The Oracle Reset – Piper Sandler Pops the Cloud Computing Bubble

A small trading floor’s screens flicker with numbers moving more quickly than conversation on a gloomy morning in lower Manhattan. Leaning over keyboards, a few analysts browse through Piper Sandler’s most recent research note while coffee cups cool next to them.

The headline is direct. The cloud narrative seems to be stretched in some way. Cloud computing has been a given for over ten years. Businesses relocated their servers to far-off data centers. Subscriptions replaced installed software. The change was viewed by investors as nearly mechanical: as demand increases, profits follow.

However, Wall Street has been experiencing a subtle tension lately. And Oracle Corporation is the business that is directly in the middle of that conflict.

Over the past few years, Oracle has undergone an oddly dramatic transformation. The company has repositioned itself as a manufacturer of massive AI-focused cloud infrastructure, having previously been primarily recognized for its enterprise databases humming away in corporate server rooms. Packed with GPUs and networking hardware, data centers are growing in industrial zones and deserts.

The scale is hard to ignore when you stand outside one of these Texas facilities. Cooling fans hum on steel walls. Equipment crates the size of tiny apartments are carried by trucks as they pass gates. Thousands of processors work nonstop inside, processing data for machine learning systems. It’s amazing. Perhaps even intoxicating.

However, Piper Sandler’s analysts recently added a note of skepticism to the discussion, raising concerns about whether the industry’s haste to develop AI infrastructure may be outpacing real demand. The argument wasn’t wholly pessimistic. However, it was sufficient to cause investors to hesitate.

It’s important to pause.On paper, Oracle’s most recent earnings figures appeared spectacular. Revenue from cloud infrastructure increased by about 44%. Executives warned that for a while, customer demand may exceed supply due to the rapid increase in demand for AI computing capacity.

Normally, skeptics might be silenced by those numbers. However, the market’s response has been oddly inconsistent. The growth seems to excite some investors. Others appear uncomfortable, recalling a time when enthusiasm for technology surpassed reality.

I think of the late 1990s. The internet had the potential to completely transform the world economy back then. servers in data centers. Startups with very little revenue received a flood of venture capital. The reasoning appeared to be unstoppable for a while.

The dot-com bubble then burst. In today’s AI spending frenzy, it’s difficult to ignore remnants of that moment. Infrastructure is receiving hundreds of billions of dollars from tech giants. Data centers are being supported by the redesign of power plants. The demand for semiconductors appears to be virtually infinite.

Excitement and anxiety seem to be sitting side by side as this develops. Oracle has taken a particularly audacious approach. The company is reportedly investing over a billion dollars a week in capital expenditures to build AI-ready facilities. Executives contend that the scale is essential. Workloads involving artificial intelligence require enormous amounts of processing power. To be fair, the contracts do exist.

According to reports, Oracle’s backlog of contracted revenue, also referred to as remaining performance obligations, has surpassed $550 billion. Customers have already committed years of business to developing AI products and services, according to that figure. However, even such figures don’t completely dispel doubts.

It’s possible that, like the development of fiber networks or electrical grids, the cloud industry is just in the early stages of a massive new infrastructure cycle. In retrospect, Oracle’s aggressive spending might seem wise if that’s the case. However, markets hardly ever move in a straight line.

Some analysts are concerned that there could be severe repercussions if a small number of significant AI clients reduce spending—possibly because their own products take longer to monetize. The capacity of a data center may appear excessive at any time. Once-demanding investors might start posing more challenging queries.

And traders take notice when those queries begin to surface in research notes from companies such as Piper Sandler.

The story also has a deeper level. Oracle’s ascent in the cloud space has come as a surprise. For many years, the industry was dominated by firms like Microsoft and Amazon. Oracle was frequently observed showing up late for the celebration. All of a sudden, it is among the world’s most aggressive developers of AI infrastructure.

As one observes this change, one can’t help but wonder how quickly narratives in technology markets can change. Today’s AI powerhouse is yesterday’s laggard. The unstoppable boom of yesterday turns into the bubble debate of tomorrow. It’s possible that neither story is totally accurate.

For the time being, Oracle’s data centers are still growing, with cranes rising against expansive skies in the Middle East, Texas, and Nevada. While engineers install processor racks, investors monitor spreadsheets to determine whether the investment will be profitable.

The AI economy, according to some, is just getting started. Some believe the excitement might eventually subside. There’s a sense that both sides may be partially right. Revolutions in technology seldom go as planned. They arrive with long stretches of uncertainty interspersed with occasional corrections and bursts of optimism.

And somewhere within that ambiguity, the Oracle reset is subtly compelling the cloud sector to pose a straightforward but unsettling query. Not if AI will be important. However, the amount of money we are prepared to spend prior to the actual profits.