Nvidia tops $5 trillion, closes at record $208

Nvidia reclaimed a market value above $5 trillion and finished the session at a record $208 per share after an intraday peak near $5.12 trillion. The stock rose as much as 4.2% during the day, a rally that added more than $200 billion to the company’s market capitalization and produced the highest closing price in the company’s history.

The buying came amid strength in semiconductor stocks, supported in part by upbeat earnings from Intel and a separate nuclear power contract that drew attention to companies tied to AI-related power demand. At its intraday high, Nvidia’s market value reached about $5.12 trillion, based on intraday market data. The $208 close remains slightly below the intraday record of $212.19 set on Oct. 29, 2025. Nvidia’s market capitalization is roughly $1 trillion larger than Alphabet’s.

Nvidia’s shares reversed a weak first quarter, when the stock fell 6.4%. April brought a rebound: shares rose about 20% over the past month and are up 101% over the past 52 weeks. By comparison, the S&P 500 gained 32.2% and the Technology Select Sector ETF rose 57.4% in the same 52-week period.

Investors are focused on Nvidia’s next earnings report, scheduled for May 20. Analysts expect diluted earnings per share of $1.70 for the quarter, up from $0.77 a year earlier, a projected increase of about 120.8%. Nvidia has beaten analyst forecasts in three of its last four quarterly reports and missed once. For the full fiscal year, analysts project $7.77 in earnings per share, up from $4.57 in fiscal 2026, and they expect earnings to reach $10.31 in fiscal 2028. Of 49 analysts covering the stock, 44 rate it Strong Buy, three rate it Moderate Buy, one rates it Hold and one rates it Strong Sell. The average price target is $268.80, implying roughly 34.6% upside from the closing price.

Nvidia released Ising, an open-source suite of AI models for quantum computing tasks such as qubit calibration and error correction. The company says Ising can cut qubit calibration time from days to hours and offers about 2.5 times faster real-time decoding and three times better accuracy than the pyMatching tool cited in comparisons. The models integrate with CUDA-Q and NVQLink, and firms including IonQ and Rigetti Computing, along with several research laboratories, have adopted the software.

In China’s AI ecosystem, DeepSeek published a preview of its V4 model adapted to run on Huawei chips. Huawei said the V4 preview is fully supported on Ascend 950–based supernode clusters and that its processors were used for part of V4-Flash training. DeepSeek previously trained models on Nvidia hardware but provided early access to local firms for the V4 preview and did not send the model to American chipmakers for tuning.

Nvidia’s late-April rally erased the weak start to the year in market-value terms and widened its lead over other large technology companies. The company’s quantum software release and parallel developments in global AI model deployments were among the notable technology events on the same day.