New York, New York
— November 25, 2025

Google’s TPU Push Is Shaking Up the AI Chip Race: Here’s Why It Matters

If you went to business school, you might remember Porter’s Five Forces, especially threat of new entrants and threat of substitutes. Those two forces are becoming increasingly relevant in the AI semiconductor world, where Nvidia has dominated for years.

But now, Google is stepping onto the field in a very real way.

TPUs: Google’s Purpose-Built AI Chip

A TPU (tensor processing unit) is a custom semiconductor designed specifically for AI workloads. Unlike GPUs, TPUs offer:

• Higher energy efficiency
• Purpose-built performance for AI
• Lower operating costs
• Tighter integration with cloud infrastructure

The tradeoff: GPUs are more flexible and support more types of workloads, but they consume significantly more power a huge issue as AI compute is draining the grid.

Why This Matters Now

Over the last several weeks, Google’s TPU momentum has been hitting the headlines. Given Googles full stack approach to AI its making waves in the markets. Specific to Google’s TPU business:

• Anthropic wants access to up to 1 million TPUs
• Meta is reportedly looking to acquire or rent TPUs
• Google Cloud is aggressively expanding TPU capacity
• CEO Sundar Pichai highlighted “tremendous demand” on the earnings call

The stock reaction reflects this strength: Google is up 80%+ in six months, outperforming nearly all large-cap tech.

This is classic Porter’s Five Forces in real time: a substitute product gaining ground in a high-profit industry dominated by a single incumbent.

Meanwhile, Nvidia…

Nvidia is still the undisputed leader with ~90% market share in AI GPUs but its stock is now down around 5% in the 3-month chart.

Not because demand is slowing.
Not because the tech is failing.
But because investors are starting to price in alternative compute pathways, including:

• AMD’s MI450
• Google’s TPUs
• (Soon) Custom silicon from Meta, OpenAI, and others

That said, this isn’t a panic moment for Nvidia. Even if they surrender some market share, their pipeline, demand curve, and execution remain exceptional.

What Investors Should Really Focus On

The real risks in the AI trade aren’t about TPUs vs GPUs.

They’re the macro risks that hit everyone:

• An AI CapEx bubble (spend growing faster than revenue)
• Debt overhang from massive data center construction
• Compute hoarding creating temporary imbalance
• Questions around AI profitability at scale

While Google’s TPU surge is meaningful, it’s not the existential threat.
The bigger picture is whether the entire AI ecosystem can convert massive investment into sustainable returns.

WSGN Takeaway

Google’s momentum is real. TPUs are legit. Nvidia will still grow, but we expect the competitive landscape to get more dynamic in 2025 and beyond.

Long-term winners? Likely several, not just one.

We’ll keep tracking this trend and breaking it down for you as the AI chip race evolves.
As always. Be good.