OpenAI has registered another senior executive departure with the dismissal of Denise Dresser from her role as CRO on Thursday, occurring just two days after Brad Lightcap announced his departure. These movements are happening while the company prepares for a possible IPO and maintains a valuation of US$ 852 billion (approximately R$ 4.37 trillion).
Furthermore, Fidji Simo also left the organization. According to CNBC, these three departures occur amid competition with Google and Anthropic, the expansion of more accessible open-source models, and investor expectations regarding OpenAI's future.
Dresser joined OpenAI in December, bringing with her over a decade of experience in leadership roles at Salesforce. Her main responsibility was to drive corporate business, a sector considered strategic in the rivalry with Anthropic.
In April, four months after her hiring, Dresser began taking on some of Lightcap's responsibilities. He had left the Chief Operating Officer position to focus on special projects and communicated this week the end of his eight years at the company to start new ventures.
Dresser's replacement will be Dali Rajic, who previously held the position of President and COO at Wiz, a cybersecurity company acquired by Google for US$ 32 billion (about R$ 164.16 billion).
Greg Brockman, CEO of OpenAI, commented that Dresser led the revenue area during a 'business formation period.' Regarding Rajic, he expressed the expectation that she can transform the company's learning into a more stable operation.
Despite the changes in leadership, OpenAI managed to present positive results in the enterprise market. Brockman reported that recurring revenue exceeded 20% in July, and penetration among corporate clients reached 32%. The enterprise division reached 2 million customers, doubling the number recorded the previous year.
In an interview with CNBC, Denise Dresser, OpenAI's CRO, stated that corporate business represented 40% of revenue and had the potential to match consumer share by the end of 2026.
The data presented by OpenAI includes the fact that the company confidentially submitted its IPO prospectus in June, although it has not yet set a date for the offering. The executive departures have attracted investor attention; Kevin McCormick, founder of SignAudit.AI, classified this wave of pre-IPO dismissals as a 'big warning sign,' and two investors told CNBC that they were surprised by Dresser's decision.
Although rapid changes are not unprecedented at OpenAI, in 2023, Sam Altman was removed from the CEO position after the board alleged that he was not 'consistently sincere in his communications.' Subsequently, he returned to the position due to the reaction of investors and employees.
This incident came back into focus during the lawsuit filed by Elon Musk against OpenAI, Altman, and Brockman, where the management of these two executives was questioned by Musk's lawyers during the trial. With new departures in management, Altman and Brockman need to maintain the confidence of investors and employees while OpenAI seeks to advance in the corporate market and prepares for a potential capital opening.
According to data from the maritime tracking service TankerTrackers, an Iranian tanker was observed on Friday for the first time since late July loading crude oil at Hargh Island.
At Hargh Island, in the Azarpad port, a supertanker belonging to the National Iranian Tanker Company (VLCC) was spotted loading two million barrels of crude oil on August 14.
TankerTrackers also noted that Iran continues to load tankers at other terminals. In the service's opinion, the country's oil production roughly matches the volumes processed at refineries or domestic consumption, which may indicate that Iran does not need to export significant volumes of crude oil.
Nvidia and OpenAI are close to finalizing an agreement to fund a vast data center complex located in Ohio, United States, but the deal structure has been modified. In the new proposal, the semiconductor manufacturer will only guarantee half of the originally planned project, instead of committing to the entirety of the construction.
According to sources close to the negotiations who spoke with The Wall Street Journal, this change reduces Nvidia's financial guarantee from US$ 250 billion (equivalent to approximately R$ 1.35 trillion) to less than US$ 120 billion (about R$ 650 billion).
This revision was implemented to address investor concerns regarding Nvidia's financial exposure, given that the company uses its balance sheet to drive demand for its artificial intelligence (AI) chips. There is a possibility that the agreement could be signed this weekend.
Nvidia's shares fell by 5% after the Journal initially disclosed details of a US$ 250 billion guarantee.
The undertaking, which has a capacity of ten gigawatts, is being developed by SB Energy, a subsidiary of the Japanese conglomerate SoftBank. Once completed, this complex will be the largest data center project ever announced.
Nvidia's initial guarantee aimed to cover both the data center lease contract and the loans needed to finance its construction. The purpose was to lower financing costs, giving creditors greater certainty that the funds would be available to maintain the project.
In addition to the infrastructure guarantee, Nvidia is negotiating a separate agreement to finance the purchase of its chips by OpenAI. This deal could reach US$ 350 billion (approximately R$ 1.9 trillion) when considering the full scope of the project.
The energy demanded by the undertaking is regulated by the United States government and will be financed separately by Japan, as part of a recent trade agreement between the two nations.
Goldman Sachs is advising SB Energy during the negotiations, while Morgan Stanley acts as advisor to Nvidia.
These negotiations take place a few days after Nvidia announced a collaboration with major financial institutions to establish what it termed 'computing financing platforms.' This initiative includes Apollo Global Management, Blackstone, BlackRock, Brookfield Asset Management, Goldman Sachs, and KKR. Together, these corporations plan to make over US$ 500 billion (about R$ 2.7 trillion) in external capital in the coming years to support the purchase of Nvidia chips at favorable rates.
This strategy allows Nvidia to help its customers finance the vast infrastructure required for AI expansion while fostering conditions to increase demand for its own components.
The current negotiation represents a restructuring of an agreement previously communicated by the two companies. Last September, Nvidia and OpenAI had established a memorandum of understanding whereby the chip manufacturer was to build a minimum of ten gigawatts of computational capacity for OpenAI. At that time, Nvidia had also agreed to contribute up to US$ 100 billion (approximately R$ 542 billion) to help OpenAI pay for the infrastructure. However, this agreement was suspended after members of Nvidia itself raised questions about the operation, as reported by the Journal.
Currently, the companies are seeking to proceed with a different architecture, where Nvidia would initially assume a smaller financial exposure, postponing the definition of how to finance the remaining ten gigawatts to a later phase.
Exports of electric vehicles from China are growing faster than the capacity of the global maritime fleet available for their transport, which is evidenced by high freight costs. In June, the cost to charter a large car carrier reached an average of US$ 70,000 per day, a significant increase compared to the US$ 42,500 recorded at the end of last year, according to data from the broker Clarksons. Over 2026, these rates registered a growth of 65%.
Due to the lack of space on specialized ships, some manufacturers were forced to use conventional containers, the same ones used for transporting electronics and furniture.
The volume of Chinese exports has skyrocketed: in 2019, China exported less than 600,000 vehicles, but this target is expected to reach 10 million in the current year. Figures for 2025 were 7.1 million units, and in June alone, the country surpassed the mark of 1 million units in a single month, representing a jump of 71.2% compared to the previous year.
Andreas Enger, CEO of Höegh Autoliners, told The Wall Street Journal that in just five years, China transitioned from a secondary automotive exporter to the world leader in this segment. He also pointed out that the sea freight for cars currently costs double what it did before the pandemic.
Although ship supply has shown some reaction, it has failed to keep pace with the increased demand. For Lasse Kristoffersen, CEO of Wallenius Wilhelmsen, which operates the largest fleet in the sector, the approximately 40% expansion of the global fleet is still insufficient to meet Chinese shipments. Furthermore, the capacity of ships dedicated to cars and trucks is only expected to increase by 7.6% this year.
With domestic sales falling by more than 20% in the first half, exports have become a decompression channel for the production surplus of Chinese automotive industries.
The use of containers is no longer an atypical situation. Kristoffersen estimates that up to 4 million vehicles annually leave China via containers or other alternatives to car carriers. The consultancy Veson Nautical predicts that this number of cars in alternative modes will double in 2026, reaching about 2 million, treating this practice as a structural characteristic of trade.
Eric Dessupoiu, Vice President of Vehicle Logistics at Ceva Logistics, clarified that car carriers are preferred because the process of loading and unloading driven cars is more economical and minimizes the risk of damage. In the case of containers, the vehicle needs to be transported to a yard near the port, placed in the box, and lifted by a crane, with the reverse procedure occurring at the destination.
Christoph Seitz, Global Vice President of Finished Vehicles at DP World, observed a divergence in approach: while Western manufacturers showed resistance to the idea, Chinese ones did not present this objection. Carriers such as the Danish Maersk and the Swiss MSC already offer this service directly to manufacturers.
Some Chinese companies opted not to rely on third parties; BYD, for example, launched its own car carrier in 2024 and currently operates eight vessels. The company did not respond to The Wall Street Journal on the matter.
Lasse Kristoffersen informed investors that companies with their own fleets have their capacity fully utilized and need to decide which customers will receive space leaving Asia, with Europe, Australia, Latin America, and Brazil being among the main destinations.
