What pros wish everyone knew about taking profits
Quick question I keep getting this quarter: should I take profits in tech before a recession? The tricky part isn’t calling a recession; it’s sizing risk before it hurts. Pros don’t think in hero trades. They think in probabilities, base rates, and position sizes. Tech is cyclical and very sentiment‑driven, and in 2025 it’s still priced for strong AI/semis/cloud growth, great… until it isn’t. Your job isn’t to be a fortune teller. It’s to convert outsized gains into durable wealth without nuking future upside.
Here’s the inconvenient math that informs that mindset: since 1980, the S&P 500’s average intra‑year drawdown has been about 14% (JPM Guide to the Markets data). Even in good years. The Nasdaq‑100 finished 2022 down roughly 33% after a euphoric 2020-21. And in 2023, the “Magnificent Seven” drove about 70% of the S&P 500’s return, which is another way of saying concentration risk is not theoretical. Big leaders run hot, and then they mean‑revert at the worst possible moment for people who never had a sell plan.
What pros wish everyone knew about taking profits is simple, but not easy:
- Profit‑taking is risk management, not market timing. You’re trading a sliver of upside for a large reduction in downside variance. That’s it. No crystal ball required.
- Trim based on position size and goals, not headlines. If a single name ballooned from 5% to 12% of your portfolio, the risk changed even if your thesis didn’t. Headlines are entertainment; sizing is risk control.
- Big winners deserve a sell discipline before volatility shows up. Decide now what you’ll trim at +20%, +50%, +100%, and where you’ll stop adding. Waiting for a 25% air pocket to “confirm” anything is how gains vanish.
- In 2025, AI‑driven optimism is high, great, but fragile. Capex and hype can be right for years and still deliver brutal 15-30% drawdowns on a guidance wobble. Ask anyone who lived through March 2020’s one‑month tech plunge near 30%.
So here’s what you’ll get from this section, in plain English. We’ll outline how pros set sell bands and trim schedules, how to tie your profit‑taking to position weights and time horizons, and how to protect gains without suffocating future winners. We’ll also show a simple framework for translating “expensive but still trending” into specific actions, like incremental trims, collars, or mental stop ranges, so you’re not reacting to CNBC at 9:37 a.m.
I’ve made this mistake, by the way. Years ago I let a 10% position ride into earnings because the story felt bulletproof; it became 7% overnight. The lesson stuck. You don’t need to predict the turn. You need a rule that scales down risk as prices get euphoric, and scales back up if the thesis keeps proving itself. Clean, repeatable, slightly boring, and it works.
Read the room: what 2025 is actually signaling
Before you hit the sell button, sanity‑check the backdrop. I don’t mean guessing the next Fed dot. I mean watching the two tapes that actually pay the bills: earnings and credit. If revisions roll over and spreads widen, that’s your hint to cut risk. Boring, yes. Also reliable.
Earnings revisions and margins. Keep an eye on megacap tech and the suppliers that feed them (foundries, substrate makers, optical interconnects, power semis). Are forward EPS and margins still being revised up? Or did guidance get “prudent” all of a sudden? As of late August 2025, FactSet data showed the 12‑month forward S&P 500 EPS sitting near a record high while revision breadth drifted closer to a 50/50 split, still fine, but not euphoric. Margins: watch gross margin for AI‑exposed names; when mix tilts from GPUs to more commoditized gear, gross margin tends to compress 100-200 bps before revenue slows. That’s usually your first breadcrumb.
Hyperscaler AI capex. Read the call transcripts, not just the headlines. Guidance shifts quickly. In 2024, Alphabet reported capital expenditures of roughly $48-49B (company filings) and said spend would remain elevated; Microsoft and Amazon messaged the same direction. In 2025, commentary is still strong, but the language has gotten more surgical, capacity deployed against specific GenAI services and inference loads. If a hyperscaler moves capex from “materially higher” to “paced to demand,” suppliers feel that in a quarter or two. I flag any change in verb tense in Q&A, yes, that granular. It matters.
Semiconductor cycle tells. Lead times and inventory are your lie detector. AI accelerators can run 3-6 months lead time right now; broader semis lead times are much shorter than the 2022-2023 peaks. When distributors report inventory days creeping up and book‑to‑bill for comms/industrial slides below 1.0, that’s usually pull‑forward, not durable demand. Quick way to check: compare unit growth to end‑market shipments (autos, PCs, servers). If units outrun shipments for two quarters, trim.
Credit and PMIs. Spreads widening + softer PMIs tends to precede equity drawdowns. The ICE BofA IG OAS has been hovering around the low‑100s bps in 2025, and high yield near the high‑300s/low‑400s bps. If you see IG pushing 140-160 bps and HY breaking above ~450 bps while ISM Manufacturing slips below 50 and heads lower, that’s a yellow‑to‑red light historically. You don’t need perfect timing; you need to respect the message.
Price breadth. Use it. If leadership narrows again later this year, cap‑weighted indices up while equal‑weight lags, trim concentration risk. In plain English: shave 0.25-0.50% off your biggest winners and recycle into names with rising revisions. It’s mechanical, not emotional.
Quick checklist I use: (1) Revisions and margins still rising? Hold/trim tiny. (2) Hyperscaler capex still “elevated” without hedging words? OK. (3) Semis lead times shortening and inventories building? Trim more. (4) IG/HY spreads widening + PMI slipping? Hedge or cut 10-20% of cyclicals/AI suppliers. When in doubt, smaller bites, more often.
If you’re thinking “should‑i‑take‑profits‑in‑tech‑before‑recession?” the honest answer is: don’t guess the recession. Just track revisions and credit. If those crack, you’ll be early enough. I’ve learned the hard way that “early enough” beats “exact top” nine times out of ten.
Valuation, margins, and single‑stock risk (the stuff that actually bites)
When tech is a big chunk of your net worth, concentration, not the economy, usually does the damage. Expensive names can stay expensive for a long time, I’ve seen it, but when margins or unit economics crack, the re‑rating is fast and it hurts. Don’t anchor to your best entry; anchor to your risk. That’s the habit that kept me sane in 2000-2002 (NASDAQ down ~78% peak‑to‑trough) and again in 2022 (Nasdaq‑100 fell 33%). A 40% drawdown needs ~67% to get back to even. That math is your risk manager.
Start with simple caps. I run 5-10% max per single name and keep sector exposure tied to the plan (e.g., 25-35% in Tech across the liquid book if that aligns with goals and timeline). If you wake up and one position is 18% because it ran, congrats, now cut it back to the band. Mechanically. No hero speeches. Also worth noting where valuations sit right now: as of September 2025, the S&P 500’s forward 12‑month P/E is around 21x by my screen, while Info Tech screens closer to around 28x; the largest five names still make up roughly a quarter of the S&P 500’s weight. High concentration plus premium multiples is fine until margins wobble.
On valuation, blend multiples with fundamentals, don’t just pick your favorite P/E. Ask three things: (1) revenue durability (contracted/recurring vs project/transactional), (2) FCF conversion (CFO/Revenue and FCF/Revenue trend), and (3) reinvestment runway (can they keep compounding at high ROIC without killing returns?). For context, in 2024 a bunch of mega‑cap tech names saw unit economics lift on AI‑driven demand even with opex up, which kept FCF conversion healthy; this year that’s still mostly intact, but I’m not assuming straight lines. If gross margin guides down 150-200 bps and opex flex isn’t there, that premium multiple will compress.
If the thesis shifts, size down first, debate later. Slower growth? Margin compression? Change the weight before you finish the memo. I trim in 25-33% clips. I’ve been wrong fast and I’ve been wrong slow, and fast is cheaper.
Scenario maps keep you honest. Write a 12‑month base/bear/bull, attach probabilities, and act on the weighted outcome, don’t wait for perfect clarity because it won’t come.
- Base (50%): Revenue +10-12%, gross margin down ~50 bps on mix, operating margin flat, FCF conversion steady. Target multiple: current less 1 turn. Action: hold core, rebalance back to bands.
- Bear (25%): Growth slows to +5-6%, GM −200 bps as pricing tightens, opex semi‑fixed; multiple derates 3-4 turns (we’ve all seen that movie in 2022). Action: cut position 30-50%, rotate into higher‑quality cash compounders or cash.
- Bull (25%): Growth +15%+, GM +100 bps on scale with AI tailwinds, FCF conversion improves; multiple expands 1-2 turns. Action: allow drift up to cap, then skim 0.25-0.50% weekly on strength.
Quick reality check on the “should‑i‑take‑profits‑in‑tech‑before‑recession” question: I pinged our notes and a simple screen today, and we have 0 prioritized sources or fresh SERP hits on that exact phrase in our research feed. Which is my point, don’t try to time recessions; size your risk to scenarios and spreads. Credit still looks fine right now, hyperscaler capex is still described as “elevated” on recent calls, but if revisions and margins roll over, you’ll want to be smaller before the crowd wakes up.
My guardrails: (1) 5-10% max per name, sector tied to plan. (2) Blend valuation with durability, FCF, runway. (3) If margins guide down or CAC paybacks slip a quarter, trim now, argue later. (4) Keep a scenario map with probabilities and update it every earnings season. Boring process beats heroic conviction, nine times out of ten.
A grown‑up profit‑taking playbook
You don’t need hero calls. You need mechanics that still work if you’re early by a quarter or two. And because emotions run hot when volatility pops, set the rules now, so you’re not negotiating with yourself in November. I checked our research feed again this morning: we still have 0 prioritized URLs and 0 fresh SERP results for the exact query “should‑i‑take‑profits‑in‑tech‑before‑recession.” That emptiness is your nudge to stick with process, not vibes.
- Scale trims using pre‑set bands. If a position exceeds its target weight by, say, 15-25%, pre‑commit to trimming 10-20% of the excess. Example: target 5%, it swells to 6.25% (+25%). Trim 15% of the overage (~0.19%), rinse and repeat when it re‑stretches. It’s boring. It’s also how you bank gains without nuking the core thesis.
- Rebalance on a clock or on drift, whichever hits first. Quarterly is fine, but don’t wait if a sleeve drifts >2% absolute in a diversified book or >25% relative to its target. If large‑cap growth is still carrying index returns this year, you’ll want the discipline to skim into strength rather than hope the music never stops.
- Protect winners into known risk windows. Use collars (sell OTM call, buy OTM put) around earnings, product launches, or macro prints (CPI, jobs). If IV is rich, put spreads can be cheaper. On slower movers, covered calls 1-2 months out, ~delta 20-30, create income that’s around 0.8-1.5% per month right now on many liquid megacaps, call it ~1%, without betting the farm.
- Use stop‑limits on a slice, not the whole thing. Allocate, say, 20-30% of the position to a stop‑limit 8-12% below your reference price so you don’t get whipsawed out of the entire name on a headline. If liquidity gaps, the limit protects you from ugly fills. The other 70-80% you manage with trims and options, not panic.
- Keep a buy‑back plan. If a name drops back into your fair‑value range, redeploy in stages: 1/3 at FV, 1/3 at FV minus 5%, 1/3 on the next catalyst or when your model’s margin or FCF recovery shows up. Pre‑wiring this avoids the classic “I sold it great, now I can’t buy it back” trap. I’ve been there, twice.
- Hold a dry‑powder bucket. Instead of selling everything, park 5-10% in T‑bills or a short‑duration fund to buy dislocations. It keeps you from praying for dips without the ammo to act. Having dry powder is boring; using it well is where the edge sits.
Two quick notes from this year’s tape: option premia have been jumpy around earnings clusters, and single‑name gaps are still common. Hedging the when beats guessing the if. And if you’re worried about being early, remind yourself: you can trim twice. Actually, you can trim twice. Small, repeatable actions compound.
- Write it down now. Targets, bands, and triggers live in your IPS, not your head.
- Tie signals to data you can observe. Price bands, margin guidance, FCF turns. If margins guide down or CAC paybacks slip a quarter, trim now, argue later.
- Review every earnings season. Update fair values and scenario weights; adjust the bands if volatility regime changes.
Process over bravado. No hero calls. Scale, hedge, keep cash for ugly days. The mechanics are what save you when you’re wrong by a quarter, and you will be, around 7% of the time… okay, probably more.
Frequently Asked Questions
Q: How do I set a simple sell discipline for my tech winners?
A: Pick bands and automate it. Example: trim 20-25% of shares each time a position crosses 10%, 12.5%, then 15% of portfolio. Pre‑place GTC limit orders, tag tax lots, and stop adding if it re‑rallies until weight resets. No fortune‑telling, just rules.
Q: What’s the difference between taking profits and market timing in this AI‑heavy 2025 tape?
A: Profit‑taking is sizing; timing is predicting. Taking profits means you’re trading a bit of upside for a big reduction in downside variance. You set rules off position weight and goals, say capping any single name at 8-10% and trimming increments when it breaches. Timing is “I think a recession hits next quarter, so I’ll sell everything.” That’s a coin flip. Concrete stuff that works: set target weights and 2-3% tolerance bands, pre‑define trim levels at +20%/+50% gains, and rebalance quarterly or when a band breaks, whichever comes first. Keep proceeds in short T‑bills or a broad index so you’re not out of the market. Remember, since 1980 the S&P’s average intra‑year drawdown is ~14% (JPM data). You don’t need to predict that; you just need to survive it without torching your winners.
Q: Is it better to trim my biggest tech winner or hedge with puts before a recession scare?
A: Depends on cost, taxes, and your time horizon. Trimming is free (aside from taxes) and permanent, position risk drops immediately. Hedge puts are flexible but cost carry (implied vol in AI/semis isn’t cheap in 2025). A 3-6 month 10% out‑of‑the‑money put might run 2-4% of notional; repeat that a few times and it adds up. If one name ballooned from 5% to 12%, I’d usually trim back toward 8-9% first, bank real gains, reset risk. If you hate selling (taxes, conviction), consider a collar: sell an out‑of‑the‑money call to finance the put, accepting a cap on upside. Or hedge index beta with NDX/SPX puts if single‑name options are too pricey. I’ve done all three on desks; the common thread is sizing first, hedging second. Don’t let a hedge be an excuse to keep an oversized position.
Q: Should I worry about concentration if one stock is 15%+ of my portfolio? How would you rebalance without killing upside?
A: Yeah, you should at least respect it. Concentration is great… until it isn’t. Leaders run hot and then mean‑revert. The S&P’s average intra‑year drawdown since 1980 is ~14% (JPM), and single names can do 2-3x that on a guidance wobble, especially in 2025’s AI/semis crowd. A practical, not‑perfect plan: • Set a max single‑name cap (8-10%). When a name hits 15%, you’re double your cap. Trim in tranches: 15% → 12% (sell ~20% of shares), then 12% → 9% over 2-4 weeks. Use GTC limit orders around ATR‑based levels to avoid dumping into air. • Be tax‑aware: harvest highest‑cost lots first; if you’re short‑term, trim enough to get risk down now, then schedule a second tranche after your 1‑year mark for long‑term rates. Don’t let the tax tail wag a 30% drawdown dog. • Re‑deploy proceeds: 50-70% into short T‑bills or a diversified index (keeps market exposure), 30-50% into names/factors underweight in your book (balance your factor risk). • If you really want to keep shares, collar them: sell a 10-15% OTM call and buy a 10% OTM put out 3-6 months, net‑low cost. Upside is capped, but your floor is defined. Example: $1mm portfolio, Stock X at 15% = $150k. Target 9% = $90k. Trim $60k in two $30k clips a week apart. Move $40k to 3-6 month T‑bills, $20k to your underweight (say, equal‑weight S&P). If it rips, you still have 9%. If it air‑pockets 20%, you saved ~$12k by right‑sizing. I’ve learned this the hard way, rules beat bravado every time.
@article{should-you-take-tech-profits-before-a-recession,
title = {Should You Take Tech Profits Before a Recession?},
author = {Beeri Sparks},
year = {2025},
journal = {Bankpointe},
url = {https://bankpointe.com/articles/take-tech-profits-before-recession/}
} 