Wall Street's Surprising Pick for AI Chip Dominance in 2026: Why Certain Analysts Are Betting Against the Giants

When it comes to AI semiconductors, everyone assumes the usual suspects—Nvidia and Broadcom—are the only plays worth making. But certain Morgan Stanley analysts, led by Joseph Moore, have a different thesis: Micron Technology is their semiconductor pick for 2026, despite what most of Wall Street believes.

The Numbers Tell a Different Story

Here’s where the consensus breaks down:

• Nvidia (NASDAQ: NVDA): 69 analysts tracking it with a median price target of $250/share (33% upside from $187)
• Broadcom (NASDAQ: AVGO): 52 analysts with a $460 target (31% upside from $350)
• Micron (NASDAQ: MU): Only 44 analysts covering it, with a $305 target (4% upside from $293)

The gap in analyst coverage itself is telling. While Nvidia and Broadcom hog the spotlight, Micron is flying under the radar—which might be exactly why certain smart money is circling it.

Why Nvidia Remains Formidable (Even If It’s Not the Pick)

Don’t get it twisted: Nvidia isn’t going anywhere. The company’s dominance in AI accelerators (80%+ market share) stems from its full-stack strategy. Nvidia doesn’t just sell GPUs; it sells entire data center ecosystems—pairing processors with networking equipment and its proprietary CUDA software ecosystem.

This creates an economic moat. As CEO Jensen Huang put it: “Our TCO [total cost of ownership] is so good that even when competitors’ chips are free, it’s not cheap enough.”

Wall Street expects Nvidia’s earnings to grow 37% annually over the next three years, making its 46x earnings valuation look reasonable for a monopoly player.

Broadcom’s Dual Advantage in Networking and Custom Chips

Broadcom controls two critical chokepoints in the AI supply chain:

1. Ethernet switching/routing: 80% market share in high-speed networking chips, with an expected 20-30% annual growth rate (per JPMorgan Chase). Its latest Tomahawk 6 and Jericho 3 chips set industry benchmarks.

2. Custom AI accelerators (ASICs): 70-80% market share. Major customers include Google, Meta, ByteDance, OpenAI, and Anthropic. The pipeline includes Apple, Arm Holdings, and xAI.

The AI accelerator market is projected to expand at 29% annually through 2033. Throw in Broadcom’s VMware virtualization subsidiary—recognized as a leader in distributed hybrid infrastructure—and you’ve got a diversified play. Expected earnings growth: 36% annually, with a current valuation of 51x earnings.

The Contrarian Play: Why Micron Gets Certain Analysts Excited

Here’s where Micron diverges from the pack. The company supplies DRAM and NAND memory—the “boring” but essential infrastructure for AI workloads.

Why it matters:

• DRAM (especially HBM) provides bandwidth for training and inference
• NAND stores datasets and models
• Micron gained 10 percentage points of HBM market share in the past year alone
• Its competitors (Samsung, SK Hynix) are losing ground

The key insight from Morgan Stanley: The AI buildout has created the worst DRAM and NAND shortage in 30 years. Prices are rising, and Micron is positioned to capture outsized margins.

Wall Street expects Micron’s earnings to grow at 48% annually—faster than both Nvidia and Broadcom—yet it trades at just 28x earnings. That’s the real valuation gap.

The Bottom Line

Nvidia and Broadcom aren’t bad bets. But certain analysts see a hidden opportunity: a memory chip maker gaining share during a supply crisis, trading at a fraction of its growth rate, and positioned to dominate a market that’s critical to every AI infrastructure buildout happening right now.

Sometimes the best investments aren’t the obvious ones.