To The Moon

Why Intel Is Ripping Higher: 18A Proof Meets the CPU Shortage Trade

$Intel(INTC)$   has seen a sustained rally recently, outperforming many large-cap U.S. semiconductor stocks. In our view, two main reasons are behind this market move.

18A is the credible catalyst

The cleanest fundamental driver is 18A plus Panther Lake moving from roadmap to spotlight, alongside Intel's marketing claim of 200 plus PC design wins versus AMD's 120 plus designs for its AI PC push.

Why does this matter for the stock? Because 18A is not only about selling more CPUs. It is about re rating Intel Foundry from "cash incinerator" to "credible manufacturing platform."

Intel Foundry's last four reported quarters add up to roughly $10.1 billion of operating losses. If 18A execution and external confidence help push Foundry toward breakeven, the valuation torque is big:

– $10.1B operating loss removed

– assume 12% tax, that is about $8.9B net income upside

– divided by 4.531B diluted shares, that is roughly $1.96 EPS

– at 12x to 16x earnings, that is roughly $24 to $31 per share of potential equity value investors can start to "option price" as credibility improves

That math is simplified, but it explains why a product milestone can move the stock more than a quarter of decent guidance.

The second story is a tradable narrative: CPU shortage

The market's other fuel is the returning "everything will be scarce" playbook. After memory and storage shortage chatter, investors are now front running a CPU shortage storyline.

Intel has previously talked about supply tightness in certain areas, which makes the rumor easier to believe. But the market is adding a newer spin: Agentic AI means CPUs matter more, because agents create more orchestration, scheduling, preprocessing, and sandbox style workloads around GPU clusters.

What FundaAI is arguing, and where the logic cracks

A recent FundaAI note argues server CPUs face a structural shortage driven by foundry constraints, claiming $Taiwan Semiconductor Manufacturing(TSM)$   can only satisfy about 80% of server CPU wafer demand in 2026 and that high end server CPU pricing could rise as much as 50%.

The direction of "more CPU work in the AI era" is understandable. The weak link is the TSMC bottleneck they implicitly point to.

The industry's loudest bottleneck at TSMC has been advanced packaging capacity such as CoWoS, because it is a critical step for HBM heavy AI accelerators. NVIDIA has described advanced packaging capacity as still being a production bottleneck even after rapid expansion. CoWoS constraints can be very real for GPUs that need logic plus HBM integration, but most mainstream server CPUs do not depend on CoWoS class HBM stacking. In plain English, a CoWoS traffic jam does not automatically prove CPUs are the thing that must go scarce.

So FundaAI's conclusion can be right as a trade, but the causal chain from "TSMC bottleneck" to "CPU structural shortage" is not clean.

The tickers the CPU shortage trade can pull higher

$Intel(INTC)$   : Intel's latest flagship server CPU is Xeon 6 Granite Rapids, and Intel has stated it is built on its latest Intel 3 process technology, which means Intel’s near term server roadmap is still anchored on Intel 3 while 18A is the longer dated rerating lever investors are really underwriting.

$NVIDIA(NVDA)$   : The market still underprices its CPU angle. NVIDIA's Vera CPU is explicitly positioned for agentic reasoning and is described as the first CPU to support FP8 precision, which reinforces the idea that CPUs can absorb inference adjacent work when GPUs are saturated. 

$Advanced Micro Devices (AMD.US)$ : EPYC has been steadily taking economic share in servers. Mercury Research data cited by Tom's Hardware showed AMD's server CPU unit share rose to 27.8% in Q3 2025 (from 27.3% in Q2), while Mercury also noted AMD's server revenue hit another record high on richer EPYC mix, though it did not disclose an exact revenue share percentage.

$Arm Holdings (ARM.US)$ : Arm is the "substitution pressure" that makes any extreme x86 shortage story harder to sustain. Arm's infrastructure leadership has said close to 50% of the compute shipped to top hyperscalers in 2025 will be Arm based, and Reuters reported that AWS's Arm CPUs have already represented over half of its added data center chip capacity in the past two years.

$Qualcomm (QCOM.US)$ : Qualcomm is not just a PC story anymore. Reuters reported Qualcomm plans to re-enter the data center CPU market with custom CPUs designed to connect tightly with NVIDIA platforms, including NVLink Fusion style ecosystems. That gives QCOM a new, higher beta way to trade the "CPU matters again" narrative beyond laptops.

Summary

Intel's surge has one durable foundation and one fragile amplifier. 18A credibility is durable because it can change Foundry's earnings trajectory. CPU shortage plus agentic AI is tradable because it needs lead times and pricing to confirm, and because CoWoS is a real bottleneck for AI accelerators while CPUs do not necessarily sit on that same choke point.

@TigerStars  @CaptainTiger  @TigerWire  @Daily_Discussion  @Tiger_chat  @Tiger_comments  @MillionaireTiger