Nvidia announced on Monday (28th) that its board of directors approved an additional $150 billion, equivalent to about R$ 795 billion, designated for the company's share repurchase program. With this decision, the total amount authorized for buybacks reaches $235 billion, approximately R$ 1.25 trillion, making it the largest program of its kind ever registered in the United States, according to the company and the market.
This new authorization follows the addition of $80 billion (about R$ 424 billion) to the program four months ago. This strategic move comes as Nvidia generates significant revenue due to strong demand for semiconductors aimed at artificial intelligence (AI).
Statements on AI-Driven Growth
Jensen Huang, CEO of Nvidia, stated that the company is rising thanks to a platform transition, which occurs in a generational cycle, towards accelerated computing and AI. The executive added that the company's cash generation allows it to invest in technologies that promote this transformation and, simultaneously, return capital to shareholders, reflecting its confidence in the long-term opportunity.
Share buybacks function as a mechanism for returning capital to shareholders; by removing securities from the market, the operation can decrease the available supply and consequently help increase the stock price. Vivek Arya, a research analyst at Bank of America, observed that such buybacks often signal that a corporation believes its shares are undervalued. Arya told The New York Times that by repurchasing its own shares, he is conveying his conviction in the company's future.
In the context of Nvidia, this move coincides with intense debates about the vast amount of capital being channeled into the AI sector. The rapid advancement of the industry has forced companies to invest billions in data centers and the development of new models. However, critics point out that these funds could be applied to other areas, such as personnel expansion, infrastructure construction, or research and development activities.
Nvidia reported that it plans to complete the execution of the remaining balance of its buyback program by the end of the 2028 fiscal year, which ends in January. In addition to buybacks, Nvidia's financial strategy involves using its cash to invest directly in technology startups. A recent presentation detailed that Nvidia's investment portfolio comprises 13 publicly listed companies and 229 private companies, with a return on investments already made exceeding three times the invested amount.
The investments include cloud computing companies that use Nvidia chips and AI model developers. The company also foresees returning excess free cash flow to shareholders both through buybacks and through progressively increasing dividends. Furthermore, Nvidia has begun participating in financing operations for AI infrastructure projects. Last month, the company signed a partnership with Wall Street institutions to partially secure up to $500 billion (about R$ 2.65 trillion) in financing for data centers, aiming to facilitate customers' access to the necessary capital to acquire its chips.
Nvidia also provided financial support for OpenAI's data center project located in Ohio, USA. Despite the magnitude of these deals, new questions have arisen regarding possible circular financing schemes, where entities within the same ecosystem mutually assist in financing. Colette Kress, the company's CFO, justified this support as essential to drive the sector's growth cycle during the earnings release at the end of August. The company also celebrated long-term lease agreements for data centers totaling approximately $20 billion (about R$ 106 billion), which will subsequently be reallocated to other companies.
Launch of Security Platform for AI Agents
Beyond financial movements, Nvidia launched a new software platform focused on the security of AI agents on Monday. The Nvidia Open Agent Safety Platform was developed to assist developers in testing and implementing AI agents in controlled environments. This tool offers functionalities to continuously monitor system behavior and define rules for their actions.
This launch occurs amid growing apprehension about AI agents that can perform tasks autonomously and which, in recent cases, have been linked to intrusion attempts and attacks. Nvidia emphasized that recent incidents involving agents deviating from expected behavior reinforced the need for this tool. Jensen Huang stressed that the extraordinary potential of AI for society will only be fully realized if AI safety is resolved.
Nvidia also advocates for the freedom of developers in creating open-source AI models, which can be downloaded and modified. The company shows interest in this ecosystem because open models have the potential to increase demand for its chips. Simultaneously, the debate about the security of these systems gained momentum after the AI agent incidents, intensifying discussions about the need to establish limits for technological development.