← All editions

TheWhiteBox · Premium edition

Could this be 2026’s big money move?

Ignacio de Gregorio Noblejas ·

Sign in or subscribe to keep reading ↓

In partnership with

THEWHITEBOX

Could this be 2026’s big money move?

Last week, we had one of those “the world is ending” days for NVIDIA, which fell by almost 7% at one point. And while March’s debacle (the DeepSeek crash) was vastly exaggerated, this one is warranted.

In pure market style, now people think Google is eating NVIDIA’s lunch with its Tensor Processing Units (TPUs), their equivalent to GPUs. Finally, the market has realized that Google is serious about entering the AI hardware business.

But I want to give you the hard data so you can decide for yourself how large the risk is for NVIDIA. We’ll discuss:

What makes TPUs different,

A very deep analysis of the performance per dollar for each option, including why NVIDIA is directly lying to you,

And we’ll cover what I believe are the most important market insights going into 2026, such as NVIDIA's “secret weapon” markets don’t have on the radar yet, AMD’s last opportunity to join the party, and a secret player that might rise to prominence unexpectedly.

Whether you’re an investor or just an AI enthusiast, here's what to watch out for in 2026 in the most profitable business in AI: chips.

Years Later… Finally A Rival

The reason for NVIDIA’s fall was a report by The Information saying that Google, the new darling of AI stocks, was heavily pushing to sell its TPUs to customers, specifically mentioning Meta and Anthropic, crucial buyers for both NVIDIA and Google, as early adopters.

We covered both “exclusives” here weeks ago, as this was already “known” among more specialized analysts before The Information disclosed it.

Rather, The Information simply made accessible ‘common knowledge’ what we analysts knew way before (as the Spanish saying goes, it was “un secreto a voces”, a “secret” that wasn’t really a secret and was being shouted for anyone who actually knows how to follow this stuff to hear, like seeing Anthropic opening a TPU Kernel Engineer, as we discussed a few weeks ago).

Also, the fact that Anthropic’s Head of Compute, James Bradbury, is an ex-Google expert on TPUs and Jax (the software equivalent to NVIDIA’s CUDA), isn’t particularly bullish to NVIDIA either.

Furthermore, TeraWulf announced a deal with Fluidstack, in which Google acted as a backstop in case Fluidstack couldn’t pay rent to Wulf… just as Fluidstack was announcing its agreement with Anthropic.

Source: TeraWulf

All this goes to prove that markets aren’t remotely aware of what is going on until mainstream media catches on. But let’s get into the weeds to describe why this matters and why it is happening.

A Move to Challenge NVIDIA

The narrative for NVIDIA bearishness is dead simple: NVIDIA was “untouchable”, and it isn’t anymore, thanks to Google’s chips, the TPUs that not only gave “birth” to the widely considered best model around, Gemini 3 Pro (this is a speculative claim at best, but it’s nonetheless “common knowledge”), but they are also (alongside Amazon’s own ASIC, Trainium), behind Claude 4.5 Opus’s training.

Enjoyed the introduction? Subscribe or sign in to continue with this edition.

TheWhiteBox · Subscriber access

Keep reading.

Already subscribed? Sign in with your newsletter email.

New here?Subscribe free ↗

Then return here to sign in and continue reading.