Wall Street received a double dose of artificial intelligence (AI) infrastructure news on Tuesday, which sent shares of both Nvidia (NVDA 4.86%) and CoreWeave (CRWV 8.21%) tumbling.
Specifically, SoftBank Group disclosed it had dumped its entire $5.8 billion Nvidia stake in October, driving Nvidia's shares lower by 2%. Hours later, CoreWeave slashed its full-year revenue guidance by $100 million to $200 million due to a data center delay. In response, the stock ended the day down by nearly 17%.
The market's reaction? Panic. But for investors with conviction about AI's multiyear buildout, Tuesday's dual sell-off in two essential infrastructure plays might represent rare buying opportunities. Here's why I'd buy both of these core AI infrastructure stocks right now.
Image source: Getty Images.
The CUDA moat competitors can't cross
Nvidia has become synonymous with AI computing, commanding the market for graphics processing units (GPUs), the specialized chips that power the training of large AI models and run them in production. But the real advantage isn't just chip performance. It's the Compute Unified Device Architecture (CUDA) software ecosystem that creates switching costs competitors can't overcome. When tech giants like Microsoft and Alphabet build entire AI infrastructures around Nvidia's tools, migrating becomes prohibitively expensive.
NASDAQ: NVDA
Key Data Points
The SoftBank disclosure on Tuesday certainly raised eyebrows. The Japanese conglomerate revealed it sold 32.1 million Nvidia shares last month to fund its $30 billion-plus investment in OpenAI. Some analysts interpreted the move as a signal that AI valuations have stretched too far. Yet, data center customers continue to expand GPU capacity, and Nvidia remains the default choice.
That said, Nvidia's shares do trade at a premium to the S&P 500 (^GSPC 1.55%) by most forward-looking valuation metrics, but that premium is well deserved, given the AI hyperbuild underway. The real question isn't whether Nvidia will remain dominant, but whether Tuesday's 2% drop creates an opportunity.
For investors who believe AI computing demand will accelerate for years rather than quarters, this SoftBank-induced dip might be a rare chance to buy the central bank of the AI revolution at a discount.
Building the plumbing beneath the chips
CoreWeave occupies a distinct layer of the AI stack, featuring specialized cloud infrastructure optimized for GPU workloads. The company positions itself as "the essential cloud for AI," building data centers specifically designed for large-scale AI computing. This offers more leverage to AI growth than chip design alone, but introduces execution risk that pure semiconductor plays avoid.
NASDAQ: CRWV
Key Data Points
CoreWeave's 2025 third-quarter revenue reached approximately $1.36 billion, accompanied by a substantial backlog of committed future business, indicating strong underlying demand. However, management lowered full-year 2025 guidance to $5.05 billion to $5.15 billion due to delays at a third-party data center. This piece of news was the main culprit behind Tuesday's double-digit pullback.
What's the investing picture look like? CoreWeave stock trades at a steep premium to the S&P 500 (64 times projected 2027 earnings), and this data center delay doesn't help matters. But demand for AI computing is only going to ramp from here. That's why I think CoreWeave stock is a buy on this dip for investors with a five- to 10-year horizon.
The infrastructure opportunity hiding in turbulence
Tuesday's dual sell-off doesn't change the fundamental AI story. Nvidia offers the dominant player approach with an entrenched moat, while CoreWeave provides leveraged exposure to the physical infrastructure buildout. Together, these two core AI infrastructure stocks provide direct exposure to this megatrend.
For investors willing to accept volatility and hold through the multiyear buildout phase, Tuesday's panic should prove to be a gift. After all, the AI computing wave isn't cresting. It's gathering strength and will continue to do so through the end of the decade.
