In a rare public conversation at the Global Alternative Investment Conference in New York, Seth Klarman – the famously reclusive author of Margin of Safety and founder of Baupost Group – sat down with CNBC for an extended interview. Over nearly 44 years, Baupost has posted losses in just five calendar years, with its worst drawdown around –10%. Klarman shared his current market views, investment principles, and the disciplined mindset that has defined his career.

This is not a transcript but a distilled summary of his key insights, organised thematically.

1. The Baupost DNA: Downside First, Always

Baupost started in 1982 as an extended family office, with $27 million from a few families who had sold their television station and a computer‑publishing business. From day one, the goal was never asset‑gathering – it was long‑term, risk‑conscious returns for those original clients.

That mission still drives everything. Klarman emphasises:

• Downside protection is weighed equally with upside potential.
• Exhaustive fundamental research on every position.
• A willingness to hold cash when opportunities are scarce.
• No portfolio leverage.
• A bias toward senior securities (debt instruments, structured private deals) that offer contractual priority in case of trouble.
• Selective macro hedging.

The firm is truly multi‑asset: public equities, credit, private investments, and real estate. Capital flows bottom‑up to wherever the best risk‑adjusted opportunities appear.

2. Is This a Bubble? “It Has Some Bubble‑Like Characteristics”

Klarman acknowledges the current market has hallmarks of a classic “new era” narrative – euphoria around a transformative technology, with examples like a shoe company getting a stock bump just by adding “AI” to its name. That echoes the dot‑com era.

Yet he also concedes that AI is genuinely game‑changing, which makes it hard to dismiss outright. The real difficulty is the uncertainty AI creates:

• No one knows who the lasting winners will be.
• Will it be winner‑take‑all, or will many players thrive?
• What price should investors pay for an unknown future?

The market now divides companies into three buckets: AI winners (everyone wants them), AI losers (everyone avoids them), and AI‑irrelevant (ignored, and thus slowly becoming cheaper). Klarman spends most of his research time on the third bucket – companies that are being neglected – and on some “losers” that may not actually be losers.

In credit markets, for example, many software‑related debt securities are heavily discounted. Klarman finds that worth studying because the implied multiples are extremely low relative to cash flows.

3. What Value Investing Really Means – and Why It’s Not Just Cheap Multiples

Klarman rejects the mechanical definition of value investing as buying low P/E or low P/B stocks. For him, true value investing means asking what a business is actually worth. Growth matters, but today many businesses can deteriorate faster than in the past – operational issues quickly erode cash flows and competitive positions.

His core question in every analysis is: “What can we actually know with confidence?” In a frothy market where investors pay 40x earnings – or even infinite multiples for some AI names – you have to be extraordinarily certain about a very distant future. Klarman sees that as virtually impossible. The irony is that higher uncertainty should demand lower multiples as a margin of safety, yet multiples have expanded. Thus, even if AI meets the most optimistic forecasts, today’s pricing may still be bubbly.

4. How Baupost Actually Plays AI – Not the Megacap Model Companies

Klarman holds Amazon and Alphabet (Google), which together represent about 10% of the portfolio that could benefit directly from AI. He likes them as massive cash‑flow machines with diversified businesses and adaptability. He bought Google years ago when its multiple lagged the broader market – a classic value entry.

More interesting: Baupost owns undeveloped land with access to power, which could be used for data centres. They partner with an experienced operator to identify and secure sites, creating valuable optionality. They also made a private investment in Asian data‑centre assets at a significant discount (roughly 40% of the public market price) via a spin‑off from a Chinese company.

As for the large language‑model giants like OpenAI or Anthropic, Klarman has stayed away. He questions whether it’s winner‑take‑all, whether these firms will ever achieve sustainable profitability given their voracious cash needs for training, and whether they fit Buffett’s definition of a “great business.” Yet he keeps an open mind – he recalls Eric Schmidt’s warning not to underestimate AI, and admits AI may ultimately exceed even today’s rosy projections.

5. His Three Personal Rules for AI Engagement

Klarman set three boundaries for himself:

1. Don’t be on the cutting edge – he’s not a technologist.
2. But stay informed – he reads broadly and listens to podcasts to avoid an information disadvantage.
3. Stay focused – he doesn’t want to miss a “fat pitch” in areas he understands, whether that’s an AI‑irrelevant stock, a distressed debt situation, or real estate.

6. Where the Real Opportunities Are Today

Distressed Debt – Klarman sees more event‑driven opportunities appearing: corporate restructurings, debt exchanges, and distressed private‑equity deals. When bonds are sold indiscriminately (on downgrades, bankruptcies, or forced selling), new capital can step in at steep discounts and wait for the turnaround.

Commercial Real Estate – Senior Living Facilities – Post‑COVID, this sector suffered severely, with many bankruptcies and low occupancy rates. Klarman believes the industry is in the early stages of recovery, making it his single favourite idea right now. He also likes industrial land, warehouses, and cold‑storage properties, as domestic manufacturing and supply‑chain shifts drive strong demand.

Data‑Centre Land – As noted, Baupost is accumulating land with power access, not necessarily to build data centres themselves, but to supply capacity to hyperscalers.

7. The Big Risks the Market Is Underpricing

Klarman is blunt: markets are underestimating multiple risks.

• Geopolitical tension – especially around the Strait of Hormuz. A prolonged closure could send oil prices well above $100, potentially to $150 or more, with severe knock‑on effects.
• US fiscal trajectory – national debt has reached 100% of GDP and is growing by over $2 trillion annually. While not an immediate crisis, the trajectory is alarming, and global confidence in US “risk‑free” assets is eroding.
• Policy recklessness – he criticises the chaotic approach to tariffs, the on‑again/off‑again nature of some announcements, and the lack of strategic thinking in foreign policy. He specifically questions the recent Iran conflict, arguing it was not a carefully thought‑out plan with a clear endgame.

Klarman notes that overseas perceptions of the US have shifted – from a stable, reliable partner to a less serious actor. He points to how the US treats its neighbours (Canada, Mexico) and allies as examples of unnecessary self‑inflicted damage.

8. The Fed and Inflation – A Dovish Bias Underneath

While AI‑related construction may be inflationary, Klarman believes the Federal Reserve (under Kevin Warsh, should he become chair) would prefer to cut rates over time. He expects a cautious, data‑dependent approach, with perhaps one or two more hikes, but ultimately sees AI as a potentially deflationary force that could give the Fed room to ease.

9. One Regret – The One That Got Away

When asked about a missed opportunity, Klarman cites Palantir. About 15–20 years ago, he received a call about a potential secondary sale of $40–50 million of venture shares. He did extensive due diligence and was ready to bid, but the seller changed their mind. If that deal had closed, it would have generated billions in profits. He shrugs it off as a tale every investor has.

10. The Investment Process and Succession

Baupost’s 40‑person investment team is divided into four silos: public equity, credit, private investments, and real estate. Klarman stays deeply involved, attending meetings and maintaining a real‑time flow of ideas. Capital allocation is strictly bottom‑up – they do not set top‑down targets but let opportunities dictate where money goes.

He reads widely and listens to thought leaders like Niall Ferguson, Dario Amodei (Anthropic), Jamie Dimon, and Warren Buffett – not for stock tips, but for perspectives on markets and the broader moment.

At 60‑something, Klarman expects to remain at the helm for another 10–12 years, actively delegating and mentoring. He insists he will step aside when he is no longer the best person to lead the firm.

Klarman’s rare public remarks offer a masterclass in intellectual humility. In an era of AI frenzy, he doesn’t claim to know where the technology is heading – instead, he doubles down on what he can know: the value of a business, the importance of downside protection, and the discipline to wait. His 44‑year track record is not a product of bold predictions, but of patient, rigorous, and diversified risk management. For anyone navigating today’s noisy markets, his enduring message is clear: start with what you can actually know – and price in the rest with a wide margin of safety.