Telegram trading bots let users buy and sell tokens directly from a chat interface. This guide explains how they work, which bots dominate the market, and what risks come with handing a bot your private key.

Telegram trading bots emerged in 2023 as a response to a specific problem in decentralized finance: the gap between the speed at which opportunities appear and the speed at which a human can execute a trade through a conventional DEX interface. When a new token launches on Uniswap or Raydium, the first buyers often capture the largest gains. By the time a trader opens their browser, connects their wallet, approves the token contract, sets slippage, and confirms the transaction, the price may have already moved 50% or more.

Telegram bots compressed that entire workflow into a single message. Paste a contract address, tap a button, and the bot submits the transaction on your behalf. The interface is a chat window. The execution happens on-chain. The speed advantage turned what started as a niche tool for memecoin traders into an infrastructure layer that now processes billions of dollars in monthly volume.

This guide explains how these bots work under the hood, which ones dominate the market, what they cost, and where the risks hide.

How Telegram trading bots work

A Telegram trading bot is a program that runs on a server, connects to one or more blockchain networks, and accepts commands through the Telegram Bot API. When a user starts a bot for the first time, the bot generates a new crypto wallet (a public-private key pair) and associates it with the user’s Telegram account. The user funds this wallet by sending tokens to the generated address.

Once funded, the user can trade by sending commands to the bot. The most basic command is a buy: the user pastes a token contract address, selects an amount, and the bot constructs a swap transaction on the relevant decentralized exchange, signs it with the user’s private key, and broadcasts it to the network. The entire process typically completes in one to three seconds on Solana and three to ten seconds on Ethereum, depending on network congestion.

The bot handles several technical steps that would otherwise require manual interaction. It automatically detects which DEX has liquidity for the token. It calculates the optimal route through liquidity pools, sometimes splitting the trade across multiple pools to reduce price impact. It sets gas parameters to prioritize transaction inclusion. On Ethereum, many bots integrate with block builders and private mempools to avoid sandwich attacks, a form of MEV (maximal extractable value) that front-runs and back-runs a user’s trade to extract profit.

The user interface is entirely within Telegram. Buttons replace the connect-wallet and approve-token steps of a traditional DEX. Portfolio tracking, profit and loss calculations, and token watchlists are all presented as inline messages or callback buttons within the chat.

The major Telegram trading bots

The Telegram bot landscape has consolidated around a handful of dominant platforms, each with different strengths.

Banana Gun is the highest-volume Telegram trading bot as of mid 2026. It operates on Ethereum, Solana, Base, and Blast. Banana Gun is known for its sniping capabilities: the ability to detect a new token listing and execute a buy transaction in the same block as the liquidity addition. The bot charges a 0.5% fee on manual buys and a 1% fee on snipes. Banana Gun processed more than $8 billion in cumulative trading volume in its first year of operation and has generated hundreds of millions in fee revenue, a portion of which is distributed to holders of the BANANA token.

Maestro was one of the earliest Telegram trading bots, launching on Ethereum before expanding to Solana and other chains. Maestro offers sniping, limit orders, copy trading (automatically mirroring the trades of a specified wallet), and anti-rug protection that attempts to detect and front-run liquidity removals. Its fee structure is 1% per transaction.

Unibot launched in mid 2023 and was among the first bots to gain significant traction. It pioneered the revenue-sharing model where a percentage of trading fees is distributed to token holders. Unibot operates primarily on Ethereum and introduced features like private transactions routed through Flashbots to protect against MEV. Trading fees are 1% for non-token-holders and 0.5% for UNIBOT holders.

BONKbot is the dominant Telegram trading bot on the Solana network. Named after the BONK memecoin community, BONKbot specializes in Solana token trading and benefits from Solana’s low transaction fees and fast confirmation times. A trade on BONKbot costs a fraction of a cent in network fees compared to several dollars on Ethereum, making it the preferred tool for high-frequency memecoin trading where traders execute dozens of small trades per day.

Trojan emerged as a competitor to BONKbot on Solana, differentiating itself through a cleaner interface and additional features such as DCA (dollar-cost averaging) orders and multi-wallet management. Trojan has grown rapidly and regularly competes with BONKbot for the top position in Solana trading volume.

What Telegram bots actually cost

The total cost of a Telegram bot trade includes three components: the bot fee, the DEX swap fee, and the network gas fee.

The bot fee is the primary revenue source for the bot operator. It typically ranges from 0.5% to 1% of the trade value. On a $1,000 trade, this means $5 to $10 goes to the bot.

The DEX swap fee is paid to liquidity providers on the underlying decentralized exchange. On Uniswap V3, this is typically 0.3% for established tokens and 1% for newer, lower-liquidity tokens. On Raydium (Solana), the standard fee is 0.25%.

The network gas fee varies dramatically by chain. On Ethereum, a swap transaction costs $3 to $15 depending on network congestion. On Solana, the same transaction costs less than $0.01. On Base, gas fees typically fall between $0.01 and $0.10.

Adding these together, a $1,000 trade on Ethereum through a Telegram bot with a 1% fee costs roughly $10 (bot fee) + $3 (DEX fee at 0.3%) + $5 to $10 (gas) = $18 to $23. The same trade on Solana costs roughly $10 (bot fee) + $2.50 (DEX fee) + $0.01 (gas) = $12.51. These costs are meaningful for small trades. A $100 trade on Ethereum through a Telegram bot loses 18% to 23% of its value to fees before any price movement occurs.

The fee economics explain why Telegram bot trading has concentrated on Solana, where the low gas costs make small, frequent trades economically viable. On Ethereum, Telegram bot trading is more practical for larger position sizes where the fixed gas cost represents a smaller percentage of the trade.

Sniping and launch trading

Sniping is the feature that originally drove adoption of Telegram trading bots. When a new token launches on a DEX, the token creator adds liquidity to a pool. The first trades against that liquidity get the lowest prices. Sniping bots monitor the blockchain for liquidity addition transactions and attempt to place a buy order in the same block.

The technical mechanics differ by chain. On Ethereum, snipers use private transaction channels such as Flashbots or MEV Blocker to submit transactions directly to block builders, bypassing the public mempool where they could be front-run. The bot must predict the exact block in which liquidity will be added and submit a transaction with sufficient gas priority to be included immediately after the liquidity transaction.

On Solana, sniping works differently because the network uses a leader-based block production model rather than a mempool auction. Bots connect to multiple RPC nodes and submit transactions with optimized compute budgets to maximize the probability of early inclusion. The speed competition on Solana is measured in milliseconds, and the leading bots invest heavily in infrastructure co-located with Solana validators. Some bots maintain dedicated connections to multiple validator operators, paying for priority access that shaves tens of milliseconds off submission times. The infrastructure arms race mirrors the high-frequency trading competition in traditional finance, where firms spend millions on co-location and network optimization to gain microsecond advantages.

Sniping carries substantial risk. Many new token launches are scams, rug pulls, or honeypot contracts that allow buying but prevent selling. A successful snipe on a fraudulent token results in a total loss. The anti-rug features offered by bots like Maestro attempt to simulate a sell transaction before executing the buy, checking whether the token contract allows selling. However, sophisticated scam contracts can pass these checks and then enable restrictions after a set number of blocks or a specific volume threshold.

The custody problem

The most significant risk of Telegram trading bots is the custody model. When a user creates a wallet through a Telegram bot, the bot generates the private key and stores it on its servers. The user receives the public address and sometimes can export the private key, but the bot retains a copy.

This means the bot operator has full access to every wallet created through the platform. If the bot’s servers are compromised, every user’s funds are at risk. If the bot operator decides to act maliciously, they can drain every wallet simultaneously. This is the exact opposite of the self-custody principle that decentralized finance was built to enable.

Several Telegram bots have experienced security incidents. In September 2023, Maestro experienced an exploit through a vulnerability in its token approval router contract, resulting in approximately $280,000 in user losses. The bot reimbursed affected users, but the incident highlighted the concentrated risk. In late 2023, Unibot experienced a contract exploit that affected users who had granted token approvals through the bot.

The practical advice for managing this risk is straightforward: treat the Telegram bot wallet as a hot wallet with limited funds. Transfer only the amount needed for immediate trading. Move profits to a hardware wallet or self-custody solution regularly. Never store a significant portion of your portfolio in a Telegram bot wallet. Some traders set a hard rule: never keep more than they can afford to lose entirely in the bot wallet.

Some newer bots have introduced partial mitigations. A few support connecting external wallets through WalletConnect, so the user retains custody of the private key and approves each transaction through their own wallet app. This approach sacrifices speed (each trade requires a manual approval step) but eliminates the custody risk. The tradeoff reflects the fundamental tension in Telegram bot trading: speed and convenience on one side, security and self-custody on the other.

The custodial risk is compounded by the lack of regulatory oversight. Traditional exchanges that hold customer funds are subject to licensing requirements, capital reserves, and regular audits. Telegram trading bots operate outside these frameworks entirely. There is no deposit insurance, no regulatory body to file complaints with, and no legal obligation for the bot operator to maintain solvency or segregate user funds. Users are trusting anonymous or pseudonymous teams with their private keys, and the only recourse in the event of a loss is whatever goodwill or reputational incentive the bot operator feels.

Revenue, tokens, and the bot economy

Telegram trading bots have created a new category of crypto revenue-generating businesses. The fee revenue is substantial: Banana Gun alone has generated more than $100 million in cumulative fees. Several bots have issued tokens that entitle holders to a share of the fee revenue, creating a form of equity-like exposure to the bot’s trading volume.

The token economics vary by project. Banana Gun distributes a percentage of trading fees to BANANA token holders who stake their tokens. Unibot distributes a share of fees to UNIBOT holders. The yield depends on trading volume, which is highly correlated with market sentiment. During bull markets and memecoin frenzies, daily fee revenue can spike by ten times or more. During quiet markets, revenue can drop to a fraction of peak levels.

This volume sensitivity makes Telegram bot tokens among the most volatile assets in crypto. UNIBOT rose from $3 to $200 during its initial hype cycle in 2023, then declined more than 90% before finding a lower range. BANANA experienced similar volatility. Traders who buy bot tokens are effectively making a leveraged bet on future DEX trading volume, particularly memecoin trading volume, which has historically been the most cyclical segment of the crypto market.

The competitive dynamics are intense. Bots compete on speed (fastest execution wins the sniping market), fees (lower fees attract volume-sensitive traders), features (copy trading, limit orders, DCA), and chain coverage (supporting more chains captures more trading activity). The low barriers to entry mean new bots can launch quickly, but the network effects of user adoption and the infrastructure investment required for competitive sniping speeds create meaningful advantages for established players.

The relationship between Telegram bots and decentralized exchange volume is symbiotic. Bots route a significant share of total DEX volume, particularly on Solana where BONKbot and Trojan together have accounted for more than 30% of all Raydium swap volume during peak memecoin periods. This makes bots a critical distribution layer for DEXs, and some DEX protocols have begun offering fee rebates or priority routing to the highest-volume bots. The arrangement benefits both sides: bots get better execution for their users, and DEXs get more volume and fees.

What this does not cover

This guide does not cover the legal and regulatory status of Telegram trading bots, which remains unclear in most jurisdictions and may evolve as regulators examine unregistered trading platforms. It does not cover the specific token contract risks of memecoin trading, including honeypot contracts, hidden mint functions, and transfer tax manipulation, which are the most common causes of loss for Telegram bot users. It does not cover the broader MEV landscape beyond its relevance to Telegram bot users, nor the technical details of Solana validator operation or Ethereum block building that underpin the sniping infrastructure.

Practical checks before using a Telegram trading bot

Check the bot’s track record. Search for past security incidents, contract exploits, or reports of fund losses. A bot that has been operating for more than 12 months without a major incident has passed a meaningful stress test, though past safety does not guarantee future safety.

Check the fee structure. Calculate the total cost of a round-trip trade (buy and sell) including bot fees, DEX fees, and gas on the specific chain you plan to trade. If the total cost exceeds 3% to 5% of your trade size, the fee drag will make it very difficult to trade profitably.

Check the custody model. Determine whether the bot generates and holds your private key, or whether it supports external wallet connections. If the bot holds your key, plan your fund management accordingly and never keep more in the bot wallet than you are prepared to lose.

Check the withdrawal process. Before trading, test a small withdrawal to confirm that you can move funds out of the bot wallet to an external address without delays or restrictions.

Check the bot’s social channels. Active developer communication, regular updates, and transparent incident response are positive signals. A bot with no public developer presence or communication channel is a higher risk.

Disclaimer

This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency investments carry significant risk, and you should conduct your own research before making any investment decisions. Information is accurate as of August 6, 2026.

Bitcoin miners are increasingly positioning themselves as providers of artificial intelligence infrastructure and high-performance computing capacity, but the stock-market reaction to fresh AI hosting deals has cooled markedly in the past couple of years. New research suggests that while contract values are growing, investors are paying less attention to the headline announcement and more attention to what happens next—execution, funding, and long-term profitability.

According to an analysis by Blocksbridge Consulting published in TheEnergyMag’s Miner Weekly, deals tied to AI infrastructure have become less “market-moving.” The report reviewed 25 AI and HPC infrastructure contracts announced between June 2024 and August 2026, finding that the average stock move on announcement day fell from roughly 24% for the earliest deals to about 10% for the most recent ones. Median gains also declined by around half over the same period, even as the reported size and value of the contracts increased.

Key takeaways

• Blocksbridge Consulting’s review shows AI/HPC deal announcement-day reactions weakening from ~24% average moves to ~10% in later deals.
• Median gains from these announcements dropped by about half despite larger contract sizes, implying investors value execution more than upfront figures.
• Revenue per contracted megawatt has inched higher over time, suggesting AI hosting agreements are becoming more financially attractive.
• Major early wins for miners tied to notable AI counterparties produced sharp stock jumps, while newer mega-deals have generated smaller, shorter-lived reactions.
• Investor caution is also visible in infrastructure-focused indices, with TheEnergyMag’s TEM AI Infrastructure Growth Index down ~28.5% from its June peak.

Why AI-hosting news is moving stocks less

The central takeaway from the Blocksbridge Consulting analysis is not that AI hosting deals are shrinking—they appear to be growing in economic importance—but that markets have started to anticipate them. As more miners and infrastructure providers offer similar propositions, investors may treat new contracts as incremental confirmation rather than a sudden re-rating of business prospects.

The report points to a nuanced shift. On one hand, annualized revenue per contracted megawatt has edged upward across the sample, an indicator that AI hosting agreements may be improving in value. On the other hand, the reduced market reaction suggests that investors now scrutinize the substance behind those deals: whether capacity can be delivered on time, how projects are financed, and how durable profitability will be once contract ramp-ups and operational costs are accounted for.

In other words, it’s possible for deals to be economically better while still failing to trigger the same stock enthusiasm as earlier announcements—because expectations adjust. When investors believe execution risk is either higher or more variable than the market used to assume, the “surprise” embedded in contract headlines becomes smaller.

Early deal spikes versus muted mega-deal reactions

The difference between early and later announcements stands out in examples cited alongside the Blocksbridge Consulting findings. According to the report’s examples, initial agreements connected to AI infrastructure sparked dramatic moves for certain miners and hosting operators.

Core Scientific’s initial hosting agreement with CoreWeave reportedly pushed its shares up by more than 40%. Applied Digital’s first CoreWeave lease gained nearly 49%, while TeraWulf’s first Fluidstack deal surged almost 60%.

But as the market has absorbed similar news, later mega-deals have tended to elicit more modest reactions. TeraWulf’s 401-megawatt lease with Anthropic lifted its shares by about 5%. CleanSpark’s $6.6 billion AI hosting agreement reportedly gained nearly 9%. Bitdeer’s new Tydal contract briefly pushed its stock up roughly 12%, but the gains reportedly faded by the close.

That pattern fits the report’s broader conclusion: investors appear more likely to react to earlier “proof points” and less likely to reprice rapidly when a company announces a larger continuation of an established AI hosting strategy. For traders and portfolio managers, the implication is straightforward—volatility around announcements may be structurally lower than it was during the market’s earlier phase of AI infrastructure discovery.

Indexes show momentum slowing, not demand disappearing

The cooling enthusiasm is also reflected beyond individual stock moves. The TheEnergyMag TEM AI Infrastructure Growth Index—tracking publicly traded companies developing AI data center and digital infrastructure businesses—has reportedly fallen about 28.5% from its June peak, even though the index remains sharply higher than a year earlier. The implication is that investors have not abandoned the sector, but they have reduced the intensity of the chase.

The same article notes that the slowdown in these AI infrastructure equities has mirrored broader risk appetite. It cites the Philadelphia Semiconductor Index falling nearly 17% from its July peak, suggesting that part of the recent softness could be tied to sector-wide sentiment rather than purely idiosyncratic execution concerns for specific mining or hosting players.

For Bitcoin miners that have broadened into AI workloads and high-performance computing, this matters because their ability to convert new contracts into steady earnings depends not only on deal economics, but also on the capital markets environment. When AI infrastructure equity momentum slows, lenders and equity investors often become more selective about who can finance expansions and meet delivery timelines—exactly the areas the Blocksbridge analysis implies investors are emphasizing more now.

What investors should watch next

If the market is indeed moving toward a more “disciplined” pricing of AI hosting deals, the next signals will likely be less about the size of the headline contract and more about execution milestones: ramp schedules, delivery progress, and evidence that annualized megawatt economics can hold up as contracts scale. Readers should watch whether announcement-day reactions continue to weaken as deals become more common, or whether new structures—potentially with clearer financing and delivery frameworks—can restore stronger sentiment.

Challenges in construction

Culper Research shorted Nvidia after predicting the $5 trillion AI giant might be re-routing Chinese demand for AI chips through sketchy deals with neighboring countries.

Almost no one believed it at the time, but as it turns out, it called the top.

Despite Nvidia CEO Jensen Huang’s guidance of “assuming zero for China” to comply with US export controls to the country, the company actually benefitted from work-arounds and created big problems for itself in neighboring nations.

“I’m forecasting China’s sales to be zero,” Huang said in November 2025 after US export restrictions halted Nvidia’s chip sales to China. “It’s zero for the next quarter, zero for the quarter after that. We’re assuming it’s going to be zero.”

By May 13, however, Culper Research sniffed a problem with that claim and sold-short Nvidia shares. It sensed that Nvidia might be re-routing its Chinese demand for AI chips via Taiwan and places like Malaysia and Singapore.

It also foresaw legal problems as regulators discovered its diversions.

With the exception of one day immediately following that report, Nvidia’s stock has never closed any day higher than its May 13 close.

Calling the top on Nvidia

It was an unexpected and remarkably accurate call in the middle of a bullish mania. The week prior to the report, Nvidia had rallied 13%, and shares were up an impressive 20% year-to-date. 

Skeptical, Culper Research wrote, “We are short Nvidia for one reason: The company has a significant China problem.”

As it turns out, Nvidia did have massive, unpublished problems in China and neighboring Taiwan. In the three months since that report, those problems became mainstream news.

On July 24, Taiwanese prosecutors searched the home and workplace of an Nvidia employee suspected of smuggling prohibited chips to China. Investigators also went through his desk at the company’s Taipei office. 

It’s the first known legal action against an alleged Nvidia employee in Taiwan’s widening AI chip-smuggling investigation. Prosecutors said the man was “strongly suspected of having committed the offences,” and cited a risk of flight and destruction of evidence.

That story surfaced on July 28. The same day, Jensen Huang quietly sat down with US Commerce Secretary Howard Lutnick in Washington, DC.

When can you best see it?