Thomas Peterffy — Optimistic On AI, Calm About Yields, But Own Land
"I am buying land so that after the democratic socialists take everything over and they ruin all the stuff… the only assets they will be able to return will be their own land."
One-line take: Recorded the day before the September FOMC. Conflict of interest up front: Peterffy founded and chairs Interactive Brokers, still owns ~75% of it, and IBKR sponsors the podcast — every IBKR comment is the insider's pitch. Rates: IBKR's spread is fixed (lend ½% over fed funds, pay ½% under), so the level doesn't matter to the business; he thinks inflation is an oil/war problem (stop Iran threatening the Strait of Hormuz) and would rather run the economy hot, but expects Warsh's Fed to hike anyway "to demonstrate that they are independent." Calm about the 10-year crossing 5% — fine as long as the rise stays gradual. AI: "very optimistic" — Nvidia at 27× earnings with 70% expected growth is "a relatively low valuation"; the bigger payoff is AI adoption (incl. open-source models) lifting productivity everywhere; hyperscalers are in a winner-take-most compute race that will crush compute prices and force write-downs; don't slow down — China won't. Prediction markets: his 10-year-old phantom-money product inspired Kalshi and Polymarket; he tried to buy Kalshi; ForecastTrader is financial/economic, not sports; he expects the majority of IBKR's earnings to come from prediction markets "in the fullness of time"; the Supreme Court decides next spring whether sports contracts are swaps. Political markets say the House flips Democratic, the Senate stays Republican. Risk: system-wide OTC leverage is opaque "spaghetti," but he's not worried now. Land: puts a 20% chance on a US turn to democratic socialism; after Hungary's communists seized everything, only agricultural land came back — so he's buying land, and gold is no substitute (it can be confiscated).
1. Stocks & names mentioned
Sponsors read by the host (LSEG, World Gold Council, BNY Investments) are excluded. IBKR is the guest's own company — he is its chairman and ~75% owner.
| Ticker | Name | Research | View | What Peterffy said | At |
|---|
| IBKR | Interactive Brokers Group | QT · SA · STK · FA | Positive | Insider view (founder/chairman, ~75% owner): ~40%/yr earnings growth since 2022 comes from a platform built for professional traders — "it is not enough to pick the right investment strategy… they also must pick the right platform." Rate-level-neutral spread, AI models plugged into accounts, a bank charter to custody ETFs/mutual funds and grow securities lending, and prediction markets as the future majority of earnings. "I don't think about the stock much" (P/E ~12× → ~30×). | 3:14 |
| NVDA | NVIDIA | QT · SA · STK · FA | Positive | "If you look at a company like Nvidia, 27 times earnings, and they stated expectations of increasing earnings by 70%, that is a relatively low valuation" — high historical multiples are cheap against AI-driven earnings growth. | 12:01 |
| Farmland | Agricultural / rural land (real asset) | — | Positive | He is buying land as insurance against a US turn to democratic socialism (20% odds): in communist Hungary only agricultural land "they couldn't basically ruin" was returned after 45 years — "land is the logical asset" for your heirs. | 41:56 |
| Kalshi | Kalshi (private — CFTC-regulated prediction market) | — | Neutral | Copied IBKR's phantom-money prediction market, got a CFTC license first ("I said what a stupid thing it is I didn't think about doing that"), and refused his takeover offer. Mostly sports volume — "a huge amount of money" — but the states are up in arms and the Supreme Court decides next spring whether those contracts are swaps or state-licensed bets. | 18:22 |
| Polymarket | Polymarket (private — prediction market) | — | Neutral | Named alongside Kalshi as having seen IBKR's phantom product and launched it "for real money" — "now they are way ahead of us of course." | 18:22 |
| NTRS | Northern Trust | QT · SA · STK · FA | Neutral | Named by the host as an incumbent custodian IBKR's new bank charter targets; Peterffy: "We offer everything that they offer… we do it better" — online short inventory and lending rates undercut banks (3% vs IBKR's 2%). | 38:04 |
| STT | State Street | QT · SA · STK · FA | Neutral | Passing: one of "the big custodians of this world" IBKR plans to compete with for ETF and mutual-fund custody once chartered. | 38:04 |
| JPM | JPMorgan Chase | QT · SA · STK · FA | Neutral | Passing: listed with Northern Trust and State Street as a big custodian; also the kind of bank whose OTC exposure he flags as the system's opaque leverage risk. | 38:04 |
| Hyperscalers | AI hyperscalers (big cloud / AI capex spenders) | — | Negative | Each must "buy all the compute capacity they can" to be one of the one or two winners; that will drive the price of compute down, "benefit tremendously everybody else," and "these companies of course will have to write down much of the compute that they bought." | 14:06 |
| GLD | Gold (SPDR Gold Shares — asset-class proxy) | QT · SA · STK | Negative | Not an alternative to land as a confiscation hedge: "gold you cannot carry with you and they will immediately say that ownership of gold is illegal" and make you hand it in. (Asset-class call; GLD as proxy.) | 43:44 |
2. Talking points
2:45 What drove 40%/yr earnings growth
- IBKR began as a platform for floor traders going electronic — "way too early"; the 2000–2010 conversion brought its first target clients ($0.5–5M daily traders) who needed multi-order management, arbitrage, pairs, shorting, algos and best execution.
- Still aimed at the professional trader; growth comes from investors learning that the platform matters as much as the strategy.
7:01 Rates don't matter to IBKR; inflation is an oil problem
- IBKR lends at fed funds +½ and pays fed funds −½, so "we don't care if fed funds are 5% or 10% or 20% or 1%."
- Inflation is driven by oil, and oil by the war — stopping Iran's threat to the Strait of Hormuz matters more than a hike; better to run the economy hot until the war ends, then see if cheaper oil cures inflation.
8:56 The Fed will hike "to demonstrate that they are independent"
- With the 10-year through 5% and the 30-year higher, he expects a hike he wouldn't favor.
- Not worried about yields: only a very sharp sudden rise would cause problems; the move since August is still gradual.
11:20 AI valuations are low against earnings growth
- Nvidia at 27× earnings with 70% expected earnings growth is a relatively low valuation; similar, smaller effects everywhere as AI permeates the economy.
- Even if frontier models stopped improving, open-source adoption across all companies would drive "huge increases in earnings and productivity" — "I'm basically very optimistic."
13:36 Hyperscalers: a compute land-grab ending in write-downs
- Only one or two will prevail, so each buys all the compute it can; compute prices will fall, benefiting everyone else, and the buyers will write much of it down.
14:47 Regulation yes, slowing down no
- China won't slow down — "a competition between the two hemispheres… I would like us to win."
- Doesn't believe in a "Hindenburg moment" from a hack; hacks have been around a long time. Assume you can be hacked and limit damage before discovery.
17:15 The phantom prediction market that inspired Kalshi
- Built ~10 years ago with phantom money because consultants warned it would block a banking license; Kalshi and Polymarket saw it and launched with real money.
- Tried to buy Kalshi after its CFTC license (rather than wait 2.5–3 years) — rebuffed; IBKR got its own license. ForecastTrader launched 2024, expanded 2025.
19:40 Prediction markets as the missing context for stock prices
- Not sports: markets only price assets, but companies live inside regional, national and global economies and a social/climate environment — questions investors need priced.
22:22 Not yet liquid — but "going to become huge"
- Hedgers exposed to these questions will pile in; experts will trade their views instead of selling them.
- Today IBKR is more exposed as market-maker — "not a very attractive business at the moment."
24:49 Kalshi's sports risk goes to the Supreme Court
- States oppose sports contracts; a ruling expected late next spring decides swaps vs state-licensed bets. "If we had bought Kalshi, they wouldn't be doing sports."
26:41 Prediction markets → majority of IBKR earnings
- "In the fullness of time… the majority of our earnings will be derived from prediction markets."
27:09 Midterms: House flips, Senate holds
- Political markets are deep enough and "more accurate than the polls" — they point to a Democratic House and a Republican Senate; the odds "don't look good" for Republicans.
29:00 Leverage: fine inside IBKR, opaque outside it
- IBKR auto-liquidates any margin violation; the worry is system-wide leverage — large margin loans and bank OTC contracts that are "a spaghetti" nobody can see centrally.
- Rising yields in the US, Japan, UK and yen moves: "a little bit of worry" but "I'm not really worried about it at this time."
32:41 AI helps IBKR — but can't forecast earnings
- All major AI models connect to client portfolios on the platform: a positive, not a threat.
- Frozen weights mean models "cannot learn" new information, so they won't transform securities analysis; prediction markets are "much more important for projecting earnings forward."
36:35 Bank trust charter → custody and securities lending
- A charter lets IBKR custody ETF and mutual-fund assets, expanding shares to lend — short-sellers are a large part of interest income.
- Versus Northern Trust, State Street and JPMorgan: transparent online borrow inventory and rates; banks match IBKR's 2% and send it business.
39:17 The multiple went 12× → 30× — "I don't think about the stock much"
- Focus is customer profitability and new products; "there's nothing I can do about" the stock price.
40:55 Why he buys land: Hungary's lesson
- Communists seized everything; after 45 years only agricultural land could be returned. He sees the same risk in the US, putting a democratic-socialist takeover (the "Mamdanis of the world") at 20%.
- Gold is no alternative — you can't carry it and it can be declared illegal. Not an inflation call: it's about leaving heirs something.
45:04 Advice: have a plan — and learn AI
- Always have a plan, open to change with new information; know each morning what to do. Learn as much about AI as possible — "the song of the future."
3. In plain English
A jargon-free summary of the thesis behind each call — what it actually is and why he makes it. (Plain-language companion to the table above; renders on each consolidated page.)
IBKR — Interactive Brokers Positive
Interactive Brokers is an online brokerage built for heavy, sophisticated traders — cheap trading, low borrowing costs and good pay on idle cash. Peterffy founded it, chairs it and still owns about three-quarters of it, so his enthusiasm is an owner's pitch, not an independent opinion.
His case: its profit on client cash is a fixed slice (it charges half a point above the Fed's rate and pays half a point below), so it earns whether rates are high or low. Next legs are AI tools connected to client accounts, a bank trust charter that lets it hold fund assets and lend more shares to short-sellers, and prediction markets (bets on economic and political outcomes), which he expects to become most of its profit eventually. The shares have re-rated from about 12 to about 30 times earnings.
NVDA — NVIDIA Positive
Nvidia makes the chips that train and run AI. Peterffy's point is simple math: a price of 27 times earnings sounds expensive, but if earnings grow 70% next year the stock gets cheap very fast. He thinks the broad market's high valuations are justified the same way, as AI spreads into every company's productivity.
Farmland — agricultural land Positive
This isn't an inflation trade. Peterffy grew up in communist Hungary, where the state took his family's land, homes and businesses. When communism fell, buildings and companies had been ruined — but farmland couldn't be, so it was handed back. He puts a 20% chance on the US drifting toward democratic socialism, and he is buying land as the one asset he thinks his heirs would eventually get back.
Kalshi — prediction market Neutral
Kalshi is a private, federally regulated exchange where people trade yes/no contracts on future events. Peterffy says it copied his earlier play-money version, got licensed first, and turned down his offer to buy it. Most of its money comes from sports bets, which states are fighting in court; the Supreme Court is expected to decide next spring whether those are federally regulated financial contracts or gambling that needs state licenses — a big risk to its business model.
Hyperscalers — the big AI spenders Negative
"Hyperscalers" are the giant cloud companies pouring hundreds of billions into AI data centers. Peterffy thinks each believes only one or two will win, so all of them buy every chip they can. That glut will eventually make computing power cheap — great for everyone who uses AI, bad for the buyers, who will have to write down the value of hardware they overpaid for. He is bullish on AI, not on the returns of the companies funding the build-out.
GLD — Gold Negative
Asked whether gold would be a cheaper way to protect against a government seizing private wealth, Peterffy said no: the government can simply make owning gold illegal and demand you hand it over, and you can't carry it away. This is a narrow view about gold as confiscation insurance, not a call on the gold price. GLD is a proxy fund.
Summary & timestamps derived from the public YouTube video (transcript in transcript.txt) for personal study. Interactive Brokers sponsors the podcast and the guest is its chairman and controlling shareholder. Not investment advice. © Master Investor Limited / Paradine Productions for source material.