Your target wallet stops being profitable
The biggest risk in copy trading is also the most boring one: the person you're copying stops winning. Top wallets on a leaderboard are selected on trailing performance, which is a lagging indicator. A wallet that was top-10 last month might be in a drawdown this month because their edge compressed, their bankroll got too big, or the market regime changed.
Rotate targets every 14–30 days. Don't trust a single wallet to keep delivering — maintain 3–5 actively watched targets and prune the ones that fall out of the top 50 after a full 30-day window. Use kolscan.io's rolling windows (daily, weekly, monthly) to detect decay early.
Position sizing blow-ups
Percentage-based sizing ('always copy 5% of the target's trade') sounds sensible until the target goes on tilt and dumps 100 SOL into a bonding-curve token with zero liquidity. Your 5% is now a 5 SOL bag of something that might be worth 0.1 SOL by the time Jupiter routes you out.
Use fixed-SOL sizing when starting. Cap per-trade size at 2–5% of your deployed bankroll. Turn on max-slippage rules so the bot skips trades where the quote is bad. If you must use percentage sizing, cap it in absolute SOL too (e.g. 5% of target's trade, max 1 SOL).
Slippage and thin liquidity
Copying a target who's buying a token with $30k of total liquidity is a trap. Their $2000 buy moved the price 15%; your mirror $200 buy moves it another 3% — you're already underwater before the candle prints.
Enforce a minimum market-cap filter (e.g. $250k+ for bonding-curve tokens, $1M+ for established pairs). Enforce max-slippage (5–8% for memes, 2% for major liquid pairs). Skip trades that can't route cleanly through Jupiter.
Rug pulls and honey pots
Top wallets sometimes buy scams deliberately (they're part of the team) or accidentally (they got rugged too). If your bot mirrors every trade blindly, you will eat the rug. On Solana specifically, honeypot tokens that let you buy but not sell are an ongoing risk vector.
Use a token blacklist (Subglow ships one for known bad actors and updates it daily). Enforce a sellability check — Subglow performs a dry-run sell quote before signing any buy on tokens under $5M MCap. Trust stretched-thin wallets' conviction less than wallets with long trade history.
MEV and frontrunning
If your copy-trade tool submits through public mempool, MEV searchers watch for mirror patterns. They see the target buy, see you buy 800ms later, and sandwich you — frontrunning your entry to push the price up and backrunning your trade to profit on the impact. Over time, this silently erodes 5–20% of your gross PnL.
Only use copy-trade platforms that submit via Jito bundles. Jito bundles land atomically in the same slot as the target, denying searchers the window to interpose. Subglow is Jito-bundled by default. Most Telegram bots are not.
Custodial bot failure
Most Solana Telegram trading bots (Trojan, GMGN in its custodial mode, Maestro, BonkBot variants) hold user funds in bot-managed wallets. Every one of these has been hacked, exit-scammed, or frozen at some point in the 2023–2025 window. User losses across the sector have been hundreds of millions of dollars.
Use non-custodial tools only. Subglow generates a dedicated keypair per user with an exportable private key — if we ever disappear, you sweep the funds yourself. If you insist on a custodial bot, never leave more in its balance than you can afford to lose completely.
Platform-level downtime at the worst moment
Markets move fastest during network stress (Solana congestion events, gas spikes, major token launches). That's also when trading platforms are most likely to have outages or degraded performance. Getting locked out of your positions during a crash is a real risk.
Run server-side stop-losses that execute without your intervention. Keep an exportable private key so you can sweep to a manual wallet if your platform goes down. Don't put your entire bankroll into one platform — split across two tools.
Tax exposure
Every copy-trade fill is a taxable event in most jurisdictions. If you're mirroring 3 wallets at 5 trades a day, that's ~1800 taxable events per year. Without careful bookkeeping and set-aside, you can end a profitable year with a tax bill that exceeds your net gains — or worse, be forced to sell SOL at a loss to cover a tax liability denominated in USD.
Export your trade history monthly (Subglow gives you a CSV button). Feed into Koinly, CoinTracker, or your tax software of choice. Set aside 25–40% of realized gains in a stablecoin pool, topped up weekly. Treat this as non-negotiable — unpaid crypto tax doesn't go away.
Phishing and private-key compromise
Copy trading increases your on-chain footprint: more transactions, more wallets tagged, more visibility. That attracts phishing. The 2024–2026 wave of fake Jupiter, Phantom, and BullX support impersonators has drained wallets belonging to otherwise careful users.
Never DM-clicked support links. Bookmark official URLs. Use a hardware wallet for your main holdings and only fund your copy-trade dedicated wallet with amounts you're actively deploying. If a 'support agent' asks for your seed phrase, you are being robbed.
Psychological blow-ups
The least-discussed risk: watching a bot you configured lose money in real time is emotionally corrosive. Many copy traders tilt after a 3-day drawdown and either yank stop-losses (letting positions drop further) or overleverage to 'catch up' (accelerating the drawdown). The automation advantage is wasted if you override it.
Set your stop-loss, take-profit, and daily loss cap before funding the wallet. Commit to not touching them for the first 14 days. If you tilt, close the dashboard for 48 hours. The bot is doing its job — the risk is you intervening badly.