24 Oct 2026
Baseline snapshot
As of the 23 Oct 2026 close, semis and copper proxies are positive YTD, while the power, nuclear and uranium proxies are negative YTD. Watch whether this divergence persists. Personal research, not advice.
← Research themes · Last updated 30 Oct 2026
Assets to hold on the way to AGI and after
Real performance data for each asset class, with a two-phase hypothesis for each.
Working thesisThrough the run-up to AGI, scarce physical inputs (compute, power, copper) and the platforms that deploy AI should outperform. After AGI, value shifts to what stays scarce (land, energy, hard assets), while long-duration bonds suffer from higher real rates.
Personal research, not financial advice. Past performance does not predict future returns. The tickers are proxies for asset classes, not recommendations. Phase theses are hypotheses with cited reasoning, not forecasts.
Asset classes: performance and two-phase thesis
25 rows
| Class | Proxy | Ticker | YTD | 1y | 5y | Last close | Run-up to AGI | Post-AGI |
|---|---|---|---|---|---|---|---|---|
| Compute / semis | Nvidia | NVDA | +9.2% | +0.4% | +586.1% | 30 Oct 2026 | Direct beneficiaries of the capex boom (Theme 10). The risk is a capex pause or digestion phase. | If AI makes chip design and fabrication cheaper, margins could compress; owners of scarce fabs (TSMC) and tools (ASML) may hold up better than designers. |
| Compute / semis | TSMC (ADR) | TSM | +34.2% | +34.9% | +272.2% | 30 Oct 2026 | Direct beneficiaries of the capex boom (Theme 10). The risk is a capex pause or digestion phase. | If AI makes chip design and fabrication cheaper, margins could compress; owners of scarce fabs (TSMC) and tools (ASML) may hold up better than designers. |
| Compute / semis | ASML | ASML | +46.7% | +45.9% | +92.8% | 30 Oct 2026 | Direct beneficiaries of the capex boom (Theme 10). The risk is a capex pause or digestion phase. | If AI makes chip design and fabrication cheaper, margins could compress; owners of scarce fabs (TSMC) and tools (ASML) may hold up better than designers. |
| Hyperscalers | Microsoft | MSFT | +8.5% | -0% | +61.6% | 30 Oct 2026 | They own distribution and cloud. Heavy capex lowers free cash flow (Theme 10). | Could capture AGI rents, or see them competed away if models commoditise. |
| Hyperscalers | Alphabet | GOOGL | -10.7% | -0.6% | +89.3% | 30 Oct 2026 | They own distribution and cloud. Heavy capex lowers free cash flow (Theme 10). | Could capture AGI rents, or see them competed away if models commoditise. |
| Hyperscalers | Amazon | AMZN | +3.3% | +7% | +35.5% | 30 Oct 2026 | They own distribution and cloud. Heavy capex lowers free cash flow (Theme 10). | Could capture AGI rents, or see them competed away if models commoditise. |
| Hyperscalers | Meta | META | -5.6% | -6.4% | +84% | 30 Oct 2026 | They own distribution and cloud. Heavy capex lowers free cash flow (Theme 10). | Could capture AGI rents, or see them competed away if models commoditise. |
| Energy / utilities / nuclear | Vistra (power) | VST | -14.8% | -27.5% | +627% | 30 Oct 2026 | Data-centre power demand is projected to rise sharply (Goldman Sachs; IEA). | Physical power stays scarce longer than software does, but regulated returns cap the upside. |
| Energy / utilities / nuclear | Constellation (nuclear) | CEG | -25.8% | -31.3% | — | 30 Oct 2026 | Data-centre power demand is projected to rise sharply (Goldman Sachs; IEA). | Physical power stays scarce longer than software does, but regulated returns cap the upside. |
| Uranium | Cameco (uranium) | CCJ | -17.8% | -28% | +182.2% | 30 Oct 2026 | More nuclear power for data centres means more demand for uranium (World Nuclear Association). | A long-lived resource, but depends on reactor buildout timelines. |
| Uranium | Global X Uranium ETF | URA | -21.9% | -37.7% | +39.8% | 30 Oct 2026 | More nuclear power for data centres means more demand for uranium (World Nuclear Association). | A long-lived resource, but depends on reactor buildout timelines. |
| Energy / utilities / nuclear | US Utilities ETF | XLU | +4.1% | -0.2% | +50.6% | 30 Oct 2026 | Data-centre power demand is projected to rise sharply (Goldman Sachs; IEA). | Physical power stays scarce longer than software does, but regulated returns cap the upside. |
| Data-centre REITs | Equinix (DC REIT) | EQIX | +34.4% | +24.3% | +43.2% | 30 Oct 2026 | Leasing demand from hyperscalers. | Exposed if hyperscalers own more capacity themselves, or if efficiency gains cut the space needed. |
| Data-centre REITs | Digital Realty (DC REIT) | DLR | +17.4% | +7.2% | +35.7% | 30 Oct 2026 | Leasing demand from hyperscalers. | Exposed if hyperscalers own more capacity themselves, or if efficiency gains cut the space needed. |
| Copper / critical minerals | Freeport-McMoRan (copper) | FCX | +24.4% | +51.4% | +81% | 30 Oct 2026 | Grid and data-centre buildout increases demand (IEA minerals outlook). | Scarce physical inputs may keep value if AI lowers the cost of other goods. |
| Copper / critical minerals | Copper Miners ETF | COPX | +0.6% | +19% | +124.2% | 30 Oct 2026 | Grid and data-centre buildout increases demand (IEA minerals outlook). | Scarce physical inputs may keep value if AI lowers the cost of other goods. |
| Gold | SPDR Gold | GLD | -10% | -3.6% | +110.1% | 30 Oct 2026 | A hedge against fiscal stress and geopolitical risk; central-bank buying (WGC Gold Demand Trends). | A store of value that does not depend on any AI outcome. |
| Bitcoin | Bitcoin | BTC-USD | -11.8% | -28.7% | +26.3% | 30 Oct 2026 | Speculative; correlated with risk appetite. | A fixed-supply digital asset, a possible hedge against debasement (Theme 1), but highly volatile. |
| Privacy coins | Zcash (ZEC) | ZEC-USD | +141.1% | +257.5% | +657.5% | 30 Oct 2026 | Working thesis: cheap AI makes chain analytics and payment surveillance close to free, so financial privacy becomes scarcer and more valuable. Zcash offers optional shielded transfers built on zero-knowledge proofs (z.cash). | Under fiscal stress (Theme 1) and heavier financial surveillance, demand for private, self-custodied money could rise. Risks: regulation (EU AMLR Art. 79, applies from 10 Jul 2027 bars EU-regulated crypto providers from handling anonymity-enhancing coins), exchange delistings, low liquidity and extreme volatility. |
| Privacy coins | Monero (XMR) | XMR-USD | +37.1% | +83.9% | +131.6% | 30 Oct 2026 | Working thesis: same privacy-scarcity case as ZEC. Monero is private by default: sender, receiver and amount are hidden (getmonero.org). | A candidate store of value for financial privacy after AGI, if it survives regulation. Risks: delistings (Binance delisted XMR in Feb 2024), the EU AMLR, thin liquidity, high volatility and possible future cryptographic weaknesses. |
| Broad equity | S&P 500 ETF | SPY | +11.7% | +12.3% | +72.4% | 30 Oct 2026 | Broad exposure to AI productivity gains. | If AGI raises growth, real interest rates should rise (Chow, Halperin & Mazlish), which hurts valuations; winners could be concentrated. |
| Broad equity | MSCI All-Country World ETF | ACWI | +11% | +12.2% | +59% | 30 Oct 2026 | Broad exposure to AI productivity gains. | If AGI raises growth, real interest rates should rise (Chow, Halperin & Mazlish), which hurts valuations; winners could be concentrated. |
| Bonds / cash | 20+yr US Treasuries ETF | TLT | -3.2% | -5.8% | -34.8% | 30 Oct 2026 | Long bonds lose if AGI expectations raise real rates (Chow, Halperin & Mazlish). Cash earns the policy rate. | Debt sustainability (Theme 1) and inflation risk. Short-dated cash is the liquidity reserve. |
| Bonds / cash | 1-3m T-bills ETF (cash proxy) | BIL | +3% | +3.7% | +20.1% | 30 Oct 2026 | Long bonds lose if AGI expectations raise real rates (Chow, Halperin & Mazlish). Cash earns the policy rate. | Debt sustainability (Theme 1) and inflation risk. Short-dated cash is the liquidity reserve. |
| Land / real estate | US Real Estate ETF | VNQ | +9.5% | +9.9% | +4.1% | 30 Oct 2026 | Rate-sensitive. AI hubs could see local price gains. | Land is fixed in supply; if AI makes labour and goods cheap, scarce land could gain relative value. This is speculative. |
Returns are computed from adjusted closes; 5y is blank where less than five years of history exists (CEG listed in 2022).Source: Yahoo Finance chart data (adjusted close) · as of 30 Oct 2026 · fetched 30 Oct 2026 · Download CSV (all columns)
Privacy infrastructure
The working thesis is that if AI makes surveillance cheap, the tools that keep data and money private become infrastructure. There is no clean, citable listed or tokenised pure play, so this is tracked qualitatively rather than with a price row.
End-to-end encryption
Messages that only the sender and recipient can read. The Signal Protocol is a widely used design (Signal docs). The pressure point is governments: Apple withdrew Advanced Data Protection in the UK after a government demand (BBC, Feb 2025; Apple).
Zero-knowledge proofs
These let you prove a statement (for example "I am over 18" or "this payment is valid") without revealing the underlying data (ethereum.org). They are used in Zcash shielded transfers and Ethereum rollups.
Self-custody
You hold your own keys instead of keeping assets with an exchange. The EU AMLR restricts regulated providers, not self-custody or peer-to-peer transfers (analysis), so self-custody is where privacy assets are likely to end up held.
What would change the view
Supporting: more AML and surveillance mandates, capital controls, and privacy features shipping in mainstream apps. Weakening: an outright ban on holding privacy coins in major jurisdictions, a cryptographic break, or a collapse in privacy-coin liquidity after the July 2027 EU deadline.
Data pendingPrivate AI labs and tokenised stocks
There is no reliable public price series for private AI equity (OpenAI, Anthropic, xAI). Tokenised-stock products are new and their legal status is evolving (SEC Commissioner Peirce, Jul 2025). Tracked qualitatively until a fetchable index exists.
What to watch
- Hyperscaler capex guidance on earnings calls (Theme 10).
- Real 10-year yields (the Chow/Halperin/Mazlish signal).
- Data-centre power contracts and nuclear restarts.
- Central-bank gold purchases.
Reading
- Transformative AI, existential risk, and real interest rates— Chow, Halperin & Mazlish, 2024Markets are not pricing AGI if real rates stay low.
- AI's $600B Question— Sequoia Capital (David Cahn), 2024-06The revenue gap facing AI infrastructure.
- AI to drive 165% increase in data center power demand by 2030— Goldman Sachs, 2025The power-demand thesis.
- Gold Demand Trends— World Gold Council, 2026Quarterly gold flows.
- Scenarios for the Transition to AGI— NBER (Korinek & Suh), 2024Wages and output under different AGI timelines.
Five-year paths
Growth of $100 over 5 years (weekly)
US$ (start = 100)
Growth of $100 over 5 years (weekly)
US$ (start = 100)
Growth of $100 over 5 years (weekly)
US$ (start = 100)
Privacy coins vs Bitcoin and gold: growth of $100 over 5 years (weekly)
US$ (start = 100)
Privacy coins against Bitcoin and gold. This is a log scale, and privacy coins are extremely volatile.
Source: Yahoo Finance chart data (adjusted close), rebased by this site [2] [3] [4]
· Data as of 26 Oct 2026 · fetched 30 Oct 2026
Global copper price
US$ per tonne
Source: FRED, Federal Reserve Bank of St. Louis (PCOPPUSDM)
· Data as of 1 Sep 2026 · fetched 30 Oct 2026
What to watch
- Divergence between AI semis and the S&P 500.
- Copper price versus copper miners.
Reading
- Mineral Commodity Summaries— USGS, 2026Supply data for copper and critical minerals.
- World Uranium Mining Production— World Nuclear Association, 2025Uranium supply.
- Energy and AI— IEA, 2025-04Electricity demand from AI.
Signals
Supporting the run-up thesis
- Hyperscaler capex keeps rising (Theme 10).
- Physical-constraint news: grid connection queues, transformer shortages, copper deficits.
Weakening it
- Capex cuts or "digestion" guidance.
- Model commoditisation (cheap open models) that erodes platform margins.
- Real rates stay low, implying markets do not expect transformative AI.
What to watch
- The YTD figures in the table: energy and uranium proxies are negative YTD while semis and copper are positive. Note that divergence in the journal.
Reading
- Situational Awareness— Aschenbrenner, 2024-06The bull case for the compute and power buildout.
- AI Index Report 2026— Stanford HAI, 2026Investment and adoption data.
Prediction markets
Live prices from public prediction markets, snapshotted by scripts/update_markets.py. A price of 30% means traders pay about 30¢ for a contract that pays $1 if the event happens; treat it as a crowd estimate, not a fact. Polymarket and Kalshi are real-money markets (Kalshi is CFTC-regulated); Manifold uses play money and is often thin. Long-dated and low-volume markets are poorly calibrated. Each market's exact resolution rules on the platform decide what counts, so read them before relying on a headline question.
Bitcoin and gold
49%
$1.4M vol · $67k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
32%
$2.7M vol · $63k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
28%
$758k vol · $62k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
18%
$1.1M vol · $44k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
18%
$1.5M vol · $62k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
13%
$4.0M vol · $94k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
>99%
$921k vol · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
15% bid 14¢ / ask 15¢
1.7M contracts · 418k open interest · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
6% bid 5¢ / ask 6¢
739k contracts · 119k open interest · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
3% bid 3¢ / ask 4¢
658k contracts · 161k open interest · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
2% bid 1¢ / ask 3¢
509k contracts · 122k open interest · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
2% bid 1¢ / ask 2¢
352k contracts · 95k open interest · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
Polymarket Will Bitcoin outperform Gold in 2026?
40%
$481k vol · $19k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
AI equities
53%
$1.6M vol · $84k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
44%
$1.4M vol · $99k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
1%
$902k vol · $153k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
1%
$1.2M vol · $77k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
Polymarket Will Nvidia’s market capitalization be between $5.00T and $5.50T at the end of 2026? · $5.00-$5.50T
28%
$8k vol · $7k liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
Polymarket Will Nvidia’s market capitalization be between $5.50T and $6.00T at the end of 2026? · $5.50-$6.00T
22%
$11k vol · $8k liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
Polymarket Will Nvidia’s market capitalization be between $4.50T and $5.00T at the end of 2026? · $4.50-$5.00T
21%
$10k vol · $3k liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
7%
$6k vol · $6k liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
9%
$2.5M vol · $26k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
Polymarket AI bubble burst by June 30, 2027?
23%
$44k vol · $13k liquidity · closes 2027-07-01 · as of 2026-10-30 17:26 UTC
36%
$8k vol · $322 liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
27%
$5k vol · $381 liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
16%
$27k vol · $2k liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
13%
$4k vol · $610 liquidity · closes 2026-12-31 · as of 2026-10-30 17:26 UTC
Private AI labs
4%
$1.1M vol · $41k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
21%
$7k vol · $13k liquidity · closes 2027-04-01 · as of 2026-10-30 17:26 UTC
Polymarket Will OpenAI IPO by June 30, 2027?
45%
$11k vol · $5k liquidity · closes 2027-07-01 · as of 2026-10-30 17:26 UTC
0%
$151k vol · closes 2026-10-01 · as of 2026-10-30 17:26 UTC
Polymarket Will Anthropic IPO by October 31, 2026?
<1%
$1.5M vol · $125k liquidity · closes 2026-11-01 · as of 2026-10-30 17:26 UTC
Polymarket Will Anthropic IPO by November 15, 2026?
2%
$325k vol · $11k liquidity · closes 2026-11-16 · as of 2026-10-30 17:26 UTC
Polymarket Will Anthropic IPO by November 16, 2026?
8%
$7k vol · $9k liquidity · closes 2026-11-17 · as of 2026-10-30 17:26 UTC
0%
$1.1M vol · closes 2026-10-01 · as of 2026-10-30 17:26 UTC
Privacy coins (ZEC, XMR)
25%
$67k vol · $1k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
21%
$28k vol · $1k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
10%
$9k vol · $1k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
8%
$2k vol · $824 liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
5%
$5k vol · $3k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
>99%
$5k vol · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
Polymarket Will Monero hit $1000 in 2026?
14%
$127k vol · $3k liquidity · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
24%
Ṁ9k play-money · 17 traders · closes 2029-12-31 · as of 2026-10-30 17:26 UTC
8%
Ṁ312 play-money · 12 traders · closes 2027-01-01 · as of 2026-10-30 17:26 UTC
Snapshot fetched 2026-10-30 17:26 UTC from public APIs (Polymarket Gamma/CLOB, Kalshi trade API, Manifold v0). Prices are what traders are paying, not my forecasts and not advice; thin or long-dated markets can be far off. Sparklines show daily prices over up to 120 days. Metaculus: API now requires an authenticated account token (HTTP 403 without one) and the website is behind a bot challenge, so Metaculus forecasts are not included. Add a ref type with a personal token later if wanted.
How to add a market
Add a ref to "markets" → "groups" → "refs" in src/themes/<slug>.json: {"platform":"polymarket","event":"<slug after /event/ in the URL>"}, {"platform":"kalshi","event":"<EVENT_TICKER>"} or {"platform":"manifold","id":"<market id or slug>"}. Optional: "markets" (list of specific sub-market slugs or tickers), "max", "sort" (date, prob or api). Then run python3 scripts/update_markets.py <slug> and rebuild.
Notes & journal
Dated observations from src/notes/agi-assets.md. Add a new entry with a ## YYYY-MM-DD Title heading.