AI Portfolio for Two Weeks: 8 Trades an AI Picked for 10,000 USD

This AI portfolio is an experiment we run in public. We asked the AI model GPT-6 Sol to analyse the market and build a portfolio for a 10,000 USD demo account in StocksTrader with a fixed horizon of 10 trading days, using leverage of up to 1:2 on stocks and ETFs and up to 1:50 on currencies and crypto. Seven of the eight positions opened on 28 September after the US market open and closed by 9 October. The AI portfolio earned 148.43 USD, or 1.48% of the deposit. A US 500 position of about the same total size, the benchmark, made 182.50 USD with dividends, or 1.83%.

AI Portfolio Status

How the AI Portfolio Experiment Works

The portfolio was built with GPT-6 Sol, and the AI does the whole selection. It assesses the market, picks the instruments, splits the budget, and calculates the position size, Take Profit and Stop Loss for every trade. Our part is technical: we check the quotes in StocksTrader before sending the orders, open the positions on the demo account and record the result.

The request sent to the AI, translated from the original:

The portfolio follows five rules:

• Fixed horizon. Positions open on Monday 28 September after the US market opens. Everything still open is closed on Friday 9 October after the US open, whatever the result.
• Early exit. If the total profit of the portfolio, counting open positions, reaches 4.5% (450 USD) before 9 October, all remaining positions are closed at once.
• No second attempt. A position closed at its Stop Loss is not reopened within the same portfolio.
• Leverage. Up to 1:2 on stocks and ETFs, up to 1:50 on currencies and crypto.
• Reference prices. Stock and ETF prices are the close on 25 September. Currencies, oil, gold and BTC are the prices at 10:00 GMT on 28 September. The actual entry prices are listed under the positions table.

Market Backdrop for the Next Two Weeks

• US stocks. SPY, a fund that follows large US companies, closed on 25 September at 771 USD, up 0.54%. The market may keep recovering over the next two weeks, though the rise is unlikely to be smooth. New inflation and jobs data could send stocks lower again.
• Fed rate and bonds. On 16 September the Fed raised its rate to a range of 3.75% to 4.00%. The yield on 10-year US government bonds stood at 5.17% a year on 25 September. When bonds pay that much, some investors prefer them to stocks. The portfolio does not count on a rate cut within the next two weeks.
• US dollar. At 10:00 GMT on 28 September one euro bought about 1.1380 USD. The high Fed rate supports the dollar, so the portfolio includes a trade on the euro falling against the dollar. Upcoming inflation and jobs data could change the rate quickly.
• Gold. At 10:00 GMT on 28 September XAU/USD fell to about 4,175 USD per ounce. A strong dollar and high bond yields hold gold back, while tension around the world keeps demand for it alive. With no clear direction, gold stays out of the portfolio.
• Oil. Brent traded near 98 USD a barrel at 10:00 GMT on 28 September after rising on fears of supply disruptions. That can help the oil companies inside the XLE fund. If the tension eases, oil and XLE could both fall quickly.
• Crypto. BTC traded near 83,000 USD on 28 September. Crypto prices can move sharply even at weekends, so the BTC position stays small.

How the AI Split the Budget

• Core block: SPY, XLE and GOOGL. SPY invests in large US companies at once, XLE in energy companies, and GOOGL in Alphabet, which develops cloud services and AI. The three positions add up to about 6,914 USD including leverage.
• Defensive block: XLV and KO. Both are bought without leverage. Healthcare and consumer goods make the portfolio less dependent on technology companies, although they can also fall together with the market. The block takes about 2,500 USD.
• Currency block: sell EUR/USD. The trade sells 4,000 EUR at 1.1380 on the view that the dollar strengthens. Its size is 4,552 USD, and it does not use the full 1:50 leverage available.
• Tactical block: BA and BTC. Boeing shares could react to news on the certification of the 737-10 aircraft, and BTC could rise if demand for crypto recovers. The BA position is 990.35 USD and the BTC position is 831 USD, 1,821.35 USD for the block.

The rest of the deposit stays free as a reserve.

The core block is built for a recovery in US stocks and for oil supporting energy companies. XLV and KO reduce the portfolio's dependence on that scenario. The EUR/USD sell can make money if the dollar strengthens even while stocks are weak. BA is a bet on certification news that may or may not arrive within two weeks, and BTC is a small bet on a rebound, kept small because its price swings sharply.

AI Portfolio Positions

Entry prices below are the reference prices listed in the rules. Take Profit and Stop Loss are the levels the AI set, and they are placed with the orders as they are. Charts for every stock and fund are on the RoboForex stock charts page, and the currency pair is on the EURUSD live chart.

If SPY, XLE, GOOGL, XLV, KO or BA opens at or beyond its Take Profit or Stop Loss on 28 September, that position is not opened, and this page says so in the update.

Actual Entry Prices

Seven of the eight positions opened on 28 September after the US market open, together with the US 500 benchmark. Take Profit and Stop Loss stay at the levels the AI set.

BA was not opened. In the first five minutes of trading, on reports of a software error in the 737 MAX, Boeing shares fell below the planned Stop Loss of 190 USD. The level had already been passed by the time the order was placed, so under the gap rule the position was left out, and the 990.35 USD planned for it stays free.

Potential Profit and Risk of the Portfolio

At the actual entry prices, with BA not opened, if all seven Take Profit levels are reached and the portfolio is not closed early, the profit comes to about 520 USD, or 5.2% of the deposit, before costs. If all seven Stop Loss levels are hit exactly at their prices, the loss comes to about 309 USD, or 3.1%. Price gaps, slippage, the spread and overnight fees can make the actual loss larger.

The total size of all open trades is 14,786.50 USD. Stocks and ETFs account for about 9,405 USD of it, well below the 20,000 USD the account could hold at the maximum leverage of 1:2.

Five dates shape the cycle. The two releases on the economic calendar carry the most risk for the portfolio.

Unexpected figures can make the price jump past a Stop Loss, so the loss on a position can end up larger than planned.

Three Trades With the Best Target-to-Stop Ratio

At the actual entry prices the same three trades keep the best ratio of the distance to the target against the distance to the stop.

• XLE, about 2.11 to 1. At the actual entry of 62.39, the distance to the target is about 2.1 times the distance to the stop. The morning Brent price supports the energy sector, but a sudden easing of tension is a bigger risk for XLE than for any other position.
• KO, about 4.92 to 1. KO opened at 87.26, only 0.76 USD above its Stop Loss, which lifts the ratio and also means a small dip is enough to close the position. It is held without leverage and earns from a different source than technology stocks.
• BTC, about 1.81 to 1. A rebound after the fall could be quick, but round-the-clock trading raises the chance that the stop is filled with slippage. That is why the size is only 0.01 BTC.

AI Portfolio vs the US 500

To check how well the AI built the portfolio, we compare it with a simple alternative: about the same total size in the US 500, 2 lots bought at 7,688.87 at the same moment as the portfolio. That is about 15,378 USD against 14,787 USD in the AI portfolio's open trades. If the AI portfolio closes early at its profit target, the US 500 comparison ends at the same moment, or at the next market open.

Matching the size keeps the comparison fair: both sides put about the same amount of money to work, and the difference in the result comes from what the AI chose to buy and sell.

AI Portfolio Result After 10 Trading Days

The AI portfolio closed on 9 October with a profit of 148.43 USD, or 1.48% of the deposit. The closing rule at 4.5% did not trigger. The US 500 position of about the same size earned 182.50 USD, or 1.83%, including 5.72 USD of dividends, so the index finished 34.07 USD ahead.

Result of Every Position

Results include overnight interest: 20.81 USD in total across the AI portfolio and 36.74 USD on the US 500 position.

Four of the seven open positions made money. The oil trade did the most work: XLE added 88.96 USD as Brent stayed high, and the EUR/USD sell reached its Take Profit as the dollar strengthened, so two parts of the AI's market view played out. The defensive block cost the most: XLV and KO both closed at their Stop Loss levels, 45.72 USD together. KO opened only 0.76 USD above its stop, and a small dip was enough to close it. BTC drifted lower and closed on schedule with a small loss. The AI portfolio ended 34.07 USD, about 0.3 percentage points of the deposit, behind the US 500.

What the First AI Portfolio Showed

The first AI portfolio made money, 148.43 USD on a 10,000 USD deposit in 10 trading days, and the US 500 made more, 182.50 USD. The AI read the market correctly on oil and the dollar, and those two trades brought most of the profit. The defensive block, built to soften a fall in stocks, lost money in a rising market, and that cost the portfolio its lead.

One cycle is too short to judge the AI, and the experiment continues. From the next cycle the AI also sets Take Profit and Stop Loss as a percentage from entry, so a gap at the open no longer changes the risk of a trade. The next AI portfolio is planned for 12 October 2026.

FAQ