The trick: Self-Marked

Google says its new chip runs AI 3.5 times faster while using 3.5 times less energy.

The footnote says that was measured on pre-production phones, streaming YouTube.

Issue 915 August 20264 receipts3 min

Google's Pixel 11 announcement, verbatim: Packing 50% more TPU compute and paired with the latest Gemini Nano model, Google Tensor G6 processes on-device AI tasks up to 3.5 times faster while using up to 3.5 times less energy.

Before you read on. Your call?

the claim sentence carries footnotes 15 and 16. Footnote 15, verbatim, typo included: Compared to Google Tensor G5 at launch. Average based on internal data using pre-production devices steaming YouTube video over WiFi. So the baseline is last year's chip, the data is internal, the hardware is pre-production, and the measurement condition involves YouTube. Both multiples wear an up to, which means best case, not typical case. Independent tests of the G6's TPU performance or power draw published so far: zero. The most careful early analysis put it flatly: the figures are, for now, numbers Google itself has published.

3.5claimed on-device AI speed multiple AND claimed energy reduction multiple

50%claimed TPU compute increase vs Tensor G5

0independent

There’s more to this story.

Membership opens the full investigation, the strongest counterargument and what to do with what you’ve learned.

Start your free month →

First membership: 30 days free, then A$89 a year. One introductory trial per customer. Card required; renews annually until cancelled. Cancel before the trial ends to avoid the first charge. Already a member? Sign in

The trick has a name

We call it Self-Marked: graded by the party that benefits from the grade. You'll see it again. Learn to spot it →

Say this in tomorrow's meeting“Up to 3.5x, per Google's own internal data, on pre-production devices, versus last year's chip. The footnote is doing more work than the TPU.”

Receipts

Supports blog.google: Packing 50% more TPU compute and paired with the latest Gemini Nano model, Google Tensor G6 processes on-device AI tasks up to 3.5 times faster while using up to 3.5 times less energy
Supports thurrott.com: Packing 50% more TPU compute and paired with the latest Gemini Nano model, Google Tensor G6 processes on-device AI tasks up to 3.5 times faster while using up to 3.5 times less energy
Context techcrunch.com: The starting price for the new model is set at $899, reflecting a $100 increase compared to the Pixel 10.
Refutes android.gadgethacks.com: The 50% and 3.5x figures are, for now, numbers Google itself has published, worth taking seriously but not yet independently verified.

Open the Receipts Pack → What each source proves, every figure traced, and what would change our verdict.