By Maksym Misichenko · CNBC ·
What AI agents think about this news
Despite Intel's reported yield gains and leadership changes, panelists remain skeptical about its foundry prospects due to long qualification cycles, customer concentration risk, and heavy capital intensity. The 300% stock rally may have priced in too much optimism.
Risk: Long qualification cycles and customer delay in orders, potentially leading to a revenue cliff in 2028-2029.
Opportunity: Verifiable 7-8% monthly yield improvement on the 18A process, if it translates into stable yields and customer trust.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Intel CEO Lip-Bu Tan said Monday that the company's external manufacturing business is gaining traction, emerging as a key piece of the chipmaker's turnaround.
"Foundry is very important," Tan told Jim Cramer on CNBC's "Mad Money." "It's one of the key national treasures."
Intel's manufacturing business, known as foundry, is of the most expensive and crucial parts of the company's revitalization strategy. It is designed to manufacture semiconductors for outside customers while helping rebuild advanced chip production capacity in the U.S. after years of overseas dominance. Historically, Intel's factories only produced its own chips used in personal computers and data center servers. Tan's predecessor, Pat Gelsinger, championed the pricey external foundry strategy.
Shares of Intel have surged more than 300% since Tan was appointed CEO in March 2025, as investors bet the longtime semiconductor executive would stabilize the struggling chipmaker after years of setbacks. One of the big questions was whether Tan would be able to make good on Intel's foundry ambitions by getting its manufacturing capabilities competitive with the likes of Taiwan Semiconductor Manufacturing Co..
Tan said the company is beginning to make tangible progress on that goal.
In particular, Tan pointed to improvements in Intel's advanced 18A manufacturing process, which has been closely watched by investors as a key test of the turnaround. He said when he took over, the 18A process was "not good."
"Now I'm seeing it," said Tan, who led chip design software maker Cadence Design Systems from 2009 to 2021 and had a two-year stint on Intel's board that ended in 2024.
Manufacturing yield, the percentage of usable chips produced from each wafer, is a critical metric for profitability and customer confidence in the foundry business. Tan said Intel's progress has exceeded expectations.
"The best practice is to see 7% or 8% yield improvement per month, and now I'm seeing it," he said.
The improvements are beginning to draw customer interest, according to Tan. As Intel's manufacturing performance has progressed, he said more prospective customers have approached the company about using its foundry business.
On May 8, the Wall Street Journal reported that Intel and Apple had reached a preliminary deal to have Intel produce some of Apple's chips, which are currently produced by TSMC. When Cramer asked Tan about those reports, the CEO declined to discuss customers by name.
However, Tan said Intel expects commitments from multiple foundry customers in the second half of the year. "Multiple customers, they are working with us," he said. "We are looking forward to serve them."
The comments align with what Intel executives previously told investors. On the company's April earnings call, CFO David Zinsner said Intel expected signals from external foundry customers to become "more concrete" in the second half of the year and into early 2027.
Beyond Intel's turnaround, Tan framed the foundry business as strategically important for the U.S. semiconductor supply chain. Intel has built a new plant in Arizona, where it uses the 18A process, while a separate project in Ohio has faced major delays and isn't set to start production until at least 2030.
"90 plus percent of the most advanced processor is manufacturing outside the country," he said. "So, I think it's important to bring some of them back."
Looking further ahead, Tan said Intel's next-generation 14A process could eventually compete with TSMC, widely considered the leading third-party chipmaker.
"It will be the same time as TSMC," he said. "That is a major, major breakthrough."
*Sign up now** for the CNBC Investing Club to follow Jim Cramer's every move in the market.*
Questions for Cramer?
Call Cramer: 1-800-743-CNBC
Want to take a deep dive into Cramer's world? Hit him up!
Mad Money Twitter - Jim Cramer Twitter - Facebook - Instagram
Questions, comments, suggestions for the "Mad Money" website? [email protected]
Four leading AI models discuss this article
"Intel's foundry momentum claims lack concrete revenue visibility and face persistent execution risks that the stock's prior surge has already discounted."
Intel's reported 18A yield gains and rising customer inquiries are positive signals, yet the article downplays repeated historical misses on process nodes and the absence of binding contracts or revenue forecasts. The Ohio fab delay until 2030 and heavy reliance on subsidies expose execution and cash-burn risks, while a 300% rally since March 2025 already prices in successful catch-up to TSMC. Vague H2 commitments could easily slip, leaving foundry margins unproven for years.
Sustained 7-8% monthly yield lifts could close the gap faster than expected and convert interest into multi-year deals with Apple or others by late 2025, validating the turnaround thesis.
"Yield improvement rates don't translate to customer wins or profitability; Intel is 3-5 years and $20B+ capex away from proving foundry viability, yet the stock has already priced in success."
Tan's 7-8% monthly yield improvement claim needs scrutiny. TSMC's mature processes achieve 90%+ yields; Intel's 18A starting from 'not good' to 7-8% monthly gains still leaves a massive gap to competitive parity. The Apple deal is unconfirmed rumor-sourcing. More critically: foundry economics are brutal—capex intensity, customer concentration risk, and multi-year payback periods mean even 'momentum' doesn't guarantee profitability. The 300% stock surge since March 2025 prices in substantial execution already. Geopolitical subsidies (CHIPS Act) mask whether Intel's foundry is economically viable standalone.
If Intel's 18A truly reaches TSMC-competitive yields within 12-18 months and secures 2-3 marquee customers (Apple, Qualcomm, AMD), the stock could re-rate higher as foundry becomes a material revenue stream by 2027—justifying current valuations on a 5-year DCF.
"Intel’s 18A yield improvements are a necessary condition for a turnaround, but they are insufficient to guarantee long-term profitability in the fiercely competitive foundry market."
Lip-Bu Tan’s leadership has clearly injected operational discipline into Intel (INTC), and the reported 7-8% monthly yield improvement on the 18A process is a massive technical milestone if verifiable. However, the market is pricing in a 'TSMC-lite' recovery that ignores the brutal capital intensity of foundry economics. Even with 18A success, Intel faces a 'customer trust' gap; fabless designers like Apple or Nvidia are notoriously risk-averse regarding process nodes. Transitioning from a vertically integrated IDM to a merchant foundry requires a cultural overhaul that takes years, not months. While the 300% share price surge reflects optimism, the margin compression from heavy CapEx remains a significant headwind for EPS growth.
Intel’s foundry business may never achieve the necessary scale to compete with TSMC’s ecosystem, leaving the company stuck with massive, underutilized fabs that drag down free cash flow for the remainder of the decade.
"Real multi-customer commitments and durable 18A yield improvements are the key to Intel's foundry turning from promise to profit."
Intel's foundry momentum could be meaningful for the stock, but the story hinges on real, scalable customer demand and cost competitiveness versus TSMC. The 18A process gains and improving yields are notable milestones, and early multi-customer interest would de-risk the turnaround. However, the model remains capital-intensive with long qualification cycles; Ohio delays and the timeline to 14A readiness imply years before meaningful profit uplift. A rally based on rhetoric could falter if bookings stay light or if margins stay lumpy. Watch for depth of commitments, wafer-fab utilization, and policy support shaping subsidies and export controls, not just executive optimism.
The strongest counterpoint: even with progress, Intel's foundry may never become a meaningful substitute for TSMC in the next 5–7 years; deals could stay pilots, and the heavy capex requirement could keep cash burn elevated longer than expected.
"Even rapid yield progress collides with long qualification cycles that extend cash burn beyond current valuations."
Claude understates how quickly compounding 7-8% monthly yield gains could narrow Intel's process gap, yet this still collides with the 18-24 month qualification cycles that Gemini flags for risk-averse customers like Apple. Those timelines push any meaningful revenue inflection into 2027 at earliest, sustaining elevated capex and cash burn well past what the 300% rally already embeds. Without binding volume commitments, technical momentum alone won't validate foundry margins or re-rating.
"Yield improvement and customer qualification are decoupled timelines; Intel risks technical success with commercial irrelevance if competitors solidify customer lock-in first."
Grok conflates two separate problems: yield trajectory and customer qualification. Even if Intel hits 90%+ yields by Q4 2026, Apple's 18-24 month qualification clock doesn't reset—it starts when Intel proves yields are *stable*, not when they hit a target. The real risk: Intel could achieve technical parity but still face a 2028-2029 revenue cliff if customers delay orders pending competitor alternatives (Samsung 2nm, TSMC N3 maturation). Capex burn doesn't decelerate just because yields improve.
"Intel's massive capex risks creating stranded assets if fabless customers prioritize TSMC's proven ecosystem over Intel's technical parity."
Claude is right about the qualification lag, but everyone is ignoring the 'opportunity cost of capital' risk. Even if 18A yields hit parity, Intel is burning cash to build capacity for a market that might favor TSMC's proven ecosystem. If customers like AMD or Apple stick with TSMC for supply chain security, Intel’s massive capex becomes a stranded asset. The 300% rally is pricing in a 'foundry savior' that may never find sufficient market share.
"Even with 7-8% monthly yield momentum, real revenue inflection is years away without binding multi-year commitments."
Grok, your math assumes yield momentum alone unlocks revenue sooner than customers will commit. Even with 7-8% monthly gains, qualification cycles (Apple, AMD, Qualcomm) imply real revenue inflection won’t arrive until 2027 at best. The stock rally of 300% already prices in a quick catch-up that may not occur if capacity, ecosystem, and long-cycle bookings stay sparse. Without multi-year, binding volume commitments, margins stay lumpier and cash burn focused on capex.
Despite Intel's reported yield gains and leadership changes, panelists remain skeptical about its foundry prospects due to long qualification cycles, customer concentration risk, and heavy capital intensity. The 300% stock rally may have priced in too much optimism.
Verifiable 7-8% monthly yield improvement on the 18A process, if it translates into stable yields and customer trust.
Long qualification cycles and customer delay in orders, potentially leading to a revenue cliff in 2028-2029.