If you thought the AI boom was reaching its peak, think again. According to industry experts, we're actually just getting started, and the next phase could be worth several trillion dollars. The catch? It's not about better chatbots—it's about Physical AI, and it turns out there's enough opportunity for multiple chip giants to prosper.

Physical AI: The Multi-Trillion Dollar Phase We Didn't See Coming

Daniel Newman, CEO of Futurum Group, recently shared analysis alongside data from Morgan Stanley that paints a compelling picture. The tech industry is transitioning away from the chatbot era and into something much bigger: Physical AI. We're talking wearables, autonomous devices, humanoid robots—the kind of stuff that actually moves around in the real world.

This isn't just incremental growth. The Total Addressable Market (TAM) for Physical AI is being pegged at multiple trillions of dollars, and Newman argues we're "so damn early" in the cycle. Why? Current datacenter capacity isn't even close to what's needed to support the wave of enterprise and physical generative workloads heading our way. Translation: sustained demand for hardware infrastructure stretching well into the 2030s.

So much for the AI bubble concerns.

The Market Is Expanding, Not Playing Musical Chairs

Here's where it gets interesting. There's been chatter that the rise of custom silicon—think Alphabet Inc. (GOOG) (GOOGL) TPUs or Broadcom Inc. (AVGO) ASICs—means bad news for Nvidia Corp. (NVDA). Tech analyst Beth Kindig of I/O Fund isn't buying it.

Kindig's take? The AI monetization wave is going to "widen," creating room for new players to capture revenue without necessarily taking market share from Nvidia. "We are not at a static point," she noted, emphasizing that while Broadcom might grab share in the inference market, Nvidia still holds a "near monopoly" on training thanks to its robust CUDA software ecosystem and relentless research and development pace.

In other words, this isn't a zero-sum game. The pie is getting bigger, and there are enough slices to go around.

Capital Spending Reaches an Inflection Point

Want to understand how serious companies are about this opportunity? Look at the money they're throwing at it. Kindig pointed to Oracle Corp. (ORCL) as a prime example. The company's 2026 capital expenditure estimates have more than doubled—from $9 billion to over $20 billion.

Some analysts have compared this spending spree to "drunken sailor" behavior, but tech CEOs are betting on a massive "inflection point" of monetization that will justify these investments. And they might be onto something. Companies like OpenAI have hit a $20 billion run rate faster than any company in history, suggesting the market is ready to absorb and validate these enormous hardware buildouts.

How Nvidia Is Performing in 2025

Speaking of hardware investments, Nvidia (NVDA) shares have climbed 40.16% year-to-date and 19.32% over the last six months. The stock closed 1.21% lower at $188.22 on Monday, but the overall trend remains strong.

Nvidia maintains robust price momentum across short, medium, and long-term timeframes, though its value ranking is less impressive—which makes sense given the premium investors are willing to pay for exposure to the AI infrastructure buildout.

The bottom line? Physical AI represents a genuine paradigm shift, not just hype repackaged. The opportunity is massive enough that both general-purpose GPU leaders like Nvidia and custom silicon specialists like Broadcom can carve out profitable niches. Current datacenter capacity can't handle what's coming, corporate spending is accelerating dramatically, and we're still in the early innings of what could be a decade-long infrastructure buildout. That's the kind of setup that creates multiple winners, not just one.