|
|
 |
Apr 20, 2026 |
|
|
|
| Supported by |
 |
|
|
|
|
|
Happy Monday! Polymarket is talking to investors about raising $400 million at a roughly $15 billion valuation. Cloud platform Vercel confirmed a security breach. And the NSA is using Anthropic's Mythos model despite the Pentagon's blacklisting of the startup.
|
|
|
|
Predictions site Polymarket is in discussions with investors about $400 million in funding at around a $15 billion valuation, which would add to $600 million that Intercontinental Exchange recently invested in the company, The Information reported. This valuation is substantially lower than rival Kalshi’s $22 billion round. Kalshi’s higher valuation stems from its existing U.S. customer base and $1.5 billion in annualized revenue, while Polymarket just recently launched its U.S. platform and began charging fees. |
|
|
|
Vercel, a cloud development platform, on Sunday said a security breach allowed an attacker to gain unauthorized access to data for a “limited subset of customers.” In a post on X, Vercel CEO Guillermo Rauch said the attack began when a Vercel employee’s Google Workspace account was compromised through a breach at the AI platform Context.ai. “We believe the attacking group to be highly sophisticated and, I strongly suspect, significantly accelerated by AI,” Rauch wrote. “They moved with surprising velocity and in-depth understanding of Vercel.” Vercel said it had notified law enforcement and is directly contacting customers it has identified as affected. In September, investors valued Vercel at $9.3 billion in a funding round that raised $300 million for the company. Rauch recently said the company is ready for an initial public offering, but hasn’t decided on the timing for one yet. |
|
|
|
The National Security Agency is using Anthropic’s powerful new AI model, Mythos Preview, even though the agency’s parent, the Department of Defense, has sought to blacklist the AI startup from government contracts, Axios reported. Axios didn’t say how the NSA is using Mythos, though other organizations with access are primarily using it to scan their own environments for security vulnerabilities. One source told Axios that Mythos Preview is being used more widely within the Defense Department beyond the NSA. The continued use of Anthropic’s AI within the Defense Department stands in stark contrast to the Pentagon’s efforts recently to label the startup a “supply chain risk,” which ordinarily bars a company from doing business with the government. The Pentagon took action against Anthropic earlier this year after the company refused to allow its AI models to be used for autonomous weapons and mass surveillance. Anthropic has filed two federal lawsuits as a result. At the same time, there are signs of a potential thaw with other parts of Donald Trump’s administration. Anthropic CEO Dario Amodei on Friday met with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent to discuss Mythos. |
|
|
|
Jeff Bezos’ rocket company Blue Origin completed the third mission for its New Glenn rocket, which mostly went as planned. On Sunday morning, the rocket’s booster successfully landed on a barge in the Atlantic Ocean off the coast of Florida, the second time Blue Origin has accomplished that feat, which is a crucial step in the reusability of the rocket. Bezos posted a video clip on X of the landing. The second stage of the rocket also deployed its payload, an AST SpaceMobile satellite intended to expand AST’s direct-to-smartphone broadband network capacity. That satellite also powered on. However, Blue Origin in a post on X said that the satellite was placed in an “off-nominal orbit”—space jargon for a mishap. “We are currently assessing and will update when we have more detailed information,” Blue Origin said in the post. New Glenn’s progress is being closely followed in the space industry. If the company can successfully and frequently launch the rocket, it could provide a serious new competitor for Elon Musk’s SpaceX. New Glenn will jockey for rocket launch business with SpaceX’s new Starship rocket. |
|
|
|
Anthropic has received expressions of interest from investors who want to put money into the AI startup at an $800 billion valuation, but the company has no current plans to raise money. Those were among the details in a profile of Anthropic Chief Financial Officer Krishna Rao published by The Information. A decision on raising more money likely won’t happen until after Anthropic’s May board meeting, a person with knowledge of the company’s plans told The Information. Some Anthropic investors believe the company could fetch a $1 trillion valuation, which would represent more than 2.5 times its $380 billion valuation in January. Rao has guided Anthropic through multiple fundraising rounds since he joined Anthropic in 2024. He has also pushed the company to expand its relationship with multiple cloud computing and chip providers. |
|
|
|
Anthropic CEO Dario Amodei met with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent on Friday, in what both sides described as a “productive and constructive” discussion, according reports in Axios and The New York Times. The meeting signals a potential shift in the relations between the AI startup and the Trump administration after months of bitter conflict. The Trump administration previously declared Anthropic a “supply chain risk”—a designation historically reserved for foreign adversaries—after the company refused Pentagon demands to allow its AI models to be used for autonomous weapons and mass surveillance. Anthropic responded by filing two federal lawsuits, and a running court battle has continued since. The catalyst for the meeting was Anthropic’s new Mythos model, which the company has described as being capable of identifying previously unknown vulnerabilities in software. Some White House officials had argued that fighting Anthropic was counterproductive and was denying the U.S. government access to some of the most powerful AI tools available, particularly given Mythos’s potential to protect government networks from cyberattacks, The Times reported. The administration and Anthropic are expected to focus on how federal departments, other than the Pentagon, can gain access to Mythos, according to Axios. |
|
|
|
SemiAnalysis, the AI and semiconductor research firm, which also publishes a widely read industry newsletter, expects to do more than $100 million in sales this year, marked growth from a year earlier. SemiAnalysis did closer to $20 million last year, purveying in-depth data and research to startups, investors and the major hyperscalers, according to this week’s Big Read, which details how founder Dylan Patel, 29, has quickly achieved considerable influence in Silicon Valley. Patel has flourished by blurring the lines between tech, media and investing, somewhat similar to how podcasts “TBPN” and “All-IN” established their fandoms. He has been an active angel investor, flouting old-school conventions around conflicts of interest and accumulating stakes in startups like Mira Murati’s Thinking Machines Lab and more than a dozen others. He is currently trying to raise a venture capital fund and previously assembled a special purpose vehicle to invest in Fluidstack, a cloud provider. |
|
|
|
Andreessen Horowitz and Thrive Capital are in talks to co-lead a $2 billion investment in coding assistant startup Cursor at a $50 billion valuation before the investment, according to people familiar with the matter. The two firms first invested in the startup in 2024. Nvidia is also expected to participate in the fundraise, which hasn’t yet closed, the people said. The startup last raised $2.3 billion in a round led by Coatue Management and Accel in November, which valued the company at $27 billion, not including the investment. Founded in 2023, Cursor makes an AI coding assistant used by software developers to write and debug code, a breakthrough that ushered in the so-called “vibe coding” era. It’s become one of the fastest growing startups in Silicon Valley, and its annual recurring revenue reportedly soared to $2 billion as of February, up from $150 million a year ago. But pressure has intensified as OpenAI and Anthropic have improved their own coding tools, Codex and Claude Code, respectively, in the past year. Some of the details of the fundraise were previously reported by Bloomberg. |
|
|
|
Lawyers for Elon Musk said Friday that they may drop their fraud claims against OpenAI in his case charging the ChatGPT maker with violating its charitable mission and breaking promises to Musk. Dropping the claims would make Musk’s legal position more consistent with a court filing from earlier this month, in which Musk’s lawyers clarified that if OpenAI is found liable, they want OpenAI to transfer more than $100 billion from OpenAI’s business to its charitable foundation, rather than paying Musk. In contrast, money awarded for fraud would likely go to Musk, who claims he was personally deceived. The judge overseeing the case expressed skepticism about awarding money to OpenAI’s charity during Friday’s conference, which happened online. “I don’t know that you have the legal right to do what you’re saying that you want to do,” Judge Yvonne Gonzalez Rogers told a lawyer for Musk. If Musk “wants to make a charitable contribution to the entity, that’s on him, I don’t care. But he can’t come in here and say that he has the legal authority to have the court issue an order sending that money to the charity.” Judge Rogers went further, saying that if Musk prevails on his breach of charitable trust claim, any money from OpenAI might not go to Musk at all, but instead to the donor advised funds through which Musk funneled his donations. She asked Musk’s lawyers to submit a legal brief on these issues. Musk has alleged that OpenAI took his donations under promises that it did not keep, amounting to fraud and “constructive fraud,” which does not require showing an intent to defraud, making it easier to prove in court. If the claims are removed, that would leave only three remaining claims—unjust enrichment, breach of charitable trust and aiding and abetting breach of charitable trust—for the trial that is scheduled to begin with jury selection on Monday, April 27. |
|
|
|
Kevin Weil, a member of OpenAI’s management team and a former executive at both Meta Platforms and Twitter, is leaving the AI firm, according to a post the executive made on X. He is one of at least three OpenAI leaders to announce their departures in recent days. Weil, who has been at OpenAI since mid-2024, was at one time Chief Product Officer but more recently led the company’s work to build AI tools for scientists. The organization he led, OpenAI for Science, is being “decentralized into other research teams,” the post said. OpenAI’s tool for scientists, Prism, and the team behind it will be moving into OpenAI’x Codex coding product, according to a person with knowledge of the move. That move comes as OpenAI looks to unify its apps including Prism and its browser Atlas into a single “superapp,” the person said. Meanwhile, research scientist Bill Peebles, who led OpenAI’s work on its video generation model Sora, also said that he would be leaving the company in a post on X. His announcement comes a month after the company decided to shut down its Sora app and model efforts due to computational constraints. And on Friday afternoon, Srinivas Narayanan, whose title as been chief technology officer for B2B applications, said he was leaving the firm. Narayanan joined OpenAI in 2023 after 13 years at Facebook. The executive changes follow a concerted effort from the company to cut “side quests” and focus its efforts behind coding and selling AI to enterprises, as the company looks to regain its lead against rival Anthropic. (This article has been updated with Narayanan’s departure.) |
|
|
|
|
Popular articles
By Julia Hornstein and Cory Weinberg
By Kevin McLaughlin and Aaron Holmes
By Sri Muppidi, Valida Pau and Cory Weinberg
|
|
|
|
|
Opportunities Empower your teams to stay ahead of market trends with the most trusted tech journalism.
Learn more
Reach The Information's influential audience with your message. Connect with our team |
|
|
|
|
|
0 comentários:
Postar um comentário