How Wallet Drainers Work in 2026: Full Technical Breakdown

The Numbers Nobody Reports Correctly

If you search for wallet drainer statistics, you’ll find a lot of articles confidently citing $500 million a year. That figure is from 2024, and it has been wrong for over a year.

Here’s the actual trajectory: losses hit roughly $295 million in 2023 across 324,000 victims, jumped 67% to $494 million in 2024, then fell sharply to roughly $84 million in 2025. Affected wallets dropped 68% over the same period, and the largest single recorded drain fell from $55.4 million in 2024 to $6.5 million in 2025.

An 83% collapse in a single year. That looks like a victory, and partly it was — Inferno Drainer alone had been linked to more than $80 million in theft across more than 16,000 malicious domains before its infrastructure was disrupted, and the industry shipped meaningfully better wallet warnings.

But read the 2026 data and the picture inverts. Signature phishing losses jumped 207% in January 2026 compared to December, while the total number of victims dropped 11% — a strategic shift away from spraying thousands of small wallets toward hitting fewer, larger targets.

That inversion is the single most important thing in this article. The economics of draining changed. The old model was volume: 300,000 victims, a few hundred dollars each. The new model is precision. And precision targets people who consider themselves too experienced to be targets.

One more figure worth sitting with: the largest single drain on record took $55.47 million in DAI from one victim. One person. One signature.

The Kill Chain: Five Stages, Three Chokepoints

Every drainer attack, regardless of which kit powers it, follows the same sequence. Understanding the sequence matters more than memorising any individual scam, because the wrapper changes constantly and the mechanics don’t.

Stage 1: Distribution — Getting You to the Page

The drainer payload is trivial. Getting a wallet-holding human in front of it is the actual work, and it’s where the innovation happens.

Current distribution channels, roughly in order of prevalence:

• Search and malvertising — paid ads on branded search terms, so the fake ranks above the real
• Compromised legitimate front-ends — the most dangerous variant, because the URL is correct
• Social media reply-jacking — replies inserted under authentic posts from real projects
• Direct messages — Telegram, Discord, X, increasingly with AI-generated personalisation
• Fake CAPTCHA pages (ClickFix) — a different attack shape covered separately below

Front-end compromise deserves emphasis because it defeats the advice most guides give. Angel Drainer was the payload behind the Ledger Connect Kit incident of December 2023, and the 2022 BadgerDAO front-end attack cost $120 million. More recently, in March 2026 BONK.fun was hit after attackers hijacked a team account and deployed a wallet drainer on the site’s own domain.

If the attacker controls the real site, “check the URL” doesn’t save you. Only the signature-stage defences do.

Stage 2: Reconnaissance — The Wallet Scan

The moment you click “Connect,” before you’ve approved anything of consequence, the drainer reads your public on-chain position. The malicious site detects your wallet, scans your token balances and existing approvals, and serves up the most profitable prompt it can craft.

This is why two people visiting the identical scam site see different requests. The kit is doing portfolio analysis in real time — indexing token balances, liquid staking derivatives, NFT holdings, and critically, approvals you already granted months ago to legitimate protocols.

A connection request by itself steals nothing. But it hands over a complete map of what’s worth stealing and which door is already unlocked.

Stage 3: The Signature — Where the Theft Actually Happens

This is the entire attack. Everything before it is theatre; everything after is settlement.

There are four mechanisms an attacker can ask you to authorise, and they are not equally dangerous.

approve() — the legacy method. An ERC-20 approve() can hand over an unlimited allowance, letting a contract spend that token forever. It’s an on-chain transaction, it costs gas, and it shows up clearly in your wallet as a transaction. Increasingly, that visibility is why attackers have moved away from it.

setApprovalForAll() — the NFT sweep. Grants control over every NFT in a collection in a single click.

permit() / Permit2 — the 2026 weapon of choice. This is the one that matters. Permit and Permit2 signatures cost no gas, appear as a plain message rather than a transaction, and still authorise a transfer. That is the trap most people walk straight into.

The mechanism: instead of an on-chain approve(), the attacker collects an off-chain EIP-712 signature via eth_signTypedData_v4. No transaction. No gas from your side. Just a pop-up that says “Sign this message.” A malicious site can present what looks like a routine signing request that actually authorises the transfer of your entire token balance — then the attacker calls transferFrom().

The data on how effective this is should end any debate: across Scam Sniffer’s 2024 cases above $1 million, permit accounted for 56.7% of losses by value, setOwner for 31.9%, transfer for 4.5%, and increaseAllowance for 3.5%.

Over half of all large-value drainer theft came through a mechanism that most wallets render as an unreadable blob of typed data with a “Sign” button underneath.

Why Permit2 specifically is so exploitable. Permit2 is Uniswap Labs’ universal token permission system, deployed in November 2022, and by 2026 it has become the most widely integrated approval contract in crypto, sitting at the identical address 0x000000000022D473030F116dDEE9F6B43aC78BA3 on every major EVM chain. If you’ve used Uniswap, 1inch, CowSwap, or MetaMask Swaps in the last two years, you have almost certainly signed a Permit2 message without registering that you did.

That familiarity is the vulnerability. A Permit2 prompt from a drainer looks exactly like a Permit2 prompt from Uniswap, because structurally it is one.

Some kits go further and chain a Permit2 signature with a legitimate-looking ETH transfer to disguise the drain — so the transaction you’re inspecting looks like the swap you asked for, while the signature riding alongside it does the actual damage.

EIP-7702 and account abstraction represent the emerging frontier here. Modern drainers increasingly exploit newer protocol integrations like EIP-7702 alongside Permit2. The pattern will keep repeating: every new UX improvement that reduces signing friction becomes a new drainer surface within months. That’s not an argument against the improvements — it’s an argument for reading what you sign, permanently.

On Solana, the equivalents are token delegate assignments or account ownership changes bundled into an innocent-looking transaction. Different primitives, identical psychology.

Stage 4: Execution

Once the signature is captured, the attacker submits it on-chain and moves assets. In sophisticated kits this is automated and ordered by value — highest-value assets first, in case the victim notices mid-drain and starts moving funds.

Stage 5: Laundering

Proceeds route through mixers, cross-chain bridges, or DEX swaps. In the highest-value 2026 incidents, conversion to Monero has been the observed endpoint — often before the theft has even hit the news cycle.

Drainer-as-a-Service: Why This Is an Industry, Not a Crime Wave

The most consequential thing to understand about drainers is that the person who stole from you almost certainly didn’t build anything.

In the Drainer-as-a-Service model, one developer builds and maintains the kit and rents it out. The affiliate — often a low-skill actor — handles the phishing. The kit handles the theft. The split copies the ransomware playbook almost exactly: developers keep about 20%, affiliates keep 80%. For that cut, the buyer gets ready-made phishing pages, a control dashboard, anonymity tools, and customer support.

The service provides prebuilt scripts, phishing templates, hosting, dashboards, and operational support. Affiliates can launch campaigns without developing any wallet-draining code themselves.

The lineage matters because it explains why takedowns don’t work. The 2023–2026 chain runs through Monkey Drainer (shut down March 2023 following ZachXBT’s investigation), MS Drainer (active 2023 to early 2024), Inferno Drainer (roughly $80 million between late 2022 and November 2023, disrupted by Group-IB in January 2024), Pink Drainer (wind-down announced 2024), and Angel Drainer.

Watch what actually happened at each “shutdown.” Inferno announced retirement in November 2023 — then resumed operations by May 2024, claiming far higher figures than its original run. In October 2024 the Inferno team transferred control of its toolkit to rival provider Angel Drainer — a merger, not an exit. Pink Drainer cleared about $85 million before its operators announced they were quitting.

One crew retires, criminals shift to the next kit, and Inferno itself came back in reloaded form. Killing one operator does not kill the market — the market is the service, not the scammer.

That’s the correct mental model. Enforcement disrupts capacity temporarily. It does not remove the category.

What Changed in 2026

Three developments have reshaped the threat since the 2025 lull.

1. Drainers weaponised live hacks

This is genuinely new and genuinely nasty. April 2026 became the worst month for crypto theft on record — over $629 million drained across more than 20 incidents, led by a $292 million KelpDAO breach and a $285 million Drift Protocol exploit.

When a hack that size breaks, security teams urgently tell users to revoke approvals. Drainer operators learned to hijack exactly that moment. Threat researchers tracked five separate exploit-phishing campaigns in April 2026 alone, where operators registered fake revoke and migration sites within hours of major hacks to target panicked users.

The ZetaChain case is instructive because it shows the timing precisely. After a cross-chain exploit on 27 April 2026, as security teams advised users to revoke approvals across all EVM chains, a malicious lookalike revoke domain appeared and was posted in direct reply threads under the legitimate community alert — pointing users toward the fake site under the guise of helping them protect themselves.

Different kits, different infrastructure, different operators, all exploiting the same window of panic. This isn’t one coordinated group — it’s a technique adopted across the entire drainer ecosystem.

2. ClickFix — malware without a download

A parallel attack that doesn’t touch your wallet’s signing flow at all, and therefore bypasses every defence built around it.

You land on a page showing a broken CAPTCHA. It offers an alternative: press Win+R, press Ctrl+V, press Enter. What’s already sitting in your clipboard is an obfuscated PowerShell line, placed there by the page itself.

The command launches PowerShell or mshta to pull the real payload. Because you pressed Enter yourself, no browser download warning fires and no confirmation prompt appears. The final stage is usually an infostealer or remote-access trojan, harvesting email logins, browser cookies, banking credentials, and crypto wallet files.

Two details reveal how carefully this is engineered. Many ClickFix prompts include a long string of leading whitespace before the actual command, so when the victim glances at the Run dialog they see something innocuous like a verification ID — the real command is scrolled off-screen to the right. And the technique abuses tools that already ship signed and trusted on every Windows machine, sidestepping security software that watches for suspicious downloads.

The growth rate is the alarming part: ESET recorded a 517% increase in ClickFix and FakeCAPTCHA campaigns in the first half of 2025 compared with the second half of 2024, and Microsoft confirmed the technique reaches thousands of devices a day globally. It is no longer Windows-only — a Huntress threat hunt in June 2026 found Go-based Mac stealer malware on a system infected via ClickFix three months earlier, harvesting credentials from the macOS Keychain.

The rule is absolute and easy: a real CAPTCHA never asks you to open a terminal. Real verification asks you to click a box, solve a visual puzzle, or wait. It should never need operating-system commands.

3. The social engineering ceiling rose

A January 2026 incident illustrates the ceiling: a caller posing as a support agent walked a hardware wallet owner through a fake verification process and got them to read out their recovery seed. The result was 1,459 BTC and 2.05 million LTC — roughly $284 million, and 71% of all crypto theft losses recorded that month.

Note carefully what this was and wasn’t. No encryption was broken. No device was compromised. The hardware wallet worked exactly as designed. The attack bypassed the technology entirely by attacking the person holding it. No hardware wallet, on any roadmap, will ever protect against a user who reads their seed phrase aloud.

Trezor, Ledger, MetaMask, Coinbase, Binance — none of them will DM you or call you first. That single sentence would have prevented a $284 million loss.

The Three Chokepoints (And Why the Third Is the Only One That Always Works)

Here is the structural insight that makes drainers defeatable:

Every drainer kit terminates at the same signature event, and every drainer kit fails if the victim refuses the signature.

The page-landing step depends on you arriving at a malicious URL. The wallet-connect step depends on you initiating the connection. The signature step depends on you pressing confirm.

Chokepoint one (URL verification) fails against front-end compromise. Chokepoint two (connection hygiene) fails when the site is legitimate but hijacked. Chokepoint three never fails, because there is no drainer variant that steals without a signature.

Everything below is ranked by which chokepoint it defends.

Defence Layer 1: Never blind-sign

If your wallet shows an undecoded blob of typed data and a Sign button, and you don’t understand exactly what it authorises, reject it. Full stop.

Use a wallet or extension that decodes and simulates. Simulate transactions using tools like Rabby or Pocket Universe before signing. These show you the predicted post-transaction state — what actually leaves your wallet — rather than raw calldata. Blind-signing because the wallet doesn’t decode is one of the failure patterns appearing repeatedly in 2026 victim reports.

Defence Layer 2: Audit approvals, including both Permit2 layers

Most people who check approvals check them wrong, because Permit2 is nested.

There are two layers. The first is the ordinary ERC-20 approval you granted to the Permit2 contract itself, which is usually unlimited. The second is the set of sub-allowances Permit2 holds on your behalf. Revoke the outer approval when you’re not actively using an application that needs it. For the inner ones, Permit2 exposes lockdown to batch-revoke approvals and invalidateNonces to kill signatures you’ve signed but nobody has submitted yet.

Revoke.cash surfaces both layers, which is far easier than calling the contract yourself.

The scale of neglected exposure here is extraordinary: Chainalysis estimated that more than $2.4 billion in active unlimited approvals still sit on lost or compromised wallets, waiting to be exploited.

An important, under-stated point: revoke even after you’ve already been drained. Revocation recovers nothing — what has left the wallet has left, and an already-exploited signature can’t be unwound. What revocation does is stop the same permission being used a second time.

Defence Layer 3: Hardware wallet, used properly

A hardware wallet means the signature request must be confirmed on a device screen your attacker doesn’t control. That’s the point.

But it is not automatic protection. If a hardware wallet was paired, the seed itself is safe and you can recover onto a clean device — however, if you approve a malicious signature on the device, the theft proceeds normally. The hardware wallet protects your keys. It does not protect you from your own confirmation.

Read the device screen. Every time. That habit is the product you actually bought.

Defence Layer 4: Wallet compartmentalisation

Keep a dedicated “hot” wallet with a small, deliberately capped balance for interacting with any new dApp. Keep long-term holdings in a separate hardware-secured wallet that never connects to anything experimental.

This converts a catastrophic loss into an annoying one. It is the single highest-leverage habit in this entire article and it costs nothing.

Defence Layer 5: Treat urgency as the attack

Given that panic-exploitation is now a standard drainer technique, invert your instinct. When you see a hack alert, a migration notice, or an urgent revoke instruction — that is precisely when to slow down, navigate manually to revoke.cash or a block explorer, and never click a linked “solution,” regardless of what account appears to be posting it.

If You’ve Been Drained: An Ordered Response

1. Move remaining assets immediately, prioritising by value. If any assets remain, they’re on a clock.
2. Assume the wallet is permanently burned. Don’t reuse the address, and don’t send anything back to it.
3. Revoke everything on the compromised address — to prevent reuse, not recovery.
4. Determine the vector honestly. Was it a signature (drainer) or a key compromise (malware/seed exposure)? If your seed was exposed or your machine was infected, a new wallet on the same machine is also compromised. Rebuild on clean hardware.
5. Document the spender contract address for incident reporting, and report to Chainabuse and the relevant security trackers. This won’t recover funds; it does help flag the contract for other users.
6. Record the loss for tax purposes. In most jurisdictions crypto theft losses have tax treatment.

Set expectations honestly: many victims recover little or none of what was stolen. The value of steps 3–6 is preventing the next loss, not undoing this one.

The Bottom Line

Drainers aren’t a hacking problem. Nothing about them breaks cryptography, and nothing about them exploits the blockchain. They compromise the human layer through front-end deception and malicious off-chain signatures.

Which means the defence isn’t technical sophistication. It’s a small number of habits held consistently:

• Never sign what you can’t read. Use a wallet that decodes and simulates.
• Audit approvals quarterly, both Permit2 layers.
• Compartmentalise — a capped hot wallet for experiments, cold storage for holdings.
• Read the hardware wallet screen rather than confirming reflexively.
• Treat urgency as hostile, especially urgency attached to real breaking news.

The 2026 shift toward fewer, larger targets means experienced users are now the market. If you read this thinking you already knew most of it, you are exactly who the current generation of operators is optimising for.

AI-Powered Crypto Heists: The Frontier of Digital Crime in 2025Seed Phrase Security Guide 2026 | How to Back Up & Protect Recovery KeysWallet Drainers 2026: New AI Tactics & Prevention GuideExposing Crypto Exit Liquidity Traps: Your Guide to Staying Safe