicated to AI - we're talking about electricity equivalent to 22% of all U.S. households. McKinsey is projecting we'll need $6.7 trillion in data center spending by 2030 just to keep up with demand. That's a staggering amount of capital, and the market's rightfully asking whether the returns justify it.

So you've got this paradoxical situation where AI investors are panicking about whether the technology will actually deliver meaningful returns, while software investors are panicking that it will deliver too well and destroy their business models. Bank of America's Vivek Arya actually made a solid point about this - he said both things can't really happen at once. The way he framed it, AI models are providing unprecedented intelligence, but actually turning that intelligence into profitable products is going to take years. That's an important distinction that gets lost in the panic.

Here's what I think is actually going on beneath the surface. This isn't really a doomsday scenario - it's more of a sector reset. Yeah, AI is definitely going to disrupt a lot of software companies. Some of them won't make it through the transition. But plenty of software companies are already partnering with major AI players and figuring out how to integrate this technology into their offerings. They're going to be fine, maybe even better positioned than before.

The real issue is that we're probably seeing the end of an era where unprofitable SaaS companies could trade at 15 or 30 times revenue just because they had fancy technology. Those days are done. AI is going to make it faster and cheaper to build software solutions, which means the moats that protected some of these companies are going to erode. Margins are going to compress. The sector's going to have to repriced, and that's painful in the short term.

What's interesting is that this paradoxical tension between AI disruption concerns and AI investment concerns probably resolves itself over the next few years. Eventually, AI and software aren't going to be separate categories anymore - they're going to converge until you can't really distinguish between them. But transitions like that are messy. They create winners and losers. They create volatility. And right now, everyone's selling first and asking questions later.

I do think some of this selling has gotten overdone. There will absolutely be software companies that figure out how to thrive in an AI-powered world. But there will also be plenty that don't make it. The real opportunity isn't in trying to pick which software company survives - it's in understanding that this rerating of the sector is probably going to take a while to play out. This paradoxical moment we're in isn't the end of the story, it's just the messy middle part where everyone's trying to figure out what the new normal looks like.