Intel: Cyclical Recovery or Secular Demise? - [Business Breakdowns, EP.149]

Episode Summary

Executive Summary: The episode argues that Intel is a high-risk, high-upside turnaround: a fallen semiconductor icon trying to regain process leadership, rebuild a foundry business, and enter AI with Gaudi GPUs and AI PCs. Todd Austin frames the sector as cyclical but secularly attractive, and says Intel’s success hinges on executing multiple major transitions at once.

Main Topics: Intel’s turnaround under Pat Gelsinger (Priority: 5/5): Intel is trying to recover from years of missteps by restoring process leadership, reorganizing the business, and rebuilding investor confidence. Todd sees the bar as low, but execution risk as high. Semiconductor cycles and secular growth (Priority: 5/5): Todd emphasizes that semis combine cyclical swings with long-term innovation, which creates unusually attractive investment setups when the cycle turns and the secular trend re-rates. What went wrong at Intel (Priority: 5/5): Intel missed mobile, fell behind TSMC on process technology and EUV adoption, made poor capital allocation decisions, and lost share in PCs and servers, eroding margins and its moat. Foundry strategy and geopolitical value (Priority: 4/5): Intel’s foundry push is framed as both a business opportunity and a strategic U.S./Western supply-chain asset, helped by Chips Act subsidies and demand for geographic diversification. AI GPU opportunity and Gaudi (Priority: 4/5): Intel is trying to build a viable AI accelerator business with Gaudi 3, but Todd says it remains early, heavily customer-concentrated, and dependent on software ecosystem development. AI PC and CPU stabilization (Priority: 4/5): A potential AI-enabled PC refresh and improving server CPUs could stabilize Intel’s legacy business, help arrest share losses, and support a return to better margins. Valuation, margins, and long-term upside (Priority: 4/5): Todd outlines a path back toward much higher operating margins if Intel can stabilize its core business and create meaningful foundry/GPU revenue streams, though major dilution from new investments will persist near term.

Key Arguments: Semiconductors are attractive because cyclical recovery can combine with a secular re-rating; investors often over-extrapolate negatives at cycle bottoms. This AI cycle is different from prior cycles because hyperscalers have huge self-funding balance sheets and are investing into an apparently vast, durable market. Intel’s reputation is currently poor, but the market’s expectations are low enough that even partial execution could surprise positively. Intel’s turnaround depends on five nodes in four years, including EUV adoption and advanced packaging/power technologies like gate-all-around and backside power. The foundry business could become strategically important and economically attractive, but only after years of heavy capital spending and ecosystem building. Gaudi and other GPU efforts are not yet proven; Intel needs real orders, not specs or blog posts, to validate the business. AI training is GPU-heavy, but inferencing may shift some workload back to CPUs and edge devices, helping Intel’s core business. A major bull case is that Intel regains stable share in PC/server CPUs, builds a multi-billion-dollar GPU business, and scales Intel Foundry into a meaningful third engine. Geopolitics and supply-chain fragility create a strong strategic argument for Intel as a U.S.-based alternative to Taiwan/Korea-based manufacturing. The biggest risk is execution: if Intel fails to deliver on process technology, ecosystem integration, or share stabilization, the massive capital program could destroy value.

Data Points: Years in semiconductors for Todd Austin: 8 cycles - He says he has lived through eight semiconductor cycles as an investor/operator analyst. Starting year of his semis coverage: 1995 - He began covering semiconductor stocks in the mid-90s during a memory-cycle downturn. Intel peak revenue: $78 billion - Referenced as Intel’s revenue at its historical peak. Intel peak operating margin: above 30% - Described as Intel’s operating margin when the business was strongest. Intel current/near-term revenue: $54 billion - Todd cites this as the business level “this year” while discussing the decline. Intel current operating margin: 7.5% - Cited as the current depressed operating margin. Gross margin decline: 17% from the peak - Todd says gross margins are down 17% from peak levels. R&D spending as a share of sales: 7.5% - He cites R&D spend at 7.5% of sales. Current stock price range: $47-$48 - Todd references the stock trading in the high $40s. Target operating margin by 2026: 25% - His stated benchmark for Intel’s recovery by 2026. Earnings power at target margin: ~$4/share - Implied earnings power if Intel reaches 25% operating margin. Implied valuation multiple: 18-19x - Applied to the $4 earnings power to estimate a mid-$70s share price. Intel Foundry revenue target: $3 billion in three years - Todd suggests Foundry could reach this level if customer deposits and ramps materialize. Gaudi GPU revenue target: $2-$3 billion annually - His near-term dream scenario for Intel’s GPU business. Long-term Foundry revenue: 10 billion+ - Todd sees Intel Foundry Solutions potentially becoming a double-digit billions business by decade-end. Gross margin long-term goal: 60% - Cited as a long-run goal discussed by Intel management. Operating margin long-term goal: 40% - Cited as a long-run goal discussed by Intel management. Chip/AI accelerator market estimate: $400 billion in 2027 - Referenced as a bullish market-size estimate for AI accelerators. Associated data center spend estimate: $1.1 trillion - If the $400 billion accelerator market is realized, this would be implied data-center spending. PC market size historically: 300 million+ units - Todd cites the PC market as having exceeded 300 million units before the pandemic distortions. PC market recent level: ~250 million units - He says PC shipments fell back to around 250 million after the pandemic boom. PC client revenue floor in 2023: $29 billion - A conservative bottoming estimate for Intel’s client business. PC client revenue by 2026: $34 billion - Todd’s conservative recovery estimate for the client segment. Data center CPU revenue in 2023: $15.7 billion - He gives this as a baseline for Intel’s data center CPU business. Data center CPU revenue by 2026: $23.4 billion - His conservative recovery estimate for the data center CPU segment. Chips Act funding available to Intel: $8-$12 billion - Todd estimates Intel could receive this amount in subsidies. U.S. Chips Act total: $53 billion - Referenced as the total government support package for domestic semiconductor manufacturing. Capital spending plan: $125 billion - Todd cites Intel’s plan to invest this amount over the next five years. Training budget cited: $50 million - Used in discussing the cost of training workers for advanced fabs. Potential market cap comparison: $200 billion - He says Intel’s market cap could support significant upside if execution works. NVIDIA market cap reference: $1.5 trillion - Used as a comparison for the scale of AI leader upside.

Pivotal Quotes: "cyclical recovery and then a secular re-rate" — Todd Austin: He uses this phrase to explain why semiconductors can be exceptional investments when cycle and innovation align. "one of the great American turnarounds in corporate American history" — Todd Austin: He quotes Pat Gelsinger’s ambition for Intel’s multi-year reset. "It is build the airplane and fly it at the same time" — Todd Austin: He describes Intel’s challenge of rebuilding process leadership, foundry capabilities, and ecosystem relationships simultaneously.

Implications: Intel is a classic turnaround with strategic importance, but success depends on flawless execution in manufacturing, foundry, and AI. If it works, the upside could be substantial; if not, the capital program may destroy value.