The above button links to Coinbase. Yahoo Finance is not a broker-dealer or investment adviser and does not offer securities or cryptocurrencies for sale or facilitate trading. Coinbase pays us for certain activity generated through this link. Prices displayed are informational.

Marvell Technology (MRVL) trades at a 0.6x PEG ratio with data center revenue accounting for over 70% of sales and growing 50% year-over-year, while Micron Technology (MU) offers an even cheaper 0.2x PEG with fiscal Q2 2026 revenue of $23.86B (up 196% YoY) and 2026 HBM supply already sold out under long-term contracts.

AI infrastructure buildout is driving both companies’ growth by creating sustained demand for custom networking silicon and high-bandwidth memory that smooths traditional memory-industry cycles.

The analyst who called NVIDIA in 2010 just named his top 10 AI stocks. Get them here FREE.

Semiconductor stocks have ridden the AI wave to eye-popping heights. Data centers now devour chips for training and running large language models, pushing the sector’s valuations sky-high. Yet smart investors know not every high-flyer delivers lasting returns.

To find stocks to buy in today's market, legendary investor Peter Lynch, who grew the Fidelity Magellan fund from $18 million to over $14 billion in assets -- a 29% compound annual return -- during his 14-year tenure, would likely reach for the same straightforward tools that any retail investor could grasp.

Lynch focused on understandable businesses with solid earnings growth and reasonable prices. One tool he leaned on heavily was the PEG ratio, his go-to screen for spotting growth stocks before the crowd caught on. It essentially asked how much are you paying for each unit of growth? For Lynch:

The analyst who called NVIDIA in 2010 just named his top 10 stocks. Get them here FREE.

PEG of 1.0x -- fairly valued

PEG below 1.0x -- potentially undervalued

PEG above 2.0x -- what Lynch often viewed this as the “danger zone”

Now let’s apply that lens to some of today’s semiconductor leaders:

Company

PEG Ratio

Intel (NASDAQ:INTC)

3.2x

Applied Materials (NASDAQ:AMAT)

2.5x

Arm Holdings (NASDAQ:ARM)

2.2x

These are high-quality businesses, but based on Lynch’s framework, investors are paying steep premiums for their growth. Surprisingly, two AI-driven chip stocks sit at the opposite end of the spectrum -- Marvell Technology (NASDAQ:MRVL) at 0.6x and Micron Technology (NASDAQ:MU).

Marvell Technology has quietly become a key supplier to the AI infrastructure buildout. Instead of competing head-on with giants like Nvidia (NASDAQ:NVDA), Marvell focuses on custom silicon, networking chips, and data center connectivity -- the plumbing behind AI. Its chips handle the high-speed Ethernet connections and optical modules that let AI servers talk to one another at blistering speeds. AI demand has expanded the business rapidly.

According to Marvell’s latest earnings release, data center revenue now accounts for over 70% of total sales, and that segment grew over 50% year over year. That’s not theoretical growth -- it’s already happening.

Here’s what the numbers tell us:

PEG ratio: 0.6x

Forward P/E: 30.1x

Expected earnings growth: ~50% annually

Although a 30x earnings multiple might sound high in isolation, but when earnings are growing at roughly 50%, the PEG compresses quickly. That’s exactly the kind of mismatch Lynch looked for.

Beyond PEG, Lynch also liked understandable businesses with clear demand drivers. Marvell checks that box. Hyperscalers like Amazon (NASDAQ:AMZN) and Microsoft (NASDAQ:MSFT) are designing custom AI chips, and Marvell helps make that possible. That creates long-term design wins, not just one-off sales.

Granted, Marvell isn’t risk-free. Customer concentration is real, and custom chip cycles can be uneven. But with AI infrastructure spending still expanding, the growth runway remains intact.

If Marvell is the infrastructure enabler, Micron Technology is the fuel supplier. Memory -- DRAM and NAND -- is essential for AI workloads, and demand is rising fast. AI has supercharged the need for its high-bandwidth memory (HBM), the specialized DRAM that feeds data to GPUs at record speeds.

That last point matters. HBM is critical for AI accelerators, and Micron has locked in supply commitments well ahead of production, with supply for 2026 already sold out under long-term contracts. In fiscal Q2 2026, Micron reported revenue of $23.86 billion -- up 196% year-over-year -- with DRAM revenue alone rising 207%.

Now look at valuation:

PEG ratio: 0.2x

Forward P/E:  5x

Expected earnings growth: 70%+ annually

That’s not just below Lynch’s threshold -- it’s a fraction of it.

Historically, memory stocks trade cheaply because of boom-and-bust cycles. Lynch knew this and often leaned into cyclical names when the numbers justified it. The difference today is AI demand is smoothing out some of that volatility. Data center customers are placing longer-term orders, which reduces the usual pricing swings. The company also generates strong free cash flow and boasts a clean balance sheet relative to peers.

That said, cycles don’t disappear entirely. If memory pricing weakens, Micron’s earnings can fall just as quickly as they rise. That’s the trade-off investors need to accept. Still, when all is said and done, a 0.2x PEG ratio suggests the market isn’t fully pricing in the current growth trajectory.

This analyst's 2025 picks are up 106% on average. He just named his top 10 stocks to buy in 2026. Get them here FREE.