TLDR
- Citi maintains Buy rating on Nvidia with $270 price target, expecting stock to outperform in second half of 2026 as demand visibility extends into 2027
- Analyst projects January-quarter revenue of $67 billion versus Street consensus of $65.6 billion, with April-quarter guidance of $73 billion versus $71.6 billion consensus
- Nvidia’s Blackwell AI systems deliver up to 50 times higher throughput per megawatt than previous Hopper generation, lowering inference costs by up to 35 times
- Susquehanna analyst Christopher Rolland maintains Positive rating with $250 target, citing strong GB300 ramp through first half of 2026
- Upcoming Vera Rubin hardware expected to deliver 10 times higher throughput per megawatt compared to Blackwell systems
Nvidia stock dropped 1.6% to $179.91 in early Tuesday trading despite bullish Wall Street commentary pointing to strong earnings potential. The chipmaker’s shares are down 2% year-to-date through Friday’s close.
Citi analyst Atif Malik told investors to add positions ahead of the company’s February 25 earnings report. The bank expects January-quarter revenue of $67 billion, above the Street consensus of $65.6 billion.
Malik projects April-quarter sales guidance of $73 billion versus analyst expectations of $71.6 billion. The firm anticipates a 34% second-half acceleration in 2026 sales compared to 27% in the first half.
The analyst pointed to continued ramp of the B300 and Rubin platform as key drivers. Citi maintained its Buy rating and $270 price target on the stock.
Most investors are looking past quarterly results to Nvidia’s annual GTC conference in mid-March. The company is expected to outline its inference roadmap using Groq’s low-latency SRAM IP at the event.
Malik anticipates fiscal 2027 gross margin guidance around 75%. He assumes operating expense growth in the high-30% range, similar to fiscal 2026.
Blackwell Systems Show Major Efficiency Gains
Nvidia celebrated new performance data from semiconductor research firm SemiAnalysis this week. The data showed Blackwell systems deliver up to 50 times higher throughput per megawatt than the previous Hopper generation.
The efficiency improvement translates to lower costs across the entire latency spectrum. At low latency where agentic applications operate, costs drop by up to 35 times per million tokens compared to Hopper.
The company’s upcoming Vera Rubin hardware will push efficiency even further. Nvidia said Rubin will deliver up to 10 times higher throughput per megawatt compared to Blackwell systems.
Susquehanna analyst Christopher Rolland issued a separate positive note Tuesday morning. He maintains a Positive rating and $250 target price on the stock.
Rolland expects better results and guidance as GB300 Blackwell hardware continues ramping through the first half of 2026. He cited ever-increasing hyperscale capital expenditure plans supporting AI demand.
Key Questions About Market Position
The analyst noted the key question is whether Nvidia can exceed its previous projection of $500 billion in cumulative revenue. That figure covered Blackwell and next-generation Rubin chips from early 2025 through 2026.
One major question facing Nvidia is maintaining market leadership as demand shifts toward inference over training. Inference generates answers or results from AI models and is expected to surge with growth of AI agents.
Citi argued that rising hyperscaler capital spending will deliver long-term returns as AI-driven infrastructure demand accelerates. The bank stated increased competition in inference is natural but expects Nvidia to remain the leader across both training and reasoning-focused inference workloads.
Rolland wrote that he still views Nvidia as having one of the largest opportunity sets ahead. The stock is looking for a catalyst to move upward again after early 2026 weakness.
Malik expects the company to provide an early outlook for 2026 and 2027 AI sales at the March GTC conference. Advanced Micro Devices stock fell 4.9% and Broadcom dropped 2.4% in early Tuesday trading.


