New York, Jan 20, 2026, 08:26 EST
Goldman Sachs stuck with Nvidia and Broadcom as its top picks in AI computing on Tuesday, saying Nvidia’s lead remains strong despite custom silicon closing the cost gap. The bank noted that Google’s Broadcom-designed TPU is making rapid gains on “inference” economics—the process of running AI models post-training—but highlighted Nvidia’s CUDA software and faster rollout as crucial advantages in the growing GPU-versus-ASIC battle. (ASICs are specialized chips made for a single task.) (Investing)
Nvidia investors face a turbulent period as the stock battles between being “still the backbone of AI” and concerns it’s “too big to run like it used to.” Shares slipped roughly 0.5% to $186.23 ahead of Tuesday’s U.S. market open.
Nvidia’s next key milestone is earnings. The company plans to release its fourth-quarter and full-year fiscal 2026 results on Feb. 25, per its investor relations calendar. (NVIDIA Investor Relations)
On Jan. 19, Motley Fool’s Will Healy labeled Nvidia a “buy now,” but cautioned the company’s massive scale might limit share-price jumps, even with strong growth. Healy noted Nvidia earns roughly 90% of its revenue from its data center segment. Analysts project about 63% revenue growth for fiscal 2026 and near 50% for 2027. He also highlighted a price-to-earnings ratio around 45. (The Motley Fool)
On Tuesday, a separate Motley Fool article took a different tack, focusing on volatility. Keith Speights pointed out that Nvidia has suffered several 20%-plus “crashes” in recent years, warning that another steep drop in 2026 wouldn’t be shocking, even if it’s short-lived. He pegged Nvidia’s forward P/E (based on expected earnings) at 24.3 and highlighted CEO Jensen Huang’s comment that “the transition to agentic AI and physical AI … will be revolutionary.” (The Motley Fool)
The bull case still hinges on demand. Nvidia’s latest quarterly report showed third-quarter revenue hitting $57.0 billion—up 62% year over year. The company projects fourth-quarter revenue around $65.0 billion, with a 2% margin of error. “Blackwell sales are off the charts, and cloud GPUs are sold out,” Huang said in the release. (NVIDIA Newsroom)
The road ahead isn’t smooth. Should major cloud clients continue moving workloads to in-house chips, or if export controls and geopolitical tensions tighten once more, Nvidia might see growth slow or margins squeeze—just as investors wrestle with defining “normal” for a company of this scale.
The debate has fractured into two camps: banks such as Goldman zeroing in on AI infrastructure’s relative positioning, while traders prepare for rapid sell-offs like those Nvidia has conditioned the market to anticipate. The next key date to watch is Feb. 25.
