Intel's "life-and-death" moment: How did Chen Liwu settle the estate and start self-rescue outside the ICU?

Article by: DaiDai, MaiTong MSX

Compilation by: Frank, MaiTong MSX

2025 will be a truly life-and-death year for Intel.

Compared to Nvidia, which is dancing at the center of the AI stage, Intel (INTC.M) is more like a foot lingering at the ICU door. With “tech evangelist” Pat Gelsinger stepping down at the end of last year in disappointment, the new CEO, Chen Lifu (Lip-Bu Tan), officially took over this aging machine burdened with heavy responsibilities.

The latest Q4 financial report also brings a familiar question back to the surface: Has this century-old chip giant already reached its end?

If we only look at the stock price performance after the earnings release, the answer might be disappointing. But if we extend the timeline, what Intel is experiencing may not be a last gasp but a critical moment of removing the oxygen tube from ICU and attempting autonomous breathing.

Objectively speaking, this Q4 report is not only the first full answer sheet after Chen Lifu’s appointment but also a comprehensive reckoning of the legacy of the Kissinger era.

Intel’s stock price change in 2025, source: CNBC / Intel Newsroom

1. Leadership Change: From Kissinger’s “Technological Utopia” to Chen Lifu’s “Capital Arena”

“Survival is the first need of civilization.” Before understanding this financial report, we must first see the narrative shift Intel is undergoing.

When Pat Gelsinger returned to Intel, he painted an almost idealistic blueprint: through a radical plan of “5 years, 5 process nodes,” to regain process leadership, rebuild U.S. domestic semiconductor manufacturing capacity, with the simple logic—if technology leads, capital will follow.

Thus, large-scale factory construction spread globally: Ohio in the U.S., Germany and Poland in Europe, with capital expenditure rapidly expanding.

But reality quickly counterattacked. The AI wave swept in, and Nvidia (NVDA.M) truly took center stage. In data centers, Intel’s CPUs no longer dominate; they are even joked about as “accessories” to GPUs. Meanwhile, huge capital expenditures drained cash flow, and the stock price continued to decline, plunging Intel into a dilemma of “technologically correct but financially bleeding.”

So, it can be said that Kissinger’s departure marked the end of Intel’s “cost-no-object pursuit of technological dominance,” while Chen Lifu’s arrival represents a completely different survival logic—former CEO of Cadence and a venture capital heavyweight, he is well-versed in risk investment, especially in “asset-liability magic.” His post-appointment strategy is extremely brutal and clear: “cut losses, then go all-in on core assets”:

• Surgical layoffs: global reduction of 15%, cutting bloated middle layers;
• Strategic contraction: pause expensive overseas factories in Germany and Poland, to recoup funds;
• Strategic focus: allocate all resources to only two targets—Arizona Fab 52 and 18A process;

INTC stock price trend from Q4 2025 to Q1 2026 (source: Yahoo Finance / TradingView)

This is a typical “cut-loss—shrink—protect core” strategy. For Chen Lifu, survival is more important than storytelling.

2. Dissecting the Financials: How much “overperformance” is truly valuable this time?

Returning to the data itself, Intel’s Q4 financial report is not entirely without merit.

On the surface, profits indeed exceeded market expectations, with EPS returning to positive territory. Especially, a profit of $0.15, though accompanied by layoffs, did pull the profit statement back from the cliff, proving that large-scale cost reduction and efficiency improvements have begun to repair the profit sheet. Intel has temporarily escaped the danger zone of “continuous bleeding.”

AI-assisted table generation

But digging deeper, problems still remain.

First, revenue still declined year-over-year. Amid the AI boom in 2025 boosting the semiconductor industry, AMD and Nvidia’s data center businesses grew rapidly, while Intel’s overall revenue continued to shrink. This indicates that the profit improvement was more from “cost savings” rather than “growth from earnings.”

Second, gross margin rose to about 38%, but this level remains very low in Intel’s history. Ten years ago, the company’s gross margin was consistently around 60%. Even in recent years, it has mostly hovered around 50%. In comparison, TSMC (TSM.M) still exceeds 50%, and Nvidia surpasses 70%.

Comparison of semiconductor giants’ gross margins (based on Q4 2025 Non-GAAP data)

Ultimately, the main reason for this gross margin improvement is not a return of product pricing power but the increased capacity utilization leading to the dilution of fixed depreciation costs. Especially in the server CPU market, Intel still needs to compete directly with AMD through pricing strategies; pricing power has not truly returned to its hands.

In other words, this is a “successful stop-loss but not yet recovered” financial report—like a “super performance” after a significant downward revision of the passing line, fundamentally unchanged in the underperformance.

However, it’s worth noting that the market’s previous biggest concern was Intel’s liquidity. From this perspective, the situation has indeed eased: as of Q4, Intel held about $37.4 billion in cash and short-term investments; repaid some debt during the quarter; and had an operating cash flow of about $9.7 billion for the year.

Meanwhile, the company has gained valuable time by selling part of its Mobileye stake, bringing in external capital for Altera, and receiving subsidies from the U.S. “CHIPS Act.” Additionally, Nvidia’s strategic investment of $5 billion in Intel also signals significant capital confidence.

Overall, Intel’s cash flow crisis is temporarily alleviated, at least enough to support it through the most cash-intensive phase before 18A process mass production.

But beware: this is not “financial security,” more like “buying time for survival.”

3. 18A and Panther Lake: the last chance to turn the tide?

From a business structure perspective, Intel’s current state is highly segmented.

Client PC business remains the company’s cash cow. The PC industry inventory cycle has basically cleared, OEMs are starting to restock, providing Intel with relatively stable cash flow, but due to product mix and foundry costs, this segment is unlikely to significantly expand margins in the short term.

Data center and AI businesses, after continuous decline, saw about 9% YoY growth in Q4. This rebound mainly comes from the improved competitiveness of the Xeon 6 platform and cloud vendors replenishing CPU resources after GPU investments. But in the long run, Intel’s market share in data centers has fallen sharply from its peak in 2021; currently, it’s more about “stabilizing the fall” rather than a true reversal.

The real pressure still comes from the Foundry business, which suffered quarterly losses of billions of dollars, mainly due to high depreciation of advanced process equipment, the ramp-up costs of 18A, and unreleased revenue from external clients.

Before 18A reaches large-scale production, foundry is more like a bleeding wound, but optimistically, if the financial report reflects the past, then 18A will determine the future and is crucial for the strategic position of the U.S. semiconductor industry.

Frankly, 18A is not just a process node; it’s the only ticket for Intel to return to the throne.

From a technical path, 18A is not a fantasy. Its RibbonFET (GAA) architecture puts Intel on par with industry leaders in transistor structure; PowerVia backside power delivery technology gives it a stage-leading advantage in energy efficiency and routing density. More importantly, 18A will be first massively applied to Panther Lake, a consumer platform.

Official data shows significant improvements in performance, gaming, and battery life, especially battery life, meaning x86 laptops are now approaching or even challenging Apple Silicon in user experience.

Meanwhile, Microsoft (MSFT.M) and Amazon (AMZN.M) have become anchor customers for 18A. Nvidia’s strategic investment is seen by the market as a “geopolitical insurance” endorsement of Intel’s manufacturing capacity. Additionally, the yield of 18A is reportedly improving at a rate of 7% per month, entering a predictable trajectory.

In comparison, TSMC is expected to apply similar technology in the A16 process by the end of 2026, meaning by 2026, Intel’s 18A will be at the forefront globally in power delivery technology, highly attractive to energy-sensitive customers (like Apple, Qualcomm, and AI inference chip makers).

Overview of Panther Lake chip architecture and 18A (source: Intel Tech Tour)

CEO Chen Lifu also stated in the earnings call: “Foundry losses peaked in 2024 and will start to narrow in 2025.” If this goal is achieved, with losses shrinking, Intel’s overall profit could see explosive growth.

These may not mean victory, but at least they show that Intel has not been completely abandoned.

4. Q1 guidance shocks the market? Understanding the game beyond fundamentals

On the surface, Intel’s sharp decline after the earnings release is almost a textbook market reaction.

The main reason is the extremely conservative guidance for Q1 2026, with revenue below consensus expectations and Non-GAAP EPS even pushed to 0. For short-term funds, this is a naked signal—“don’t expect profit improvement next quarter.” In a market accustomed to “AI high-growth narratives,” such guidance naturally triggers selling.

But if we interpret this only as a deterioration of fundamentals, we might miss a more important layer. A more reasonable explanation is that this is a typical “Kitchen Sink” strategy by the new CEO: new leadership often performs a “Kitchen Sink”—completely releasing bad news, lowering expectations, and paving the way for future surprises.

AI-assisted table generation

From this perspective, Q1 guidance is more a strategic conservatism than a sign of operational failure. What’s truly worth noting is a geopolitical undercurrent gradually emerging beneath the financial report.

At the industry level, Intel faces a nearly hellish competitive environment:

• AMD (AMD.M): Zen 6 architecture imminent, still enjoys TSMC’s most advanced process priority, with stable product cadence and clear roadmap;
• Nvidia: Blackwell still in high demand, continuously drawing global data center capital expenditure;
• ARM / Qualcomm: PC-side encroachment on x86 territory, with Apple M series and Qualcomm X Elite always like a sword hanging overhead;

In this landscape, Intel is almost impossible in the short term to defeat all rivals through “business competition” directly. This also determines a reality: Intel’s valuation logic is quietly shifting from performance-driven to “system value.”

Forecast of AMD vs. Intel CPU market share (Q2 2025)

And this is the key to understanding Nvidia’s $5 billion investment.

On the surface, Nvidia’s $5 billion investment in Intel seems counterintuitive. One is the absolute leader in AI chips, the other still bleeding in foundry. But stepping out of the financials and looking at supply chain security, this deal makes perfect sense:

Investing in Intel, supporting its foundry business—especially advanced packaging and U.S. capacity—is essentially buying a long-term “geopolitical insurance” for itself. It’s not about immediately shifting orders from TSMC but preparing an activated backup system in advance.

And this is exactly the scenario the White House favors—two American semiconductor giants forming a certain “symbiotic structure,” reducing the industry’s dependence on overseas single points.

This also means that, even in fierce business competition, Intel is still regarded as an indispensable infrastructure node that cannot fall.

Final thoughts

Overall, this financial report is neither a sign of full recovery nor a verdict of the end.

What Intel has truly achieved is not to continue grand technological utopian narratives but to return to a more realistic and brutal path: shrinking scale, preserving cash, and betting on a single core variable.

Among these, 18A and Panther Lake are like Intel’s “qualification tests”—if they win, there’s still a chance for re-pricing; if they lose, this century-old giant will be completely marginalized as an industry supporting role.

Ultimately, Intel is no longer that “landlord’s fool” who can squander capital at will but a heavy-asset company at the ICU door, needing meticulous financial management.

Whether it can truly walk out of the hospital depends not on this financial report but on its execution over the next 12–18 months.