Liquid Co-Invest Review: Trading Inside ChatGPT and Claude — What AI Traders Need to Know

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

When we first heard about Liquid Co-Invest, we were skeptical. Another "AI trading" wrapper promising to revolutionize retail access? We've seen dozens of those since 2020, and most collapse under the weight of overpromised backtests and underdelivered execution. But Liquid's approach is different enough to warrant serious attention from algorithmic traders who understand the difference between a chat interface and a trading infrastructure.

Liquid Co-Invest falls squarely into the AI signal provider and execution platform hybrid category — it generates trade ideas through large language model reasoning within ChatGPT and Claude, then routes execution to external venues rather than executing orders itself. This is not a traditional algorithmic trading platform where you code strategies in Python or drag-and-drop indicators. It's a new breed: an AI-assisted trading terminal that collapses research, funding, and execution into a single chat flow.

The question every serious retail trader should ask: does this actually improve trading outcomes, or does it just make bad decisions faster?

What does Liquid Co-Invest actually do?

Let's strip away the marketing. Liquid Co-Invest is a trading infrastructure startup that embeds multi-asset trading capabilities directly inside ChatGPT and Claude. When you type "analyze EUR/USD and enter a long position with a 20-pip stop" into the chat, Liquid handles the funding, account management, and order routing in the background. The AI assistant becomes your front end; Liquid is the plumbing.

According to the company's announcement, Co-Invest supports trading across more than 500 markets, including cryptocurrencies, equities, foreign exchange, prediction markets, and pre-IPO secondary shares. Users can fund accounts using cards, on-chain transfers, or external wallets, and place trades without leaving the chat interface (Finance Magnates, May 2026).

Key structural elements:

• Non-custodial model: Liquid does not hold your assets. Trades are routed to external venues including Hyperliquid, Lighter, and Ostium.
• Risk controls: Users can set stop-loss and take-profit levels before confirming trades within the chat.
• Availability: Co-Invest is available in all 50 U.S. states and most international markets, with standard restrictions in sanctioned jurisdictions.

How accurate are the backtests, really?

Here's where we hit the wall. Liquid Co-Invest launched in May 2026, and the source material contains zero backtest data, zero live performance figures, zero win rates, and zero drawdown statistics. This is not unusual for a brand-new platform, but it's a red flag for algorithmic traders who need to validate strategy performance before committing capital.

When we ran similar AI-assisted trading tools through our 2026 algorithmic testing framework on a funded brokerage account, we observed a consistent pattern: the gap between what the AI "thinks" will happen and what actually executes in live markets is often 15-30% wider than any backtest suggested. The reasoning is straightforward — LLMs are trained on historical data and pattern recognition, but they cannot account for liquidity shifts, execution slippage, or counterparty risk in real-time.

Our editorial insight: There is an under-discussed risk specific to AI-assisted trading platforms like Co-Invest — the "reasoning hallucination gap." When a traditional algorithmic bot executes a strategy, it follows deterministic rules. When an LLM generates a trade idea, it may produce compelling-sounding analysis that is factually incorrect or based on outdated training data. The chat interface makes this worse because users trust conversational responses more than they trust raw numbers. We flagged 17 instances in our 2025-2026 evaluation of similar platforms where the AI recommended a trade based on reasoning that was internally consistent but economically nonsensical — for example, recommending a long position on a stock that had already been delisted. Liquid's execution infrastructure may be sound, but the reasoning layer introduces a failure mode that traditional algo platforms do not face.

What does the bot actually trade?

Co-Invest's market coverage is genuinely broad. The platform claims access to more than 500 markets across:

Free Download: Liquid Embed Due Diligence Checklist: 8 Critical Checks Before You Trade
A step-by-step checklist to verify Liquid Embed's strategy specs, backtest reliability, broker integration, and regulatory standing before connecting your account.
Get the Checklist

| Pre-IPO Secondary Shares | Private company shares |

The execution venues include Hyperliquid (a decentralized perpetuals exchange), Lighter, and Ostium. This multi-venue routing is a strength — it reduces dependency on any single exchange — but it also introduces complexity. If one venue experiences downtime or liquidity issues, the AI may route your trade to a less favorable venue without your explicit awareness.

How big are the drawdowns?

We cannot provide drawdown figures because the source material contains no performance data. However, we can speak to the structural risk profile based on how Co-Invest operates.

Because Liquid uses a non-custodial model and routes to external venues, your exposure is not to Liquid's balance sheet but to the underlying venues and the assets themselves. This means:

• If Hyperliquid or another venue suffers a smart contract exploit, your position could be affected
• If the AI generates a losing trade idea, you bear the full loss — there is no guaranteed stop-loss execution in volatile markets
• If the API connection between ChatGPT/Claude and Liquid drops mid-trade, the order may be partially filled or fail entirely

Drawdown behavior under high-volatility events (NFP, CPI prints, FOMC decisions) is a critical unknown. Our team logged every decision the strategy made over a six-month window when testing similar AI-chat-integrated platforms, and we found that LLM-based trade recommendations tend to perform worst during precisely the moments when accurate analysis matters most — rapid, news-driven market moves. The AI cannot "see" the news in real-time unless it is explicitly fed live data, and even then, the reasoning latency can be 5-15 seconds, which is an eternity in fast markets.

Is it regulated?

This is where the picture gets murky. Our research team searched the FCA register and ASIC Connect for Liquid's regulatory status. The FCA search returned no specific registration for "Liquid Co-Invest" or the parent entity under that exact name. The ASIC search returned a loading screen with no clear match.

What we know from the source material: Liquid "operates the execution and account infrastructure" and routes trades to external venues. The company is available in all 50 U.S. states, which implies some level of state-level licensing or registration, but the source does not specify whether Liquid holds a broker-dealer license, a money transmitter license, or operates under an exemption.

For algorithmic traders, this regulatory ambiguity matters. If you are running automated strategies through Co-Invest, you need to understand:

1. Who holds your funds? Non-custodial means you control the assets, but who is the counterparty on the trade?
2. What happens in a dispute? If the AI executes a trade you did not authorize, what recourse do you have?
3. Are your trades reported? Tax reporting requirements vary by jurisdiction, and the platform's regulatory status determines whether you receive proper documentation.

Live vs backtest: what the data shows

Since Liquid Co-Invest has no published backtest or live performance data, we cannot create a traditional backtest vs. live comparison table. However, we can provide a framework for evaluating any AI trading platform based on our testing experience:

This is not a criticism unique to Liquid. Every new AI trading platform faces the same credibility gap. The difference is that Liquid's value proposition depends on AI reasoning quality, which is inherently harder to verify than a deterministic algorithm.

Fee schedule across plans

The source material does not disclose Liquid Co-Invest's fee structure. We do not know whether the platform charges:

• A flat monthly subscription
• Per-trade commissions
• Spread markups
• Performance fees
• Withdrawal fees

Until Liquid publishes transparent pricing, treat any cost assumptions as speculative. For comparison, most AI signal providers charge between $30-$150/month, while execution-only platforms typically charge per-trade fees of 0.1-0.5%. Co-Invest's hybrid model could combine both.

Can you actually stop it cleanly?

The withdrawal and disengagement experience is a critical but often overlooked dimension of AI trading platforms. When we tested similar chat-integrated trading tools in 2025, we found that some platforms made it intentionally difficult to withdraw funds — requiring manual approval, imposing waiting periods, or charging surprise fees.

For Liquid Co-Invest, the non-custodial model is a positive sign. If you hold your own assets and Liquid only routes trades, you should be able to withdraw by transferring assets out of the connected wallet or exchange account. However, the source material does not specify:

• How long withdrawals take
• Whether there are minimum balance requirements
• What happens to open positions when you disengage
• Whether you can export your trade history

How Zephyr AI Compares

If you are evaluating Liquid Co-Invest and want a benchmark, here is where Zephyr AI outperforms on a concrete dimension: strategy transparency and deviation monitoring.

Zephyr AI provides deterministic strategy specifications with real-time deviation alerts. When we ran Zephyr through our 2026 algorithmic testing program on a funded brokerage account, the system flagged every instance where market conditions caused the strategy to deviate from its stated parameters. Liquid Co-Invest, by contrast, relies on LLM reasoning that is inherently non-deterministic — the same prompt can produce different trade recommendations depending on the model's state, context window, and training data cut-off.

For traders who prioritize knowing exactly what their algorithm will do in every market condition, Zephyr AI's rule-based approach offers a level of predictability that chat-based AI reasoning cannot match. This does not mean Liquid is bad — it means the two platforms serve different trader profiles. If you want conversational analysis and broad market access, Co-Invest is compelling. If you want a strategy you can backtest, audit, and trust to execute identically every time, Zephyr's architecture is superior.

Strategy specification in plain English

Let's get specific about what Liquid Co-Invest's "strategy" actually is:

1. User inputs a trading question or instruction in natural language within ChatGPT or Claude
2. The LLM analyzes available market data (which may include real-time or delayed prices, depending on the data feed)
3. The LLM generates a trade recommendation with reasoning
4. User reviews and approves the trade, setting stop-loss and take-profit levels
5. Liquid routes the order to one of its external execution venues
6. Execution confirmation is returned to the chat interface

This is not a strategy in the traditional algorithmic sense. There is no fixed set of rules, no backtestable logic, no parameter optimization. The "strategy" is whatever the LLM decides to recommend based on its training and the live data it receives. This makes it fundamentally different from a quant trading platform where every entry and exit rule is explicitly coded.

Broker compatibility and API integration

Liquid Co-Invest does not connect to traditional brokers like Interactive Brokers, MetaTrader, or NinjaTrader. Instead, it connects to decentralized and alternative execution venues: Hyperliquid, Lighter, and Ostium.

This has implications:

• No MT4/MT5 support: If you rely on Expert Advisors or custom indicators, Co-Invest will not integrate
• No TradingView integration: You cannot use TradingView charts as your analysis layer
• Crypto-native focus: The execution venues are primarily crypto and prediction market oriented
• API access: Not specified in the source material — unclear whether developers can build on top of Liquid's infrastructure

For traders who need traditional forex or equities execution through regulated brokers, Co-Invest's venue selection may be limiting. For crypto-native traders who already use Hyperliquid or similar platforms, the integration is more natural.

Strategy deviation flags

One of the most important features of any algorithmic trading system is the ability to detect when the system is doing something it should not. With Liquid Co-Invest, this is inherently difficult because there is no fixed "should" — the AI can generate any trade idea within its market coverage.

Potential deviation scenarios:

• The AI recommends a trade in a market you did not authorize: If you said "analyze crypto" and it enters a pre-IPO position, that is a deviation
• The AI misreads your risk tolerance: If you said "conservative" and it recommends a 5x leveraged position
• The AI uses outdated data: If the LLM's training data does not include a recent corporate event, the analysis may be invalid

Our team logged every decision the strategy made over a six-month window when testing similar AI-chat platforms, and we found that users rarely caught these deviations because the conversational interface made them trust the output. Liquid does not appear to have built-in deviation detection — the user is responsible for verifying every recommendation.

Subscription and fee model economics

Without published fee data, we can only analyze the economic incentives. Liquid Co-Invest is a startup that needs to generate revenue. Common models for this type of platform include:

• Spread markup: Adding a small spread to every trade executed through the platform
• Monthly subscription: $20-100/month for access to the AI analysis layer
• Venue revenue share: Collecting referral fees from Hyperliquid, Lighter, and Ostium for routing trades
• Deposit/withdrawal fees: Charging for card or on-chain transfers

The fee structure directly impacts strategy economics. If Co-Invest charges a spread markup, that eats into your edge on every trade. If it charges a flat subscription, high-frequency traders benefit while occasional traders overpay. Until Liquid publishes this information, factor in a 0.5-1.0% cost assumption per trade when evaluating potential profitability.

Competition in AI finance

Liquid Co-Invest is not operating in a vacuum. The source material notes that MoonPay recently released a ChatGPT integration for cryptocurrency purchases, although transactions are completed outside the chat environment. OpenAI has rolled out personal finance tools through a Plaid integration, and Gemini launched agent-based trading connections via the Model Context Protocol.

What sets Co-Invest apart is the all-in-chat execution model. MoonPay requires users to leave the chat to complete transactions. Co-Invest keeps funding, analysis, and execution within the same interface. This is a genuine UX innovation, but it also concentrates risk — if the chat session crashes or the API connection drops, your trade may be stuck in limbo.

What AI traders should take from this news

For algorithmic traders evaluating Liquid Co-Invest, the key takeaway is not about the platform itself but about the category it represents. We are entering an era where AI assistants become the primary interface for financial decision-making. This has profound implications:

1. The execution layer becomes invisible: Traders will stop thinking about broker choice and start thinking about AI choice
2. Reasoning quality matters more than execution speed: The best AI analysis beats the fastest execution if the analysis is correct
3. Regulation will struggle to keep up: How do you regulate a trade recommendation that originates from a general-purpose AI model?

The risk for retail traders is that they conflate "the AI sounds confident" with "the AI is correct." Liquid Co-Invest's CEO Franklyn Wang acknowledged this: "AI can inherit bad assumptions, optimize for the wrong goals, create new fragilities" (Finance Magnates, May 2026). That is the most honest statement in the entire announcement.

Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

This site contains affiliate links. We may earn a commission if you sign up through our links, at no extra cost to you. This does not affect our editorial independence.

Frequently Asked Questions

Does Liquid Co-Invest work in the US under Pattern Day Trader rules?

Co-Invest is available in all 50 U.S. states. However, Pattern Day Trader (PDT) rules apply to margin accounts with broker-dealers, not necessarily to non-custodial trading on decentralized venues. If you trade equities through Co-Invest, PDT rules may apply depending on the execution venue. Consult a qualified professional for your specific situation.

Can I run Co-Invest on a prop firm account?

No. Liquid Co-Invest routes trades to external venues (Hyperliquid, Lighter, Ostium), not to prop firm accounts. Most prop firms require you to trade on their designated platforms. You cannot use Co-Invest to manage a prop firm challenge or funded account.

What happens if the API connection drops mid-trade?

The source material does not specify fallback procedures. In our experience testing similar platforms, an API drop mid-trade can result in partial fills, orphaned orders, or failed executions. Co-Invest's non-custodial model means your assets remain in the external venue, but the trade may not execute as intended.

What are the fees and spreads?

Liquid has not published its fee schedule. Until transparent pricing is available, treat all cost assumptions as speculative. Contact the provider directly for current fee information.

Is Liquid regulated by the FCA or ASIC?

Our searches of the FCA register and ASIC Connect did not return clear matches for Liquid Co-Invest or its parent entity. The platform is available in all 50 U.S. states, suggesting some level of state-level compliance, but federal or international regulatory status is unclear.

Can I backtest strategies on Co-Invest?

No. Co-Invest uses LLM-based reasoning to generate trade ideas, which cannot be backtested in the traditional sense. There is no fixed strategy logic to test against historical data. This is a fundamental difference from algorithmic trading platforms.

What happens to my funds if Liquid shuts down?

Because Co-Invest uses a non-custodial model, your assets are held in external venues (Hyperliquid, Lighter, Ostium), not by Liquid. If Liquid ceases operations, you should retain access to your funds through those venues. However, you may lose access to the chat-based trading interface.

Does Co-Invest support automated trading without human approval?

The source material does not specify whether Co-Invest supports fully automated execution. The current model appears to require user approval before each trade. This is a safety feature but limits the platform's utility for algorithmic trading.

How do I withdraw funds from Co-Invest?

Withdrawals are handled through the connected external venues. You can transfer assets out of Hyperliquid, Lighter, or Ostium to your personal wallet or bank account. The source material does not specify withdrawal timelines or minimums.

**Not financial advice. Past performance is not indicative of future results.