AI bubble tactics for all 16 investor types
'This time is different' vs 'it always ends the same' — 16 personas, 16 calibrations.
The AI theme is the hardest cycle to run — a genuine paradigm shift and market mania running at the same time. The right answer is different per persona.
Below are 16 action cards: which trap is biggest for your type, which lever is strongest, and where the boundary is between conviction and obsession.
Why now?
AI infrastructure (chips, power, data centres) and AI applications (SaaS, robotics) have different risk/reward shapes. Buying them as a single bundle is the most common trap.
Single-name 'winner' bets are unbacktestable, so position sizing, written stop levels and a hard vision cap are your three safety nets.
Best-fit personas for an AI cycle
- 1🐉RILCThe Visionary
Long-horizon vision bets are natural — if the cap holds.
- 2🦅RDLCThe Precision Sniper
Strongest at building a real AI thesis from primary data.
- 3🦉RDLAThe Asset Architect
Diversification and rules keep distance from mania.
- 4🐢PDLAThe Armored Turtle
Index ownership absorbs AI without the mania.
- 5🦫PDLCThe Solitary Value Hunter
The 'patient collector' wins when prices reset.
Most exposed to AI mania
- 1🔥RISCThe Moth to the Flame
Intuition + concentration + day-trading amplifies the bubble.
- 2🐝PISAThe Curious Bee
'I want to try everything AI' is the most expensive instinct.
- 3🐿️PISCThe Cautious Squirrel
Fear vs greed paralysis kills the rules.
- 4🐰PILAThe Peaceful Gardener
Tip-following plants too many AI seedlings.
- 5🐬RILAThe Trend Navigator
Trying to ride every AI sub-theme breaks prioritisation.
What each of the 16 types should do
Tap your code to jump to that type's full guide.
Check whether 1-2 AI compounders really have a thesis.
- · Pick one bucket — infrastructure or applications — and go deep
- · Re-run a 'substitution risk' scenario each quarter
- · Keep 10-20% cash as 'post-bubble' ammo
- · Avoiding the cycle entirely because it 'looks expensive'
- · But also: 50% in a single name
Cap AI weight, leave the allocation grid alone.
- · Cap AI-related assets at 5-10% of the core sleeve
- · Add one line of 'AI-infra via REITs/utilities' indirect exposure
- · Run rebalancing on schedule
- · Surging into one sub-theme on conviction
- · Letting a single AI ETF cross 30% of total assets
Volatility is the stage — rules harden the most.
- · Use a tighter −1.5% stop on AI trades only
- · No trades on event days (earnings, product launches)
- · Sweep profit weekly into the long index sleeve
- · Options or leveraged AI ETFs as a 'one-shot recovery'
- · Adding 'just one more' trade during runs
Regime shift — keep your rule base.
- · Isolate AI-related strategies into a separate 'theme' bucket
- · Backtest a new signal once a quarter
- · Do not change the per-strategy capital cap
- · Pushing social/YouTube signals into models on impulse
- · Boosting capital from a 1-year AI-only backtest
Your natural stage — cap and trim are mandatory.
- · Defend the 40% cap on vision allocation
- · Trim quarterly into other sleeves
- · Keep 2-3 candidate winners, not 1
- · 'This time is really different' cap breaches
- · Skipping short-seller reports
Narrow your AI sub-themes.
- · Pick 1-2 from infra / apps / semiconductors
- · Cap any new AI theme at 5% of assets
- · Run a monthly per-theme P&L check
- · One 'all-of-AI' ETF as total exposure
- · Adding new themes at the top
Highest-risk cycle — outside rules first.
- · Trade AI only inside the high-risk account
- · Auto-sweep profit into long index funds
- · Enforce the 24h cooldown rule
- · Single-name AI concentration + leverage
- · 'My cycle' breaches of the risk cap
Rotate inside AI — but keep stops short.
- · Use a −3% stop on AI trades only
- · Auto-sweep some profit into MMF / short bonds
- · Cut 'no longer fun' themes on rule
- · 'This time I'll hold' attachment to one AI name
- · Spilling over into AI-tagged coins / NFTs
Wait for the 'safety margin' price on AI quality.
- · Add 2-3 AI compounders to watchlist
- · Tranche-buy only at your margin-of-safety price
- · No selling before next earnings
- · Impulse buys because 'it's too late'
- · Entering on YouTube tips
Hold core diversification; cap the AI satellite.
- · AI-related assets at 5-10% satellite cap
- · Auto-rebalance as planned
- · Defend the 6-month reserve
- · 'Grow the core this time' weight breaks
- · Breaking diversification with single AI adds
Only trade AI when the signal is unambiguous.
- · Re-write the AI entry trigger in one line
- · 24h rest after each take-profit
- · Monthly rule vs P&L review
- · Entering 'because everyone is'
- · AI-only rule deviations
Stay on the system — let the index hold AI for you.
- · Get natural AI exposure via global index
- · Hold DCA ratios steady
- · Block 'one more AI ETF' adds
- · Adding AI ETF to DCA
- · 'This time is different' rule edits
One AI quality name is enough.
- · Quarterly tranche-buy one AI compounder
- · Explain its thesis to a third party
- · Defend the 6-month reserve
- · Letting it run past 50% on conviction
- · Adding more names from tips
Plant 1-2 AI seedlings in the garden, not 10.
- · Cap AI-related names at 5-10% of assets
- · Re-test theses for tip-driven AI buys
- · Add only to the seedlings that grow
- · Adding AI names from acquaintance tips
- · Hanging on to AI names you've lost interest in
'Small but real' size — keep stops short.
- · AI trades sized at ≤3% of assets
- · Pre-written stop levels
- · Defend the emergency reserve
- · 'Just a bit more' size creep
- · Avoiding markets for 6 months out of fear
Keep AI inside the 'sampler' budget.
- · Buy AI only inside your monthly experiment budget
- · Pick one AI sample to hold for 1 year
- · Isolate the main book
- · Spilling into AI-tagged coins / NFTs
- · Breaking the experiment budget
Frequently asked questions
- Should I buy AI names now?
- This page lists per-type behaviours, not timing. Combine it with your own financial situation and time horizon.
- Are themed AI ETFs safer than single names?
- They look diversified but they cluster against one cycle. Depending on them for total exposure is one bet, not many.