Moneda AI does the research. You make the call. Invest in stocks, ETFs and crypto, and hold dollars and euros, in one app. Capital at risk.
Germany
Joined November 2024
- Tweets518
- Following13
- Followers767
- Likes566
Pinned Tweet
Wall Street has analysts. Now you do too.
Moneda AI and investing are live. Ask about any stock, ETF, commodity or crypto. Get an answer built from filings and live prices, with the source for every number.
Then invest in the same app.
It does the homework. You make the call.
There’s no best AI for money.
It’s like asking which doctor is best. You wouldn’t book a heart surgeon for a sprained ankle.
We track 47 finance AI tests. No single model leads them all.
So don’t ask which AI. Ask: good at what?
What do you ask AI about money?
Today we’re launching benchmarks(dot)finance.
One site to track AI models, finance AI products and every finance benchmark we could find. Every score sourced and dated.
Most benchmark sites test coding or general intelligence. Finance needed its own.
ALT The benchmarks.finance homepage. Headline: Finance AI leaderboards, November 2026. It tracks 419 sourced scores for 234 models on 47 finance benchmarks, plus 7 Moneda Bench benchmarks.
One test checks how well an AI reads company filings. Another, live market data. Another, how easily it can be tricked into helping with financial crime.
Every score is sourced and dated. Moneda built the site, and our own AI is on it too.
benchmarks.finance
2
15
The S&P 500 holds 503 stocks. On 30 October, ten of them made up 37.3% of it.
That's by design. The index weights each company by market value, so the biggest companies get the biggest slices.
What's less obvious is how much of the top is tied to AI. Six companies in that top ten make AI chips, rent out AI computing or spend tens of billions a quarter building it. Together, about 26% of the index.
The companies buying AI computing and the companies selling it sit in the same top ten.
Going global changes less than you'd think. MSCI World holds 1,248 stocks from 23 developed countries, yet 72.8% of it is US companies, and all of its top ten are the same US names.
For scale: from 1990 to 2015, the S&P 500's top ten were about 18% to 23% of it. In 2000, at the height of the dot-com boom, about 27%.
The label counts the companies. The weights tell you what you own.
See what's inside your own fund, how weighting works, and how each of these companies describes its link to AI, in our new guide: moneda.com/blog/index-fund-c…
No one can guarantee returns. Moneda AI will tell you that, then show you what the risk looks like for your own profile.
I asked @moneda_com AI what specific shares to take in order to get 15-25% net in the coming years.
He immediately said that no one can guarantee it, the market on average gives about 10%, and on the short horizon it's almost a coin. He looked at my profile (conservative, horizon up to 3 years, the goal is preservation) and honestly wrote: for such goals, the race for 15-25% means sitting in heavy shares and being ready for -20-30% on paper in a bad year.
Offered a balanced option (not as advice, but as a starting point):
35% VOO, 20% QQQ, 15% dividends (type JNJ/PG), 10% VXUS and 20% cash/short treas.
Then he asked how much such a briefcase could fall in a bad year. Chased the backtest in 15 years - the worst drawdown is about 25% (2020 + 2022), the average annual ~11.7%. And he immediately added that there was no 2008 in the window, so in a more serious crisis there may be 30-40%, and 25% is the floor, not the ceiling.
This is one of the options for requests that Moneda AI can respond to in detail.
Realistic assessment, binding to your profile, historical context and honest conversation about risks.
You can ask about portfolios, individual papers, drawdowns, comparisons - and get an answer with sources, not just confident text.
1
5
196
Today we’re launching benchmarks(dot)finance.
One site to track AI models, finance AI products and every finance benchmark we could find. Every score sourced and dated.
Most benchmark sites test coding or general intelligence. Finance needed its own.
ALT The benchmarks.finance homepage. Headline: Finance AI leaderboards, November 2026. It tracks 419 sourced scores for 234 models on 47 finance benchmarks, plus 7 Moneda Bench benchmarks.
We’re also introducing Moneda Bench: seven new benchmarks for the gaps we found.
Live market data. Not making up numbers. Company filings. Cited sources. Portfolio decisions. Explaining finance without giving advice. Questions models haven’t seen.
1
3
123
Every benchmark links to the paper or source behind it, and products are compared feature by feature. Moneda built the site, and our own AI is on it too.
benchmarks.finance
4
99
Moneda retweeted
Do you have a lot of questions about cryptocurrency? AI from Moneda knows all the answers
Three crashes. Three completely different ways to lose money.
2008: the S&P 500 fell 57% and took five and a half years to get back to its old high. Bonds rose about 5%.
2020: the fastest bear market on record. Down 34% in a month, back at a record by August. Bonds rose about 7.5%.
2022: inflation hit 9.1% and the Fed raised rates seven times. Stocks fell 19%. Bonds fell 13%, their worst year since the index started in 1976.
Why? When rates rise, new bonds pay more, so the ones you already own are worth less.
Bonds are usually the part of a portfolio that holds up when stocks fall. In 2008 and 2020, they did. In 2022, they didn’t.
Curious how your own would have done in each one?
Build a test portfolio paper with Moneda AI and replay it through all three: get.moneda.com/launchx
Not investment advice. Capital at risk. Past performance is not a reliable indicator of future results. Availability varies by country.
BREAKING: Intel ($INTC ) is down 29%. The biggest drop on record since its 1971 IPO.
Oct 30, 2026 −28.6%
Aug 2, 2024 −26.1%
Sep 22, 2000 −22.0%
Oct 19, 1987 (Black Monday) −19.2%
Jun 7, 2002 −18.5%
Mar 16, 2020 (COVID crash) −18.0%
ALT Image of a motherboard with an "Intel downside" text overlapped with the weekly price chart of Oct 30th crash.
📈 A tech giant reported revenue up 25% on a year ago.
📉 Its shares fell about 10% the same evening.
The market wasn’t reacting to revenue. It was reacting to cash: the company spent so heavily on AI that more went out than came in.
Earnings are full of moments like this, where good news lands badly and bad news gets cheered. None of it is hidden. It only makes sense once you know how to read it.
Five terms that change how you read an earnings headline:
Guidance: what the company expects next. A great quarter with weak guidance can still sink the shares.
EPS: profit per share. A “beat” means it came in above analysts’ estimates. Beat by less than hoped, and the shares can still fall.
Margin: how much of each sale the company keeps. Sales can rise while margins shrink.
Free cash flow: the cash left after running and investing in the business. A company can make a profit and still burn cash.
Buyback: the company buying its own shares. Fewer shares means higher EPS, even if profit hasn’t grown.
Which earnings term have you nodded along to without being sure what it meant?
ALT A person examines a smartphone displaying an article discussing a tech company's revenue growth and stock decline due to heavy AI spending.
Moneda AI and investing are live. Free to start:
get.moneda.com/launchx
Not investment advice. Capital at risk. Figures shown for illustration. Availability varies by country.
3
75