Nvidia dominates AI in 2025 with revenue of US$187 billion and peak of US$5 trillion

Nvidia consolidated dominance in the AI market in 2025 with record revenue of US$187.1 billion and a valuation that briefly surpassed US$5 trillion. The company has led innovations in advanced chips and expanded global partnerships, driving the artificial intelligence sector on an industrial scale.
The year marked Nvidia’s transition into the spotlight in Wall Street and the technology industry. Investidores closely followed each announcement and strategic move by CEO Jensen Huang. The company faced volatility in the market, but quickly recovered positions thanks to continued demand for its processors.
The focus remained on the evolution of AI, with advances in model training and inference. The company has expanded its presence in data centers, robotics and accelerated computing. Esses developments have positioned Nvidia as a key player in the global artificial intelligence infrastructure.
Keynote announcements at CES 2025
Nvidia started the year with a bang at CES 2025, in Las Vegas. Jensen Huang introduced a lineup of innovations that included new compact chips for AI and more powerful graphics cards for gaming. The talk covered topics such as physical AI, focused on robotics, and AI systems for desktops.
Software updates were highlighted as a basis for the company’s future performance. The announcements generated expectation in the market. Participantes observed the potential for integration between hardware and practical applications in various sectors.

Initial shock with the DeepSeek R-1 model
January brought a significant impact to the sector. Chinese startup DeepSeek launched the R-1 model, claiming training with lower-than-state-of-the-art computing resources. The news caused a sharp drop in Nvidia shares, with a loss of almost US$600 billion in market value in a single day.
The episode sparked debates about dependence on high-performance chips. Nvidia argued that model training and execution differ in requirements. More advanced Chips improve overall inference performance. The market stabilized over the following months. Stocks only regained previous levels in June.
Reveal of Blackwell Ultra chips at GTC
The March GTC, on San Jose, turned into a large-scale event. Jensen Huang’s talk filled SAP Center and occupied large areas of the city center. The Nvidia introduced the Blackwell Ultra chip, successor to the Blackwell GPU, and the GB300 superchip.
The GB300 combines two Blackwell GPUs with an Grace CPU. Huang highlighted the focus on the era of AI reasoning. Modelos like DeepSeek’s R-1 would gain superior performance with this new generation. The structure of the event reflected the growing relevance of the conference.
Strategic partnership in Coreia of Sul
October marked an informal meeting between Jensen Huang, the chairman of Samsung Electronics Jay Y. Lee and the executive chair of Hyundai Motor Group Euisun Chung. Dinner took place at a fried chicken and beer restaurant in Seul, a local tradition known as chimaek.
The moment preceded announcements of cooperation in AI infrastructure. Nvidia has strengthened ties with South Korean technology and automotive industries. Acordos involved supplying GPUs for national artificial intelligence projects.
Financial growth and historical valuation
Annual revenue reached US$187.1 billion, driven by demand for data center solutions. Market capitalization surpassed US$5 trillion in October, an unprecedented feat for any listed company. The value stabilized in the US$4 trillion range at the end of the year.
The trajectory reflects the central role of Nvidia in the expansion of AI. Investimentos in partnerships and innovations sustained performance. The year consolidated the company’s leadership in providing essential infrastructure for global technological advancement.
Nvidia ended 2025 with a strengthened positioning. Avanços in hardware and software kept the company at the center of discussions about the future of artificial intelligence. The impact extends to varied sectors, from cloud computing to industrial applications.






