|
|
|
Aug 26, 2026
|
|
|
|
|
Supported by
|
|
|
|
|
|
Happy Wednesday! SpaceX plans to build a $100 billion Starship launch site on the coast of Louisiana. Another OpenAI sales executive is leaving to go back to Salesforce. DeepSeek generates about $70 million in revenue in the first seven months of this year, roughly tenfold its full-year 2025 revenue.
|
|
|
SpaceX announced on Tuesday plans to build a $100 billion launch site for its Starship rocket on the coast of Louisiana. SpaceX plans to begin construction of the site, which the company has named “Starbase, Louisiana,” in 2027 and is targeting 2029 for the first Starship launch from the site. The company said the site will be built to support thousands of launches per year and Elon Musk wrote on X that the spaceport will allow SpaceX to launch “more than 30 Starship flights per day.” The site will allow SpaceX to expand its launch facilities outside of Texas and Florida. The deal was announced at an event held at the planned site with Louisiana Governor Jeff Landry and other officials on Tuesday.
|
|
|
OpenAI recruited heavily from Salesforce to sell its AI to big businesses. Now, some sales people are returning to Salesforce, deepening the shake-up in OpenAI’s enterprise sales team. Peter Doolan, the former chief customer officer of Slack who joined OpenAI in April, has signed an offer to return to Salesforce, according to a person with direct knowledge of the matter. Last week, Kaylin Voss agreed to rejoin Salesforce from OpenAI, where she led sales for the Americas, The Information reported. Another 22 former Salesforce employees who are currently employed at OpenAI are in active conversations to return to Salesforce, according to the person. Doolan will serve as Salesforce’s chief customer officer of its new AI labs initiative, which works to sell products from the company’s acquisitions. It bought supply chain automation startup Regrello in 2025 and this June said it would buy AI customer support startup Fin. Voss will take on the role of executive vice president of sales for the new team. The exits follow OpenAI’s announcement earlier this month that it was hiring Dali Rajic, president and chief operating officer of Alphabet-owned cybersecurity firm Wiz, to replace Denise Dresser as its new chief revenue officer. Dresser had recruited Doolan, Voss and other Salesforce employees to build OpenAI’s enterprise sales team as the AI firm worked to expand beyond a consumer-focused chatbot and catch up with Anthropic’s lead in selling AI to businesses. Doolan was the global head of AI transformation at OpenAI. He had been at Salesforce for 10 years before serving as chief customer officer at the Salesforce-owned messaging app Slack. Dresser tripled the size of OpenAI’s sales team, with enterprise sales accelerating following OpenAI’s release of its latest model to be released, GPT-5.6, in July, according to a person familiar with the matter.
|
|
|
DeepSeek generated about 475 million yuan ($70.7 million) in revenue in the first seven months of this year, roughly tenfold its full-year 2025 revenue, as the Chinese AI lab goes full steam ahead in its second round of funding, The Information reported on Wednesday. DeepSeek recorded a net loss of about 715 million yuan between January and July, compared to 935 million yuan for the entire 2025. The company is having ongoing discussions with existing and new investors about the funding round, which aims to raise 50 billion yuan at a valuation of 500 billion yuan. While DeepSeek’s revenue is still minuscule compared to U.S. AI developers such as Anthropic and OpenAI, the company is achieving a healthy gross profit margin. That’s because it has managed to keep the costs of running its models low, such as improving the efficiency of its AI infrastructure so that the AI systems can perform tasks using fewer chips. DeepSeek’s gross margin for the first seven months was 44.6%, and the gross margin for selling access to its models through application programming interface was 82.9%. DeepSeek’s revenue growth could give a boost to its funding talks and a potential initial public offering down the line. The company raised 50 billion yuan in its first-ever funding round closed in June and almost immediately started its second funding round. It has hired investment banks to help prepare a listing in Shanghai next year.
|
|
|
OpenAI on Tuesday released an evaluation of its new Jalapeño AI server chip, claiming it is faster and more powerful than Nvidia’s flagship Blackwell chips. The results appear promising for OpenAI, which hopes to reduce its reliance on Nvidia’s hardware someday, but it’s still early: The chips haven’t yet been tested as thoroughly against Nvidia’s latest chip, Rubin, nor have they been used at scale. And Jalapeño is focused on running existing AI models, not training them, whereas Nvidia’s graphics processing units are designed to do both. While Jalapeño is optimized to run OpenAI models, it can also run other models, the company said. OpenAI developed it relatively quickly with a team of around 100 people, said Richard Ho, OpenAI’s vice president of hardware. That’s in part because OpenAI used its most advanced AI models to help design the chip with Broadcom, he said. OpenAI intends to use the chip to run its own products, such as Codex and ChatGPT, according to someone who works on it. It couldn’t be learned whether the chips, whose production Broadcom planned to finance, would end up in the data centers of Microsoft and Oracle, which would then provide them to OpenAI. Microsoft had discussed buying billions of dollars worth of Jalapeno chips, while Oracle has separately said it could allow some of its cloud customers to “bring their own chips” to its data centers. OpenAI doesn’t yet run its own facilities, though it has separately worked out a large-scale agreement to use Nvidia hardware in a facility it would lease in Ohio starting in 2028, with around $100 billion in credit support from Nvidia. The chip leader also has invested tens of billions of dollars in OpenAI. While almost all of OpenAI’s work is powered by Nvidia hardware today, OpenAI has agreements to use AI chips made by Amazon, AMD and Cerebras as well. For now, Nvidia still dominates both the AI training and inference markets globally and even appears to have gained a greater share of the inference market over the past year. Nvidia didn’t respond to a request for comment about OpenAI’s claims about Jalapeno.
|
|
|
LayerZero, a blockchain infrastructure startup that counts Citadel Securities, DTCC, and Intercontinental Exchange among its partners, said it’s launching a new blockchain-based exchange targeting financial institutions this fall. Unlike major crypto exchanges, LayerZero’s new exchange, called ATLAS, will not have a consumer-facing application and will instead serve as the underlying exchange for other trading platforms and financial institutions. “We can provide this product—this market—here, and they open that up to all of the brokers, who then open those up to all of those customers,” said Bryan Pellegrino, co-founder and CEO of LayerZero. It will start by offering trading of spot crypto tokens and perpetual futures, with plans to support prediction contracts, futures and options later. LayerZero is the biggest blockchain bridge provider, allowing traders to move assets between different blockchains, and transferred nearly $9 billion in the past month. It was valued at $3 billion in a fundraising round in 2023 and counts Tether, a16z crypto, Sequoia Capital among its investors. The exchange will be built on Zero, the new blockchain announced by LayerZero in February. Citadel Securities, DTCC and Intercontinental Exchange, the parent company of the New York Stock Exchange, are among the partners exploring uses for the new blockchain. “We will have some of the world’s best market makers on ATLAS on day one providing liquidity,” said Pellegrino, who declined to specify which firms would participate. In April, LayerZero’s blockchain bridge suffered a $292 million hack which prompted customers of Aave, the largest decentralized lending platform, to withdraw deposits at Aave. LayerZero said it changed its configuration after the hack to require more than one single-verifier for transactions, and it changed its operational infrastructure where the compromise happened. It didn’t pay any damages following the hack.
|
|
|
|
|
| |
