AI investing: the hype vs what the evidence shows
79% of pro fund managers lost to the index in 2025

79% of active large-cap US fund managers underperformed the S&P 500 in 2025.
professionals, paid to do this full time.
and the hot idea this year is that a free chatbot will beat them for you.
that's AI investing in one line. the money pouring INTO AI is real. the idea that AI makes you a better investor is mostly hype.
The consensus: AI investing is the easy win
the pitch comes in three flavours:
- buy AI stocks, because the spending is enormous
- use robo-advice, because algorithms don't panic
- ask ChatGPT what to buy, because it has read everything
each has a grain of truth. so let's look at the evidence.
What's actually happening with AI stocks
the spending is genuinely huge. Alphabet, Amazon, Meta and Microsoft are on track to spend about $650 billion on AI in 2026, according to Bloomberg estimates reported by eWeek.
but look at how markets reacted:
→ the four lost a combined market value of over $950 billion after announcing those plans (eWeek), on doubts the spending will pay off → in early February, US software stocks were hit by fears that a new Anthropic AI tool would eat into their business → in December, the Bank of England warned that US stock valuations were close to their most stretched since the dot-com bubble, with AI-focused tech shares at risk of a sharp correction
so "AI is the future" and "AI stocks are a safe bet" are two very different claims. the first can be true whilst the second blows up your portfolio.
meanwhile retail investors are piling into AI tools. an eToro survey of 11,000 retail investors, reported by Reuters, found 13% already use tools like ChatGPT or Gemini to pick or change investments, and about half would consider it. the same report cites a forecast of robo-advisory revenue growing from $61.75 billion to $470.91 billion by 2029.
What the evidence says about AI stock picks
this is where the hype runs out.
- beating the market is brutally hard. the latest SPIVA scorecard found 79% of active large-cap US funds trailed the S&P 500 in 2025, up from 65% in 2024
- chatbots don't owe you anything. Europe's markets regulator ESMA warned that public AI tools can give inaccurate or misleading investment advice, and have no obligation to act in your best interest
- generic models have blind spots. eToro's UK managing director told Reuters that general AI models can misquote figures and dates, favour popular narratives and lean too hard on past price moves
- feeling smarter isn't being smarter. we saw the same trap with coding tools: developers felt faster with AI and were actually slower
and just like AI health chatbots, the model can "know" plenty whilst the human using it still makes the wrong call.
bad: "ChatGPT, which stocks will go up?" good: "summarise this company's last annual report, list the risks, then argue the bear case" (the case for the price falling)
What a finance degree won't teach you about AI investing
finance courses cover valuation, portfolio theory and the efficient markets debate. what's usually missing:
- how to audit an AI-generated analysis, number by number, against the actual filing
- how hype cycles show up in valuations today, beyond one lecture on the dot-com bubble
- what a robo-adviser actually does under the hood, and why that's a different thing from AI picking winners
The critical-thinking move: 5 questions for any AI investing claim
- what's the benchmark? if it isn't compared with a plain index fund, it means nothing. 79% of pros lost to one last year
- who profits if I believe this? platforms, fund sellers and app makers all benefit from AI excitement
- is the number from the filing or the chatbot? check every figure against the company's own report
- what's already priced in? a great AI company can still be a bad investment at the wrong price
- what happens in a downturn? a strategy that only works when everything goes up isn't a strategy
What to do this month
- read the SPIVA coverage and pull out two numbers you'd use in an interview
- run a chatbot audit. ask an AI to summarise one company's annual report, then check five figures against the real document
- write a one-page bear case for one big AI stock. it's much harder than the bull case, and much more useful
- read the ESMA factsheet on using AI for investing. it takes ten minutes
we'll keep tracking this on our finance hub.
so TLDR: AI is reshaping markets but it isn't a shortcut to beating them, so treat it as a research assistant, not a stock picker.
if an AI told you to buy a stock tomorrow, what's the first thing you'd check?
Sources
- Active managers stumble again in 2025 as large caps dominateInvestmentNews (reporting the S&P SPIVA U.S. Scorecard) · 04 MAR 2026
- Google, Microsoft, Meta, and Amazon Plan $650 Billion AI Spending Push in 2026eWeek · 09 FEB 2026
- U.S. software stocks hit by Anthropic wake-up call on AI disruptionReuters (via CTV News) · 04 FEB 2026
- US tech valuations close to most stretched since dot-com bubble, Bank warnsLBC (Press Association) · 02 DEC 2025
- 'ChatGPT, what stocks should I buy?' AI fuels boom in robo-advisory marketReuters (via The Star) · 25 SEP 2025
- Using Artificial Intelligence for Investing: What you should considerEuropean Securities and Markets Authority (ESMA) · 25 MAR 2025
Get the next issue by email
theVANTAGEd is written by students to bring critical thinking to AI across real industries. One issue every two weeks.
