Deep Values

The Shanghai Composite looks stuck near 3,000, yet its history shows 9% real annual returns against 6.6% for the US. With the Dow up 200% since 2020, learn why index math can mislead you about which market is really working. Chapters (0:00) The Great Divergence (1:00) American Surge vs Chinese Stagnation (4:03) Dow Jones Index Mechanics (6:02) US Value Engine vs A-Share Financing (6:41) A-Share Ma...

Deep value research on United Airlines (UAL) as of 2026-11-05: $80B list-price plane order. With revenue near zero in 2020, Scott Kirby placed the largest aircraft order in aviation history. We run that bet through a Graham-Buffett lens to test whether a century-old capital destroyer can finally compound. Chapters (0:00) Confidential Blueprint (1:20) Asymmetric COVID Bet (4:09) 30-Year Horizon (5:...

Musk called saving pointless, then admitted 72 hours later that serving Grok cost too much. See why $38 billion data centers make cash flow and moats the real test of the AI abundance story. Chapters (0:00) The Abundance Paradox (3:39) Collapsing Cost of Intelligence (5:01) Two Realities of AI (6:01) Grok Bot Unit Economics (7:25) Capex Bottleneck Reality (8:07) Grok Bot Unit Economics (8:48) Cape...

See why AI's bottleneck moved from chips to power, land and grid access. With 438 GW of requests queued in Texas alone, who owns those scarce assets matters more than who buys the GPUs. Chapters (0:00) Power Is the New Compute (1:01) The Bottleneck Has Moved Upstream (1:43) Collision of Constraints in Texas (2:21) AI Physical-Economy Value Ladder (4:23) The Bottleneck Has Moved Upstream (5:28) Lan...

Why a 97-year-old says he would rot without a task, and what people past 90 regret overpaying for. Time is the one asset you can't buy back, and you are pricing it today. Chapters (0:00) The Ultimate Long-Term Hold (2:58) Investment Verdict on Human Capital (3:57) The Ultimate Long-Term Hold (5:01) Daily Operating Engine (5:41) Attitude as an Economic Moat (6:41) Key Operational Metrics (8:03) Hid...

Bearish on Apple (AAPL) as of 2026-11-04: 12.6% yearly profit growth. See why a legendary value investor is trimming Apple to fund Google and PDD, and how to test any great business against its price. With Treasuries at 5.2%, a $5 trillion Apple needs 12.6% annual profit growth just to match bonds. Chapters (0:00) Deep Value Research Note (1:55) Apple Valuation Decline (3:58) Apple to Google Shift...

See how Graham and Buffett would judge Anthropic's $80 billion six-month loss and Google's $500 billion wipeout. With Anthropic's S-1 filing out and AI spending soaring, the margin of safety question can't wait. Chapters (0:00) AI Capital Cycle Analysis (1:40) Financial Reality of 2026 (2:23) AI Paradox Bull Bear Case (4:07) AI Capital Cycle Analysis (4:47) AI Paradox Bull Bear Case (5:48) Financi...

The viral 1875 chart marks 2026 as a sell year, yet its sell dates are just a repeating 9-10-8 year loop. See what its record really shows, and why price versus value matters more than any calendar. Chapters (0:00) Introduction to the Benner Cycle (1:23) The Arithmetic Engine Explained (2:40) Demystifying the Internal Math (5:21) Agricultural Boom-Bust Flywheel (8:25) Evidence of Famous Market Hit...

Bearish on Intel (INTC) as of 2026-11-20: -372% Intel margin of safety. Intel trades at $95.53, but our base intrinsic value estimate is $20.24. That gap means the market price assumes 36.2% annual owner-earnings growth for five years. SanDisk is the opposite case: it trades close to fair value, but only if today's NAND margins hold up through the cycle. Here is why neither one offers a margin of...

Neutral on Roblox (RBLX) as of 2026-11-19: 83% FCF lost to SBC. Roblox reported $1.35B of free cash flow in FY2025 and a $1.07B GAAP net loss in the same year. Subtract the $1.13B of stock-based compensation and only $224M is left, so we look at what that means for owner earnings before today's Q3 2026 numbers come out. Chapters (0:00) Roblox Deep Value Research Overview (2:16) The Creator-User Ne...

Neutral on Amazon (company) (AMZN) as of 2026-11-18: -$42.4B TTM free cash flow. Amazon brought in $168.3B of operating cash flow over the last 12 months, yet reported free cash flow was -$42.4B. Here's how one call, whether AI capex counts as maintenance or growth, moves the estimated intrinsic value from $25.68 to about $199.53, and which return metric to track from here. Chapters (0:00) Am...

Neutral on Alphabet (GOOGL) as of 2026-11-17: -$9.1B Q3 free cash flow. Alphabet generated $42.3B of operating cash in Q3 2026, spent $51.4B on AI capex, stopped buybacks and took on $29B of new debt. We compare its owner earnings, balance sheet and cloud growth with Microsoft and Meta to see whether this AI build can earn its cost of capital. Chapters (0:00) The AI Capital Regime Change (1:03) GA...

Neutral on Microsoft (MSFT) as of 2026-11-16: -70% FCF collapse. Microsoft spent 49% of its quarterly revenue on AI data centers, and free cash flow fell 70%. We work out its owner earnings, run a Graham-Buffett DCF and show why a Tier A business can still fail the price test. Chapters (0:00) Microsoft Classification and Stance (1:00) The Thesis: Quality vs Price (2:06) Business Model and Monetiza...

Neutral on Alphabet (GOOGL) as of 2026-11-15: $205B FY2026 capex. Google Cloud is growing 81.8% a year, but a $205B capex program pushed Alphabet's Q2 free cash flow below zero, and its $9.11 headline EPS falls to $2.85 once paper gains are stripped out. We look at owner earnings, the depreciation flag and a reverse DCF to see how $320.58 compares with our estimated intrinsic-value range. Chapters...

We take apart Anthropic's 416-page IPO prospectus with Damodaran's valuation: revenue up 1,400%, but $473B in compute commitments that work like debt, a BB+ rating and a five-class share structure. His optimistic base case puts intrinsic value at about $1.48T, while expected IPO pricing of $1.5T to $2T leaves no margin of safety. Chapters (0:00) Anthropic's IPO: A Deep Value X-Ray (0:46) Great Tec...

The Fed held rates at 3.75%-4.00%, but two governors dissented and voted for another hike, and Chair Warsh says conditions still aren't tight. Here is what that 10-2 split means for discount rates, margin of safety and companies with heavy debt. Chapters (0:00) FOMC Rate Pause Big Picture (2:21) Target Rate and 10-2 Vote (3:45) Strong Economy, Stubborn Inflation (4:25) Key Risks: The Bear Cas...

We read Anthropic's S-1 with a Graham-Dodd lens: a 3,300% headline growth rate, a growth slowdown in Q3, more than $800B in compute commitments and a single share that gives seven founders control. Then we look at where autonomous-vehicle economics look most durable, at least under the episode's assumptions. Chapters (0:00) A Deep Value Lens on AI and AVs (1:05) Generative Euphoria vs Physical AI...

Bearish on Robinhood Markets (HOOD) as of 2026-11-11: -28% Margin of safety. We audit Robinhood's Q3 2026 numbers: $2.25B in free cash flow, a 46% operating margin, zero debt, and the share dilution behind the buybacks. Then we show why a 42.8x P/E and a 17.8% implied growth rate put today's price above our estimated intrinsic-value range. Chapters (0:00) Robinhood Q3 2026 Dossier (0:38) Business...

We run Anthropic's leaked S-1 numbers through a Graham-Dodd lens: 85% inference margins, $100B+ in cash and $500-600B in compute commitments. Then we look at what a trillion-dollar AI valuation leaves for a patient investor's margin of safety before the IPO. Chapters (0:00) Anthropic Deep Value Research (1:59) The Capital-Efficient Compounder (2:37) Selling a Dollar for Ten Cents (3:33) The Capita...

A former Alibaba executive was told to multiply her forecast by seven, and her team hit it. Learn what that hyper-growth era cost her health and her family, and why value investors should audit their own human capital as carefully as a balance sheet. Chapters (0:00) The Economics of Hyper-Growth (0:55) Thesis: Growth's Unfunded Liabilities (2:39) The Scaling Engine (5:21) By the Numbers: The Bull...

Bullish on PayPal (PYPL) as of 2026-11-08: $6.58B Free cash flow TTM. We put PayPal's $6.58B in free cash flow, its 31% share retirement and three DCF scenarios side by side to see whether a 9.3x P/E is a value trap or a real margin of safety. Revenue is growing only 2% and Apple Pay is squeezing checkout, so the cash numbers tell you more than the story the market is telling. Chapters (0:00) PayP...

Deep value research on Microsoft (MSFT) as of 2026-11-07: $190B AI capex plan. A leaked memo shows Microsoft offering buyouts to about 60,000 employees, roughly 27% of its workforce, while planning up to $190B in AI data center spending. We use a Graham-Buffett lens to show what swapping payroll for depreciating hardware means for owner earnings, maintenance capex and the moat. Chapters (0:00) The...

Bearish on Anthropic (ANT) as of 2026-11-06: $431B Fixed compute. We tear down Anthropic's S-1 to show why its $80.7B loss is mostly a non-cash accounting charge, and why $431B in fixed compute commitments matters far more. Then we stress-test the rumored $1.5T valuation with Graham-Dodd math to see whether any margin of safety is left. Chapters (0:00) Anthropic Pre-IPO Diagnostic (0:40) From AI L...

Executive Order 14330 opens the $14 trillion 401(k) pool to private equity just as the average PE deal sits 60% below its 2021 valuation. We compare that mark-to-myth trap with ASML's transparent EUV monopoly, and show why even a wide moat needs a mid-cycle, normalized valuation. Chapters (0:00) The PE Trap and the Ultimate Monopoly (0:43) The Noise vs the Signal (1:47) EO 14330: The Exit Valve (2...

Li Lu argues that the US-China AI race will be decided by where top talent chooses to live and work, more than by compute alone. We use a deep value lens on his thesis, covering a 1951 immigration quota that helped hand China a rare earth monopoly, the 95% of chip fabrication done in East Asia, and why recursive self-improvement in AI could break traditional ways of estimating intrinsic value. Cha...