As of November 30, 2025, VR/AR has grown but is not yet a really important part of our existence for a large portion of the global population, in the way smartphones or PCs are.

Key evidence:

• Headset and user base size: Global VR/AR headset shipments remain relatively small compared to major computing platforms. Estimates put annual VR/AR device shipments in the tens of millions at most, versus well over a billion smartphones shipped per year and several billion active smartphone users worldwide.
• Meta’s Reality Labs scale vs. adoption: Meta has spent tens of billions of dollars on Reality Labs, but the unit has continued to post large operating losses and has not produced a mass‑adopted platform on the scale of mobile. Public filings and earnings coverage consistently frame VR/AR as a long‑term, early‑stage bet rather than a ubiquitous platform already integrated into daily life for most people.
• Usage patterns: VR usage is still concentrated in gaming, niche professional training, design, and some enterprise/industrial contexts. AR is most pervasive in lightweight forms (e.g., smartphone AR filters, basic try‑on apps), but these are typically features inside existing platforms, not a standalone, central computing environment akin to mobile or desktop operating systems.
• Lack of mainstream daily dependency: For the majority of people, daily essential tasks—communication, work, commerce, entertainment discovery—are still dominated by smartphones, laptops, and TVs. VR/AR devices are not yet required or standard for work or social life in the way that phones and PCs are.

Because the quote is framed as "we should assume that VR and AR is going to be a really important part of our existence" (i.e., a strong, near‑inevitability claim), the best evaluation as of late 2025 is that this has not yet materialized at population scale. It might still occur in the longer‑term future, but judging today, the prediction about VR/AR becoming a pervasive, major computing platform for a large share of humanity is not yet true, so the outcome is assessed as wrong (as of now) rather than inconclusive.