Intel is not a "safe" investment. It is a turnaround play on the most capital-intensive business in the world. The downside is bounded by physical assets ($105B PP&E) and strategic importance (US government has too much at stake). The upside depends entirely on whether Intel Foundry can become a credible TSMC competitor. Leopold's use of call options (not shares) tells you he sees this as a binary outcome with favorable odds — he wants convexity, not safety.
Leopold holds 20.2M shares worth of call options on Intel and exactly 1 actual share. This is not a Buffett-style "buy and hold forever" position. This is a calculated bet that Intel's stock will move significantly higher within a specific timeframe (options expire). The structure tells us several things:
1. US chip reshoring is a national security imperative. Leopold's entire worldview (documented in "Situational Awareness") centers on AGI racing. If AGI is the most important technology ever, and all advanced chips are made in Taiwan (a geopolitical flashpoint), then the US MUST build domestic semiconductor manufacturing. Intel is the only credible vehicle for this.
2. CHIPS Act creates a funding floor. Intel has received $8B+ from CHIPS Act ($5.7B commercial + $3.3B Secure Enclave). Plus $2B from SoftBank, $5B from NVIDIA. This is ~$15B in non-dilutive (or equity-based) funding that reduces the cash burn risk. The government literally owns shares and warrants in Intel — it will not let it fail.
3. Intel 18A changes the game. If 18A is competitive with TSMC N2, Intel has a credible product to sell to external customers. The first 18A products (Core Ultra Series 3) are shipping. This is proof of concept — the node works in production.
4. The valuation is absurdly low relative to assets. Intel's gross PP&E is $212B. Its market cap is $226B. You're essentially paying $14B for the entire Intel Products business (which earned $12.7B in operating income last year), the Intel brand, 80,000 employees, and all the IP. This is deep value territory.
| Period | Shares Equiv. | Reported Value | Implied Price | Action |
|---|---|---|---|---|
| Q1 2025 | 20,237,400 | $460M | $22.71 | NEW POSITION (Call options) |
| Q2 2025 | 20,237,400 | $453M | $22.40 | HOLD |
| Q3 2025 | 20,237,400 | $679M | $33.55 | HOLD |
| Q4 2025 | 20,237,400 | $747M | $36.90 | HOLD (+ added 1 share) |
| Current (Mar 2026) | 20,237,400 | ~$947M | $46.78 | Unrealized: +106% |
Leopold's portfolio is heavily concentrated in AI infrastructure and power: Bloom Energy (#1 position), EQT, CORZ, IREN (all power/compute plays). Intel fits the pattern — it's US-based AI infrastructure. Leopold is betting the US government will pour resources into domestic chip manufacturing, and Intel is the primary beneficiary.
| Year | Revenue ($B) | Gross Profit ($B) | Gross Margin | Op Income ($B) | Net Income ($B) | EPS (Diluted) |
|---|---|---|---|---|---|---|
| FY2018 | 70.8 | 43.7 | 61.7% | 23.3 | 21.1 | $4.48 |
| FY2019 | 72.0 | 42.1 | 58.6% | 22.0 | 21.0 | $4.71 |
| FY2020 | 77.9 | 43.6 | 56.0% | 23.7 | 20.9 | $4.94 |
| FY2021 | 79.0 | 43.8 | 55.5% | 19.5 | 19.9 | $4.86 |
| FY2022 | 63.1 | 26.9 | 42.6% | 2.3 | 8.0 | $1.94 |
| FY2023 | 54.2 | 21.7 | 40.0% | 0.1 | 1.7 | $0.40 |
| FY2024 | 53.1 | 17.3 | 32.7% | (11.7) | (19.2) | ($4.50) |
| FY2025 | 52.9 | 18.4 | 34.7% | (2.2) | (0.3) | ($0.06) |
| Segment | Revenue ($B) | Op Income ($B) | Op Margin | YoY Revenue |
|---|---|---|---|---|
| CCG (Client Computing) | 32.2 | 9.3 | 29% | -3% |
| DCAI (Data Center & AI) | 16.9 | 3.4 | 20% | +5% |
| Intel Products Subtotal | 49.1 | 12.7 | 26% | -1% |
| Intel Foundry | 17.8 | (10.3) | -58% | +3% |
| All Other | 3.6 | 0.3 | 7% | -1% |
| Consolidated | 52.9 | (2.2) | -4% | flat |
| Year | Total Revenue ($B) | Intersegment ($B) | External ($M) | Cost + OpEx ($B) | Operating Loss ($B) | Op Loss % |
|---|---|---|---|---|---|---|
| FY2023 | 18.5 | ~18.0 | $547 | 25.6 | (7.1) | -38% |
| FY2024 | 17.3 | ~17.1 | $159 | 30.6 | (13.3) | -77% |
| FY2025 | 17.8 | 17.5 | $307 | 28.1 | (10.3) | -58% |
| Year | Operating CF ($B) | CapEx ($B) | FCF ($B) | FCF Margin |
|---|---|---|---|---|
| FY2019 | 33.1 | 16.2 | 16.9 | 23.5% |
| FY2020 | 35.4 | 14.3 | 21.1 | 27.1% |
| FY2021 | 30.0 | 18.7 | 11.3 | 14.3% |
| FY2022 | 15.4 | 24.8 | (9.4) | -14.9% |
| FY2023 | 11.5 | 25.8 | (14.3) | -26.3% |
| FY2024 | 8.3 | 23.9 | (15.7) | -29.5% |
| FY2025 | 9.7 | 14.6 | (4.9) | -9.3% |
FCF has been deeply negative for 4 consecutive years. FY2025 showed improvement ($-4.9B vs $-15.7B) because capex dropped sharply ($14.6B vs $23.9B), but operating cash flow is still far below historical levels.
The global semiconductor foundry market is ~$150B annually and growing rapidly driven by AI. It is dominated by a single company:
| Foundry | 2024 Revenue | Market Share | Leading Node | HQ |
|---|---|---|---|---|
| TSMC | ~$87B | ~62% | N3/N2 | Taiwan |
| Samsung Foundry | ~$15B | ~11% | 3nm GAA | South Korea |
| GlobalFoundries | ~$7B | ~5% | 12nm (mature) | USA |
| UMC | ~$7B | ~5% | 14nm (mature) | Taiwan |
| SMIC | ~$8B | ~6% | 7nm | China |
| Intel Foundry (external) | $0.3B | <0.5% | Intel 18A | USA |
Intel 18A is Intel's most advanced process node, now in high-volume manufacturing (HVM) at facilities in Arizona and Oregon. It introduces two major innovations:
Intel's implementation of gate-all-around (GAA) transistor architecture. GAA replaces FinFET, which has been the standard since ~2012. GAA transistors wrap the gate around the channel on all four sides, providing better electrostatic control and lower leakage. TSMC is also moving to GAA with its N2 node.
A first-of-its-kind innovation that moves power delivery to the backside of the wafer, separating it from signal routing on the front side. This improves power efficiency, reduces signal congestion, and enables tighter transistor packing. TSMC's equivalent (backside power) is not expected until N2P or later.
Intel's process technology roadmap has a troubled history:
| Node | Original Target | Actual HVM | Delay | Notes |
|---|---|---|---|---|
| 10nm (Intel 7) | 2016 | 2019 | ~3 years | Massive yield problems, nearly killed Intel |
| 7nm (Intel 4) | 2021 | 2023 | ~2 years | Used EUV, significant catch-up |
| Intel 3 | 2023 | 2024 | ~1 year | Refined Intel 4 |
| Intel 20A | 2024 | Cancelled | N/A | Skipped in favor of 18A |
| Intel 18A | 2025 | 2025 (H2) | On time | First GAA + PowerVia in production |
| Intel 14A | 2026-2027 | TBD | TBD | "Potential pause" if no external customer |
For Intel Foundry to matter, external revenue needs to go from $307M to billions. Here's what the path looks like:
| Milestone | Timeline | Impact |
|---|---|---|
| First significant 18A customer tape-out | 2026-2027 | Proof of credibility |
| NVIDIA/Broadcom/Qualcomm places an order | 2027-2028 | Game-changer validation |
| External revenue reaches $5B | 2028-2030 | Segment approaches breakeven |
| External revenue reaches $15-20B | 2030+ | Intel Foundry valued as standalone entity |
| Item | FY2025 ($B) | FY2024 ($B) | FY2023 ($B) | Notes |
|---|---|---|---|---|
| Assets | ||||
| Cash & Equivalents | 14.3 | — | — | Rebuilt from near-zero |
| Short-term Investments | 23.2 | — | — | Highly liquid |
| Accounts Receivable | 3.8 | — | — | |
| Inventory | 11.6 | — | — | WIP: $7.8B, Finished: $2.8B |
| Current Assets | 63.7 | 47.3 | 43.3 | +35% YoY |
| PP&E (Net) | 105.4 | 107.9 | 96.6 | Slight decline (slower capex) |
| Goodwill | 23.9 | 24.7 | 27.6 | Some impairment |
| Intangible Assets | 2.8 | 3.7 | 4.6 | Amortizing down |
| Total Assets | 211.4 | 196.5 | 191.6 | |
| Liabilities | ||||
| Current Liabilities | 31.6 | 35.7 | 28.1 | |
| Long-term Debt | 44.1 | 46.3 | 47.0 | Slowly declining |
| Total Liabilities | 85.1 | — | — | |
| Equity | ||||
| Common Stock + APIC | 65.2 | — | — | Massive from share issuances |
| Retained Earnings | 49.0 | — | — | |
| Stockholders' Equity | 114.3 | — | — | |
| Non-controlling Interest | 12.1 | — | — | Altera minority interest |
| Total Equity | 126.4 | 105.0 | 110.0 | |
| Metric | Value |
|---|---|
| Stockholders' Equity | $114.3B |
| Less: Goodwill | ($23.9B) |
| Less: Other Intangibles | ($2.8B) |
| Tangible Book Value | $87.6B |
| Shares Outstanding | 4,994M |
| Tangible Book Per Share | $17.54 |
| Current Price | $45.25 |
| Price / Tangible Book | 2.58x |
| Category | Gross ($B) | Notes |
|---|---|---|
| Land & Buildings | 65.4 | Fabs in AZ, OR, NM, OH, Ireland, Israel |
| Machinery & Equipment | 111.9 | EUV lithography, ion implanters, etc. |
| Construction in Progress | 34.5 | New fabs being built (Ohio, expansion) |
| Total Gross PP&E | 211.9 | |
| Accumulated Depreciation | (106.5) | 50% depreciated |
| Net PP&E | 105.4 |
| Metric | FY2025 |
|---|---|
| Long-term Debt | $44.1B |
| Current Debt | $2.5B |
| Total Debt | $46.6B |
| Cash + ST Investments | $37.4B |
| Net Debt | $9.2B |
| Debt/Equity | 37% |
| Current Ratio | 2.02x |
| Interest Coverage (FY2025 OCF/Interest) | ~10x |
| Program | Amount | Status | What Intel Gave in Return |
|---|---|---|---|
| Commercial CHIPS Act (DFA) | $5.7B (accelerated) | Received in full (Aug 2025) | 275M shares + warrants for 241M shares at $20 |
| Prior CHIPS incentives | $2.3B | Received (pre-Aug 2025) | Accounted as government grants |
| Secure Enclave | $3.3B | Being disbursed over time | 159M escrowed shares at $20/share |
| CHIPS capital incentives (2025) | $769M | Recognized in FY2025 | — |
| Ohio fab grants | $123M | Recognized in FY2025 | — |
| Total CHIPS-related | ~$12.2B | Most received | Up to 675M shares + warrants |
| Partner | Amount | Shares Issued | Price/Share | Strategic Angle |
|---|---|---|---|---|
| SoftBank | $2.0B | 87M | $23.00 | AI infrastructure bet |
| NVIDIA | $5.0B | 215M | $23.28 | x86 + GPU co-development partnership |
| Altera sale (51%) | $4.3B | — | N/A | SLP affiliate; $500M deferred |
| Total non-CHIPS capital | $11.3B | 302M shares |
Pat Gelsinger (CEO Feb 2021 - Dec 2024): Gelsinger launched the foundry strategy, committed to the ambitious "5 nodes in 4 years" roadmap, and secured CHIPS Act funding. He was a semiconductor veteran (former Intel CTO) who understood the technology deeply. But he was forced out in December 2024 after the stock crashed ~60% and the board lost confidence in the pace of the turnaround.
Lip-Bu Tan (CEO since March 2025): Tan is the former CEO of Cadence Design Systems, one of the world's three major EDA (electronic design automation) companies. He is deeply embedded in the semiconductor ecosystem — he knows every chip designer, every foundry, every design tool. His appointment signals a shift toward commercial execution over technology ambition.
| Date | Employees | Change |
|---|---|---|
| Dec 2024 | 99,500 | |
| Sep 2025 (Altera deconsolidation) | ~93,500 | -3,000 (Altera) |
| Dec 2025 | 80,100 | -19,400 (restructuring) |
| Target (2026) | ~75,000 | -5,000 more |
2025 restructuring charges: $2.2B (primarily employee severance). This is Lip-Bu Tan cutting costs aggressively to stem the bleeding.
| Scenario | Probability | Impact on Intel | Stock Impact |
|---|---|---|---|
| AGI arrives, Intel Foundry succeeds, manufactures AI chips for multiple customers | 15-20% | Transformative — Intel becomes Western TSMC | $150-300+ (3-7x) |
| AGI arrives, Intel Foundry has modest external success, products business stabilizes | 25-30% | Gradual improvement, foundry losses narrow | $60-100 (1.5-2.5x) |
| AGI arrives, Intel Foundry fails to gain external customers, products business declines | 25-30% | Slow bleed; government keeps funding but no growth | $25-45 (flat to down) |
| AGI delayed/slower, Intel Products stabilizes, foundry remains internal | 15-20% | Intel becomes a smaller, focused CPU company | $30-60 (depends on margins) |
| Catastrophic: Intel Foundry fails, products decline, debt spiral | 5-10% | Government bailout/nationalization risk | $10-20 |
The answer is: both, simultaneously. Intel's products business (CPUs) is threatened by AGI shifting compute to GPUs and AI accelerators. But Intel's manufacturing business could be essential to AGI if the US needs domestic chip production. The stock is priced for the bear case. Leopold is betting on the bull case. The truth is probably somewhere in between — and that "somewhere" is still higher than $45.
There are too many floors under this company: $105B in physical fab assets, US government ownership stake, national security importance, $37B in cash/investments, and a products business still generating $12.7B in operating income. The zero-risk probability is effectively nil. But "not going to zero" is different from "good investment." Intel can absolutely stagnate at $20-30 for a decade if the foundry fails.
| Risk | Severity | Probability | Impact |
|---|---|---|---|
| Intel 18A yields disappoint / fail to attract customers | HIGH | 25-35% | Foundry thesis dies, stock $20-30 |
| Continued massive foundry losses ($10B+/yr for 3+ more years) | HIGH | 40-50% | Equity erodes, more dilution needed |
| Intel 14A paused or cancelled | MEDIUM-HIGH | 30-40% | Intel falls behind TSMC again permanently |
| PC market secular decline (AI cloud replaces local compute) | MEDIUM | 20-30% (10yr) | CCG revenue declines 20-40% |
| NVIDIA/AMD/ARM take server market share from x86 | MEDIUM | 40-50% | DCAI revenue declines 10-20% |
| TSMC opens US fabs, removing Intel's geographic advantage | MEDIUM | 60-70% | Weakens national security argument |
| Further dilution from share issuances / warrants | MEDIUM | 70-80% | EPS drag, ownership dilution |
| Taiwan conflict triggers TSMC disruption (helps Intel) | LOW (but tail) | 5-15% (5yr) | Intel becomes monopoly foundry |
| US government pulls back CHIPS Act support | LOW | 5-10% | Loss of funding, sentiment crash |
| Accounting/fraud risk | VERY LOW | <2% | N/A — Intel has clean accounting |
TSMC is building multiple fabs in Arizona (Fab 21). If TSMC produces leading-edge chips domestically at scale, Intel's key advantage — "only US-based leading-edge foundry" — disappears. TSMC Arizona is expected to produce N4/N3 chips by 2025-2026 and N2 by 2028. This directly undermines Intel's national security narrative.
| Year | Shares Out (B) | Change |
|---|---|---|
| 2018 | 4.61 | Buybacks |
| 2019 | 4.42 | Buybacks |
| 2020 | 4.20 | Buybacks |
| 2021 | 4.06 | Last year of buybacks |
| 2022 | 4.11 | Dilution begins |
| 2023 | 4.19 | +2% |
| 2024 | 4.28 | +2% |
| 2025 | 4.99 | +17% (massive issuances) |
| Potential (warrants) | 5.23+ | +5% if warrants exercised |
Because Intel has distinct businesses with very different characteristics, sum-of-parts is the right framework.
| Component | Bear Case | Base Case | Bull Case | Methodology |
|---|---|---|---|---|
| Intel Products (CCG + DCAI) | $80B | $120B | $160B | $12.7B op inc x 6-12x EV/EBIT |
| Intel Foundry (internal) | ($30B) | ($15B) | $0 | NPV of losses until breakeven |
| Intel Foundry (external optionality) | $0 | $10B | $60B | Probability-weighted future revenue |
| Cash + Investments (net of debt) | ($9B) | ($9B) | ($9B) | Net debt |
| Other (Mobileye stake, IP, etc.) | $5B | $10B | $15B | |
| Total Enterprise Value | $46B | $116B | $226B | |
| Equity Value | $37B | $107B | $217B | -net debt |
| Per Share (5.0B shares) | $7.40 | $21.40 | $43.40 |
| Company | Market Cap | Revenue | EV/Revenue | Op Margin | Notes |
|---|---|---|---|---|---|
| TSMC | ~$900B | ~$87B | ~10x | ~45% | Dominant foundry, the gold standard |
| NVIDIA | ~$3.2T | ~$130B | ~24x | ~60% | AI chip monopoly, different business |
| AMD | ~$180B | ~$25B | ~7x | ~22% | Fabless, directly competes with Intel CPUs |
| Samsung (semi division) | — | ~$60B | ~2x | ~15% | Foundry + memory, struggling at leading edge |
| Intel | $226B | $52.9B | 4.9x | -4% | IDM + aspiring foundry |
Intel trades at a significant discount to TSMC (4.9x vs 10x revenue) and AMD (4.9x vs 7x). The discount reflects Intel's unprofitability, execution risk, and declining market position. If Intel achieves TSMC-like foundry margins on even a portion of its manufacturing, the multiple re-rates dramatically.
This requires one of the following scenarios by 2030-2032:
| Scenario | Revenue | Op Margin | Op Income | Multiple | EV | Price/Share |
|---|---|---|---|---|---|---|
| A: Products stabilize + foundry works | $80B | 25% | $20B | 15x EV/EBIT | $300B | ~$55 |
| B: Foundry gets $15B external revenue | $85B | 28% | $24B | 18x | $432B | ~$80 |
| C: Foundry is real ($30B external), products grow | $110B | 30% | $33B | 20x | $660B | ~$125 |
| D: Intel becomes Western TSMC ($50B+ external) | $140B | 35% | $49B | 22x | $1,078B | ~$200 |
Methodology: Tangible book value ($87.6B) x 70-100% recovery in liquidation / 5.25B shares = $11.70-$16.70. Round to $12-18.
What this assumes: Intel's physical assets retain 70-100% of book value. Goodwill and intangibles are worthless. The business generates no future earnings. This is the "everything goes wrong but the government prevents bankruptcy" scenario.
Confidence: HIGH. The physical assets are real. The government will not let Intel go below this level. This is the hard floor.
Methodology: Intel Products at 8-10x operating income ($102-127B) + foundry at cost ($0) + net cash position + other assets. Divided by 5.0B shares = $35-55.
What this assumes: Intel Products continues earning $10-13B/yr in operating income. Foundry losses gradually decline but the segment never becomes independently valuable. Modest dilution.
Confidence: MODERATE. The products business is solid, but market share erosion and margin pressure are real risks.
Methodology: Intel Products at $55B revenue + Intel Foundry external at $15-50B revenue. Combined 25-35% operating margins = $18-37B operating income. At 15-20x = $270-740B enterprise value. At 5.25B shares = $50-140/share. Add optionality premium for foundry re-rating = $120-200.
What this assumes: Intel 18A/14A are competitive. Intel wins 2-3 major external foundry customers. US government continues support. No catastrophic geopolitical event (or a Taiwan event that HELPS Intel). Shares held below 5.5B.
Probability: 15-25%. This is the Leopold thesis. If you believe it, call options are the right vehicle.
| Scenario | Probability | Stock Price | Leopold's Return (from ~$23) |
|---|---|---|---|
| Foundry transforms Intel | 15% | $200 | ~8.7x on shares, likely 20-40x on options |
| Foundry has modest success | 25% | $80 | ~3.5x on shares, likely 5-10x on options |
| Status quo, slow improvement | 30% | $45 | ~2x on shares, options near breakeven |
| Foundry fails, products decline | 20% | $25 | Flat on shares, options lose 50-80% |
| Catastrophic | 10% | $15 | -35% on shares, options worthless |
| Expected Value | 100% | $72 | +3.1x shares, likely 5-10x on options |
Intel peaked at ~$62 in early 2026 after the CHIPS Act funding acceleration and Lip-Bu Tan appointment. It previously traded at $68 in early 2020 before the multi-year decline began. The all-time high was ~$75 in 2000 during the dot-com bubble.
Intel is NOT a Buffett-style investment. The business is in decline, the foundry is bleeding cash, and the path to recovery depends on flawless execution in the most capital-intensive industry on Earth. There is no "little chance of losing money" at $45/share — the fair value range is $35-55, meaning there's meaningful downside risk.
However, Intel has three things most turnaround stories lack: (1) irreplaceable physical assets worth $100B+, (2) US government financial and strategic support, and (3) a credible technology position (18A) that could unlock a massive new market. The downside is bounded; the upside is genuinely enormous.
Leopold's approach is correct for this type of bet. Call options give you convexity: limited downside, unlimited upside. If you believe there's a 15-25% chance Intel becomes the Western TSMC, and you can buy that optionality cheaply, the expected value is strongly positive. This is not an investment for the core portfolio — it's a sized bet on a specific outcome.
| Metric | Assessment |
|---|---|
| Can we compute a floor? | YES — physical assets + government backstop provide a floor |
| Floor price | $12-18/share (tangible book value with discount) |
| Floor confidence | MODERATE-HIGH — assets are real, but ongoing losses erode equity |
| Can it go to zero? | NO — too strategically important, too many physical assets |
| Can it stagnate for a decade? | YES — if foundry fails and products decline, $20-30 for years |
| Current price vs floor | $45 is 2.5-3.8x the floor — NOT cheap by our framework |
Analysis date: March 12, 2026. Data sources: SEC EDGAR 10-K (FY2025, filed 2026-01-23), XBRL Company Facts API, yfinance, 13F filings for Situational Awareness LP.