Gold hit $3,680 this week and keeps ripping to new all time highs every other day.
If you've been watching, this isn't your typical precious metals chop.
Gold is up 8% in the past month alone, moving almost monotonically up and to the right with zero pullbacks.
For most investors, this creates a classic FOMO dilemma.
Do you chase a parabolic move? Is this the start of a multi-year bull run? Or are you buying the top from whoever's been accumulating for months?
The challenge with gold is there are too many moving parts to track manually.
Central bank buying, geopolitical tensions, dollar weakness, real rates, inflation expectations, ETF flows, jewelry demand ... each driver pulls in different directions and changes daily.
I wanted to use ChatGPT Deep Research to cut through the noise and figure out what's actually driving this move.
And ultimately determine how much room left do we have to run?
Here's my framework for trading assets without cash flows like gold and crypto: find the marginal buyer. Who's pushing prices higher right now? How much more can they buy? When do they tap out?
I structured my Deep Research to investigate:
- How gold is moving vs other asset benchmarks
- Key drivers and who's buying
- Risk scenarios that could derail the move
- What to watch next
- How this fits into broader macro trends around rates and central banks
This type of global macro analysis used to take days of research across dozens of sources. Now we can synthesize it in minutes.
The Prompt
We're going to use our favorite two part prompting trick for Deep Research to produce very detailed guardrails for the AI to pull our research...
Prompt 1: Use ChatGPT to Craft the Deep Research Prompt
I am doing investigative research on Gold from the perspective of an investor. I need help from you to craft a detailed ChatGPT Deep Research prompt in our preferred role - input - output style - task - rules framework.
Here are the areas I want to cover:
Current price & recent movement
How has Gold been moving for the last 1 week, 1 month, 1 year and last 5 year? How does it compare vs other asset benchmarks?
Key drivers
What’s pushing the price up (or holding it back) recently? Usually some mix of macroeconomic, geopolitical, and market-structure factors. Who is buying it? Explain macro drivers: inflation, interest rates, U.S. dollar strength/weakness, real yields. Can you detail central bank buying trends (e.g. PBOC, RBI, others). Discuss investor demand: ETFs, futures positioning, speculative vs. structural. Cover geopolitical/safe-haven demand: wars, elections, political instability.
Risk & uncertainty
What could go wrong? What might reverse the trend? What historical patterns (e.g. September seasonality) are relevant?
Forecast / what to watch next
What are analysts expecting? What upcoming events or data might move gold?
Context / framing
How this fits into bigger macro trends (inflation, interest rates, global stability, central banks, etc.). Also comparisons with other investments.
This research should be a balanced 'investment note' style output that remains as unbiased as possible and relies heavily on raw data to support any claims. Before you begin, please ask me any clarifying questions so we are aligned.
Prompt 2: The Deep Research Prompt
Role
You are a professional macro & commodities research analyst tasked with preparing an unbiased, data-driven deep-dive investment research note on Gold (XAU/USD). Your role is to synthesize raw data, authoritative sources, and consensus analysis into a structured institutional-style research report.
Input
Investigate the current state of Gold from the perspective of an investor. Cover:
Current price & recent movement
Latest spot/futures price with retrieval date
Performance over 1 week, 1 month, 1 year, and 5 years
Compare vs. S&P 500, Nasdaq, Bitcoin, U.S. 10Y Treasury, and Oil
Key drivers
Macro drivers: inflation, interest rates, USD strength/weakness, real yields
Central bank buying trends (e.g., PBOC, RBI, other EM/DM central banks)
Investor demand: ETFs, futures positioning, speculative vs. structural demand
Geopolitical/safe-haven demand: wars, elections, political instability
Risk & uncertainty
Potential reversals or downside risks
Seasonal/historical patterns (e.g., September/October seasonality)
Vulnerabilities to shifts in policy, FX, bond yields, risk sentiment
Forecast / what to watch next
Analyst and sell-side forecasts (Goldman, JPM, UBS, World Bank, IMF, etc.)
Upcoming events, data releases, or catalysts that could move Gold
Context / framing
How Gold fits into bigger macro themes (inflation, rates, central banks, global stability)
Comparisons with other investment classes for portfolio role (hedge, diversifier, speculative asset)
Output
A 10+ page deep-dive research note in neutral institutional style with:
Executive Summary (1–2 pages): concise overview of price action, main drivers, risks, and forward view
Main Report Sections: structured by the 5 categories above
Tables & Charts: where possible, include charts (price movement, ETF flows, futures positioning, central bank reserves, etc.) and tables (performance comparison, forecasts, risks)
Data citations: include retrieval dates and source links for each metric
Balanced framing: avoid bias; present both bullish and bearish cases
Task
Write a professional, data-driven, research-style investor note on Gold that could be used by portfolio managers, strategists, or macro investors to understand the current setup, risks, and outlook.
Rules
Use authoritative sources where possible (World Gold Council, IMF, BIS, CFTC, FRED, central banks, major investment banks, etc.)
Always state the date of the latest data used
Include at least one performance comparison table and one drivers/risk summary table
Present both bullish and bearish scenarios clearly
Write in neutral research report tone (not salesy, not casual)
Begin with an executive summary before diving into detailed sections
Aim for 10+ pages equivalent length (long-form, not just bullet points)
The Result:
Full output here: Link
ChatGPT Deep Research delivered what manual analysis would've taken days to compile.
Here's what the AI surfaced that changes how you should think about gold's move:
Central banks are the story, not the Fed. Everyone's focused on rate cuts, but the real driver is 1,000+ tons of annual central bank purchases for three consecutive years. That's 25% of global supply absorbed by buyers who don't trade on sentiment. China alone bought for 10 straight months through August (!) This looks very much to me like systematic dedollarization.
(AI confirms a hunch I've had with $GLD long's through the summer)
My marginal buyer framework reveals significant runway. Gold now represents 27% of global reserves, surpassing U.S. Treasuries for the first time since 1996. Central banks hold $4.5 trillion in gold vs $3.5 trillion in Treasuries. These aren't momentum traders who'll dump on the first correction. Central banks are strategic accumulators building 20 year+ positions.
Investment flows confirm the structural shift. Investment demand jumped 78% year-over-year in Q2. ETF inflows hit largest H1 since 2020 after two years of outflows. But here's the key: this coincided with equity markets at records. Gold isn't just a defensive trade anymore ... it's becoming a core allocation.
Unusual Risk Adjusted Returns: Gold's 39% YTD return beat Bitcoin's 22% while exhibiting far lower volatility. It outperformed during simultaneous dollar weakness (-3% YTD) and equity strength (S&P near 6,600). That's unusual correlation breakdown.
For positioning: I'm personally long on Gold. Feels to me very asymmetric. If the economy runs hot, Gold acts as an inflation hedge. If the economy turns, Gold acts as a defensive instrument. And meanwhile we have central bank purchases backstopping dips in either situation as we see from our AI research.
This type of multi-variable macro analysis (geopolitical + monetary + flow + currency dynamics) is exactly where AI adds leverage. Too many moving parts to synthesize efficiently by hand.
Who's in New York?
I'd love to meet any readers based in New York :)
I'll be speaking at the Generative AI In Investment Management Conference in New York on September 30, 2025 (organized by Columbia University)
I'll be on a panel alongside hedge fund PMs, founders of leading AI investment research products like Boosted.ai, and investment management professors at Columbia.
If you're keen on attending, the organizers have offered a 10% discount if you use promo code Wang10off at this link.
Hope to see you in person!
Posts from Me:
