As of 30 November 2025, the conditions and time horizon of Chamath’s prediction have not clearly played out yet.

1. What the prediction requires The normalized prediction is conditional and medium‑term:

• If hyperscalers keep spending tens of billions per quarter on AI infra without commensurate new revenue,
• then, within the next few years, Nvidia’s valuation will prove unjustified and both Nvidia and the big spenders will suffer significant share‑price punishment.

The key pieces to check are: (a) continued huge AI capex, (b) lack of matching revenue/earnings, and (c) a major, sustained drawdown in Nvidia and hyperscaler stocks within a few years of May 2024. We’re only ~18 months into that window.

2. What has actually happened so far

Spending:

• Google/Alphabet has sharply ramped technical‑infrastructure and AI spending, guiding ~$85–90B of annual capex, mostly for servers, data centers and networking to meet AI cloud demand.【0search7】【0news16】
• Amazon plans around $100B of AWS capex in 2025, with quarterly capex in the low‑$20Bs to $30B+ range, the “vast majority” on AI infrastructure.【0search3】【0search8】
• Microsoft’s capex has risen into the tens of billions per quarter as well, largely to support Azure and AI workloads.【2news13】

So the “tens of billions per quarter” spending part is clearly happening.

Revenues/earnings:

• AWS became a >$100B‑per‑year business in 2024 with ~$107.6B in sales and ~$39.8B operating income, 19% YoY growth, as Amazon simultaneously prepares ~$100B of AI‑heavy capex in 2025.【0search5】【0search3】
• Alphabet’s Google Cloud revenue grew 32% YoY to $13.6B in a recent quarter, with a $106B backlog, and management explicitly ties its elevated capex to strong AI and cloud demand.【0search7】
• Microsoft reports that Azure surpassed $75B in annual revenue, growing in the low‑to‑mid‑30% range, and total Microsoft Cloud revenue is growing ~27% YoY, framed by the company itself as driven by cloud and AI.【0search0】【0search2】

In other words, substantial incremental revenue and profit are already showing up, even though investors still argue about whether these returns are fully “commensurate” with the capex.

Stock‑price behavior: Starting near Chamath’s podcast date (31 May 2024) and looking through late 2025:

• Nvidia (NVDA) closed at about $109.6 on 31 May 2024; recent prices are around $177–180, a gain of roughly 60%+.【1search0】【0finance0】 Nvidia even briefly reached a $4T market cap in July 2025, becoming the world’s most valuable company due to AI chip demand.【3search0】【3search1】
• Microsoft (MSFT) rose from about $411 at 31 May 2024 to roughly $490–520 in late 2025.【2search0】【0finance1】
• Alphabet (GOOGL) went from about $171 to around $280–320, and its market cap is now approaching $4T, with the stock up ~70% in 2025 alone on AI optimism.【1search1】【0finance2】【3news12】
• Amazon (AMZN) climbed from about $176 to the low‑$230s.【1search3】【0finance3】
• Meta (META) moved from about $464 to the mid‑$600s.【2search4】【0finance4】

There have been sizable pullbacks and sector wobbliness in late 2025 (e.g., Nvidia giving back hundreds of billions from a peak above $5T, short‑term 10–15% drops on AI‑ROI fears),【0news14】【3news13】 but from the May 2024 baseline these names are still dramatically higher, not “punished.”

3. Why this is not yet a clean “right” or “wrong” call

• The timeframe Chamath pointed to was “within the next few years” from mid‑2024. As of late 2025, that window is only partially elapsed; markets could still reassess the ROI of AI capex and re‑rate these stocks lower in 2026–27.
• The prediction is explicitly conditional on hyperscalers not showing incremental revenue/earnings from their AI spending. So far, they are in fact reporting large and growing cloud/AI revenue and profits; whether that is “commensurate” is partly subjective, and the condition has arguably not been met yet.
• Because the antecedent (“no incremental revenue”) is not clearly satisfied, we can’t empirically test the “then they will be punished” part. At the same time, the observed outcome so far (Nvidia + hyperscalers up sharply) runs opposite to the spirit of his warning, but that could still change before his vague “few years” horizon fully expires.

Given this, the forecast cannot yet be definitively scored as right or wrong. The AI capex boom and Nvidia’s valuation have so far been rewarded, not punished, but the specified time window and the conditional nature of his statement mean the final verdict is still open.

Conclusion: it’s too early to declare a definitive outcome, so the appropriate rating is **“inconclusive (too early).”