Best Way to Invest in AI Infrastructure Right Now — 2026 Strategy Guide
Photo by Paris Bilal on Unsplash
META_DESCRIPTION: Anthropic hits $30B ARR with explosive growth. Here’s how to position your portfolio in AI infrastructure stocks and ETFs before the next wave. Three actionable strategies inside.
How to Position Your Portfolio for AI Infrastructure Explosion
The AI revenue story just became impossible to ignore. Anthropic’s annual recurring revenue hit $30 billion this week—adding $11 billion in just five weeks. That’s not a typo. While skeptics debate whether AI will ever be profitable, the cash register is already ringing to the tune of tens of billions, and it’s accelerating.
For investors, the question isn’t whether AI is real anymore. It’s whether you’re positioned to capture the infrastructure build-out that’s happening right now. The companies selling shovels during this gold rush—chip makers, cloud providers, and specialized compute infrastructure—are posting revenue growth that looks more like early-stage tech companies than mature industrials.
By the end of this article, you’ll have three specific investment strategies with ticker symbols, entry considerations, and clear risk parameters. No theory. Just actionable positioning for what’s shaping up to be one of the defining investment themes of this decade.
Understanding the Opportunity: Why AI Infrastructure Is Just Getting Started
The numbers tell a compelling story. Anthropic grew from $100 million ARR at the start of 2024 to $30 billion today—roughly 18 months of exponential expansion. If they maintain even a fraction of this velocity and hit $100-150 billion by year-end, the infrastructure demands will be staggering.
Here’s what matters for investors: Anthropic doesn’t build its own chips or data centers. Neither does OpenAI. Neither do most AI companies. They rely on Nvidia for GPUs, Broadcom for custom AI accelerators, Google Cloud and Microsoft Azure for compute, and companies like Micron for high-bandwidth memory.
The recent Broadcom-Google partnership announcement with Anthropic isn’t just a press release—it’s a signal that hyperscale AI deployment is transitioning from experimental to operational. When you see those partnerships formalized, it means purchase orders, multi-year contracts, and predictable revenue streams flowing to infrastructure providers.
The catalyst is clear: AI companies are compute-constrained, not demand-constrained. Anthropic’s CEO has said as much. The limiting factor for growth isn’t whether customers want the product—it’s whether they can get enough computing power to serve them. That bottleneck creates sustained pricing power for infrastructure providers.
Top 3 Ways to Invest: ETFs, Stocks, and Alternatives
Strategy 1: Core Position — Nvidia (NVDA)
Nvidia remains the unavoidable play here. They supply the H100 and upcoming B100 GPUs that power most frontier AI models. Despite the run-up, the company is still growing data center revenue at 200%+ year-over-year, and Anthropic’s acceleration suggests that demand story has legs.
Entry consideration: Look for pullbacks toward $800-850 range for entry or scale in with 2-3% portfolio allocation on any 5-10% dip. This is a multi-year hold.
Risk level: Medium-high volatility, but reduced existential risk given proven revenue model and competitive moat.
Strategy 2: Diversified Infrastructure — VanEck Semiconductor ETF (SMH)
If single-stock concentration makes you nervous, SMH gives you weighted exposure to Nvidia, TSMC (chip manufacturing), ASML (chip equipment), and Broadcom (custom AI silicon) in one vehicle. The expense ratio is reasonable at 0.35%, and you get natural rebalancing.
This works especially well in tax-advantaged accounts where you’re buying and holding. The ETF smooths out individual company volatility while keeping you anchored to the sector.
Entry consideration: Dollar-cost average over 4-6 weeks, or buy on any market-wide risk-off event (geopolitical flare-up, Fed surprise).
Risk level: Medium, lower than single stocks, still correlated to tech sector overall.
Strategy 3: High-Conviction Specialty Play — Broadcom (AVGO)
Broadcom is underappreciated in the AI infrastructure stack. They design custom AI accelerators (ASICs) for hyperscalers who want alternatives to Nvidia’s off-the-shelf GPUs. The just-announced partnership with Google and Anthropic validates that