AI Boom or AI Bubble? Why Software Is Being Unfairly Punished

Now everyone’s asking the same question: why did software stocks get hammered so badly yesterday, even when the broader market was rebounding? And why is money flowing into hardware but avoiding software?

My take is simple — the market is reacting to fear.

xAI just dropped a very powerful model, and suddenly it’s like Thanos snapped his fingers on software stocks. The narrative becomes: if AI can do everything, what’s the point of traditional software?

So people sell first, think later.

This fear isn’t new. It’s always been there. The recent optimism was just a pause — now we’re back to doubting software again.

But let’s be clear: this selloff is not rational.

Look at the data:

$iShares Expanded Tech-Software Sector ETF(IGV)$  

Software ETF IGV: ~-5% intraday, worst drop since 2023

SaaS index: ~-40% YTD

Meanwhile:

Earnings: still fine

Profit forecasts: still going up

2027 growth expectations: revised higher

But valuations collapsed to ~20x vs ~34x historical.

So fundamentals up, prices down.

That’s sentiment, not reality.

Now here’s where it gets interesting — and where most people are missing the bigger picture.

Everyone is chasing “AI infrastructure winners”.

$NVIDIA(NVDA)$  

Names like NVIDIA and newer players like CoreWeave are being treated as if their growth is guaranteed.

$CoreWeave, Inc.(CRWV)$  

Let’s talk about CoreWeave for a moment.

On the surface, it looks like the perfect AI story:

Revenue growth: +400%+

Backlog: tens of billions

Massive deals with hyperscalers

Pure-play GPU cloud

Even recently, it secured multi-billion dollar deals with companies like Meta to supply AI compute capacity

So the market says: “This is the future.”

But dig one level deeper.

CoreWeave is:

Highly leveraged

Burning cash

Spending $30–35B capex to keep up

Still loss-making despite explosive growth

It’s basically a capital-intensive infrastructure bet.

Even worse:

Heavy customer concentration (e.g. Microsoft historically dominant)

Growth depends on a few giant buyers continuing to spend

So the real question is:

👉 If downstream doesn’t make money, who pays CoreWeave?

$Oracle(ORCL)$  

Now compare that with Oracle.

Market says:

Oracle = “old software”

CoreWeave = “AI infrastructure future”

But reality is almost the opposite.

Oracle today:

Cloud infrastructure already >50% of revenue

Profitable

Strong cash flow

RPO (backlog) in the hundreds of billions

CoreWeave:

Smaller scale

Negative earnings

Massive debt + capex

Backlog looks big, but comes with execution risk

Even CoreWeave’s backlog (~$25–60B range depending on period) is growing fast, but still far smaller and riskier compared to Oracle’s contracted visibility

Here’s the irony:

The market is pricing:

CoreWeave → like a premium infrastructure winner

Oracle → like a declining software company

That’s a mismatch.

Because structurally:

CoreWeave = high growth, high risk, high capital intensity

Oracle = moderate growth, high profitability, strong visibility

Yet valuations don’t reflect that difference properly.

$Microsoft(MSFT)$  

Same story with Microsoft.

People are ignoring the fact that:

It owns both infrastructure (Azure) AND downstream applications

It monetizes AI directly through software

While CoreWeave depends on others to monetize AI.

Now step back and think bigger.

The AI value chain must eventually balance:

Upstream (chips, infra)

Midstream (cloud)

Downstream (software, apps, users)

Today:

Upstream margins = very high

Downstream = being questioned

But this cannot last.

Because if software doesn’t make money:

👉 Infrastructure demand collapses

👉 And companies like CoreWeave feel it first

So what we’re seeing now is a classic early-cycle distortion:

Hardware = over-earning

Software = over-feared

But long term, both must win.

My conclusion:

Software stocks are being oversold, while some infrastructure names are being over-assumed.

CoreWeave is a great company, but it also highlights the risk:

👉 Growth without profitability

👉 Demand dependent on a few big buyers

👉 Heavy capital requirements

Meanwhile, companies like Microsoft and Oracle:

👉 Already profitable

👉 Already embedded in the AI stack

👉 Already monetizing

So if you’re holding software names, this is probably not the time to panic.

If anything, this is where patience matters most.

Because in every cycle:

👉 The market chases what looks strongest today

👉 And misprices what will matter tomorrow

Just remember:

Don’t go all-in

Size properly

Build gradually

The real money is made when perception and reality reconnect.

@TigerObserver  @TigerPM  @Tiger_comments  @TigerStars  @Daily_Discussion