Introduction
After an intense multi-year rally, artificial intelligence investing has entered a quieter phase as markets question how quickly massive compute spending will translate into profits. This pause has pushed several high-profile AI-linked stocks lower, even as companies continue to describe AI infrastructure as essential spending rather than optional experimentation. In that setting, three names often highlighted as “buy-the-dip” candidates are Microsoft, Broadcom, and Nebius, each tied to a different layer of the AI stack: cloud platforms, custom silicon, and AI-first compute services.
Microsoft: Selloff Meets a Cloud-Driven AI Narrative
Microsoft has remained central to the AI buildout through its cloud business, especially Azure, which benefits directly from higher demand for AI training and inference workloads. Despite reporting strong results for the second quarter of fiscal 2026, the stock has pulled back sharply, described here as roughly 30% below its all-time high. The argument made by bullish investors is valuation: Microsoft’s price-to-earnings level is portrayed as unusually low compared with recent history, implying the market may be discounting AI spending risk more heavily than the company’s current profit engine justifies.
The core question for investors is timing. AI infrastructure can lift revenue quickly for cloud providers, but the biggest gains may depend on enterprise adoption cycles that play out over years. That gap between spending now and payoffs later is one reason Microsoft and other AI leaders can see volatility even when near-term financial performance looks healthy.
Broadcom: Custom AI Chips and Hyperscaler Demand
Broadcom is positioned differently, with a key growth driver in its custom AI chip business. The thesis is that hyperscalers want alternatives to one-size-fits-all accelerators in certain workloads, and custom silicon can improve efficiency or cost performance depending on use case. Broadcom’s shares have also declined, described here as about 20% off recent levels, even as growth expectations remain high.
Analysts cited in this view project rapid expansion across the next two fiscal years, with expectations of about 53% revenue growth in fiscal 2026 and about 39% in fiscal 2027. If realized, those figures imply a steep growth curve powered by hyperscaler partnerships and continued AI infrastructure demand. The key risk is execution and concentration: custom chip cycles are capital intensive and customer-driven, so momentum depends on sustained hyperscaler spending and successful product delivery.
Nebius: AI-First Cloud With Aggressive Capacity Buildout
Nebius is framed as a smaller but faster-growing compute platform focused on AI developers who want a full-stack environment for building and running models. The growth claim rests on rapid capacity expansion through new data centers and site additions. In this account, Nebius had an annual run rate of roughly $1.25 billion at the end of 2025 and is expected to reach $7 billion to $9 billion by the end of 2026 as more infrastructure comes online.
The company’s buildout is described as scaling from two sites in 2024 to seven in 2025, with a target of 16 operational sites by the end of 2026. Supporters argue that demand for AI compute remains strong and could continue rising as AI tools spread across industries. Nebius shares are described as down about 25% from highs set in October 2025, with the claim that the pullback creates an attractive entry if growth ramps as projected.
Conclusion
The current AI pullback reflects a market that wants near-term returns while AI leaders prioritize long-term infrastructure investment. Microsoft offers a mature cloud platform with AI demand tailwinds but valuation sensitivity. Broadcom ties its upside to hyperscaler-driven custom silicon adoption. Nebius represents a higher-risk, higher-growth compute buildout story. The main implication is that investors are now weighing the same question across the AI stack: whether today’s spending wave turns into durable earnings power fast enough to justify the next leg higher.
