OpenAI said Monday it’s releasing its buzzy AI video-generation tool, Sora, later in the day.
The AI video-generation model works similarly to OpenAI’s image-generation AI tool, DALL-E: A user types out a desired scene, and Sora will return a high-definition video clip. Sora can also generate video clips inspired by still images and extend existing videos or fill in missing frames. The Microsoft-backed artificial intelligence startup, which burst into the mainstream last year thanks to the viral popularity of ChatGPT, introduced Sora in February.
It’ll debut to U.S. users as well as to “most countries internationally” later today, according to OpenAI’s YouTube livestream, and the company has “no timeline” yet for launching the tool in Europe and the U.K., as well as some other countries.
OpenAI said users don’t need to pay extra for the tool, which will be included in existing ChatGPT accounts such as Plus and Pro. Employees on the livestream and OpenAI CEO Sam Altman demonstrated features like “Blend” (i.e., joining two scenes together at the user’s direction), as well as the option to make an AI-generated video endlessly repeat.
Until now, Sora has mainly been available to a small group of safety testers, or “red-teamers,” who test the model for vulnerabilities in areas such as misinformation and bias.
Reddit users asked OpenAI executives in October about Sora’s release date, questioning whether it was being delayed “due to the amount of compute/time required for inference or due to safety.” In response, OpenAI’s product chief Kevin Weil wrote, “Need to perfect the model, need to get safety/impersonation/other things right, and need to scale compute!”
“We obviously have a big target on our back as OpenAI,” Rohan Sahai, OpenAI’s Sora product lead, said on the livestream, adding that the company needs to prevent illegal use of the technology. “But we also want to balance that with creative expression.”
OpenAI closed its latest funding round in October at a valuation of $157 billion, including the $6.6 billion the company raised from an extensive roster of investment firms and Big Tech companies. It also received a $4 billion revolving line of credit, bringing its total liquidity to more than $10 billion.
It’s all part of a serious growth plan for OpenAI, as the Microsoft-backed artificial intelligence startup battles Amazon-backed Anthropic, Elon Musk’s xAI, Google, Meta, Microsoft and Amazon for the biggest slice of the generative AI market, which is predicted to top $1 trillion in revenue within a decade.
Earlier this month, OpenAI hired its first chief marketing officer, indicating plans to spend more on marketing to grow its user base. And in October, OpenAI debuted a search feature within ChatGPT that positions it to better compete with search engines like Google, Microsoft‘s Bing and Perplexity and may attract more users who otherwise visited those sites to search the web.
With Sora, the ChatGPT maker is looking to compete with video-generation AI tools from companies such as Meta and Google, which announced Lumiere in January. Similar AI tools are available from other startups, such as Stability AI’s Stable Video Diffusion. Amazon has also released Create with Alexa, a model that specializes in generating prompt-based short-form animated children’s content.
Video could be the next frontier for generative AI now that chatbots and image generators have made their way into the consumer and business world. While the creative opportunities will excite some AI enthusiasts, the new technologies present serious misinformation concerns as major political elections occur across the globe. The number of AI-generated deepfakes created has increased 900% year over year, according to data from Clarity, a machine learning firm.
OpenAI has made multimodality — the combining of text, image and video generation — a prominent goal in its effort to offer a broader suite of AI models.
News of Sora’s release follows protestors’ decision to leak what appeared to be a copy of Sora over concerns about the ChatGPT maker’s treatment of artists.
Some members of OpenAI’s early access program for Sora, which it said included about 300 artists, published an open letter in late November critiquing OpenAI for not being sufficiently open or supporting the arts beyond marketing.
“Dear corporate AI overlords,” the protestors’ open letter stated, “We received access to Sora with the promise to be early testers, red teamers and creative partners. However, we believe instead we are being lured into ‘art washing’ to tell the world that Sora is a useful tool for artists.”
The letter added that hundreds of artists provided unpaid labor for OpenAI through bug testing and feedback on Sora, and that “while hundreds contribute for free, a select few will be chosen through a competition to have their Sora-created films screened — offering minimal compensation which pales in comparison to the substantial PR and marketing value OpenAI receives.”
“We are not against the use of AI technology as a tool for the arts (if we were, we probably wouldn’t have been invited to this program),” the open letter stated. “What we don’t agree with is how this artist program has been rolled out and how the tool is shaping up ahead of a possible public release. We are sharing this to the world in the hopes that OpenAI becomes more open, more artist friendly and supports the arts beyond PR stunts.”
In late November, an OpenAI spokesperson responded to the protestors’ actions in a statement to CNBC.
“Hundreds of artists in our alpha have shaped Sora’s development, helping prioritize new features and safeguards,” the OpenAI spokesperson said at the time. “Participation is voluntary, with no obligation to provide feedback or use the tool. We’ve been excited to offer these artists free access and will continue supporting them through grants, events, and other programs.”
Meta on Tuesday announced it will eliminate its third-party fact-checking program to “restore free expression” and move to a “Community Notes” model, similar to the system that exists on Elon Musk‘s platform X.
The company said Community Notes will be written and rated by contributing users to provide more context to posts across its platforms, and the feature will roll out in the U.S. over the next couple of months. The announcement marks Meta’s latest attempt to smooth over relations with Republican President-elect Donald Trump before he takes office.
“We’ve reached a point where it’s just too many mistakes, and too much censorship,” Meta CEO Mark Zuckerberg said Tuesday in a video announcement. “The recent elections also feel like a cultural tipping point towards once again prioritizing speech, so we’re going to get back to our roots and focus on reducing mistakes, simplifying our polices and restoring free expression on our platforms.”
Zuckerberg said the third-party fact-checkers have been “too politically biased” and have “destroyed more trust than they’ve created, especially in the U.S.”
Meta said it will simplify its content policies by removing restrictions on subjects like immigration and gender and implement a new approach to policy enforcement that will focus on illegal and high-severity violations. The company is moving its trust and safety and content moderation teams from California, a historically Democratic state, to Texas, a historically Republican state.
“We’re going to work with President Trump to push back on governments around the world that are going after American companies and pushing to censor more,” Zuckerberg said.
Federal Trade Commission Chair Lina Khan addressed Meta’s announcement in an interview Tuesday on CNBC’s “Squawk Box,” stating, “We should have an economy where the decisions of a single company or a single executive are not having extraordinary impact on speech online.”
Joel Kaplan, Meta’s head of global policy, appeared Tuesday on Fox News’ “Fox and Friends” and said Meta thinks the Community Notes system on Musk’s platform X has been working “really well.” Musk, who has been a vocal advocate for Trump online and donated millions of dollars to his campaign, has been in close contact with the president-elect since the election.
Last week, Meta said that Kaplan would become the company’s top policy officer, succeeding Nick Clegg, who was a former British deputy prime minister and a leader of Britain’s centrist Liberal Democrats party.
Kaplan, who has held several policy-related positions at Meta since joining the company in 2011 when it was still named Facebook, is well known within the Republican Party. He was a White House deputy chief of staff under former President George W. Bush and also once worked as a law clerk for former Supreme Court Justice Antonin Scalia.
In December, Kaplan revealed in a Facebook post that he joined Vice President-elect JD Vance and Trump during their recent visit at the New York Stock Exchange.
“We want to make it so that, bottom line, if you can say it on TV, you say it on the floor of Congress, you certainly ought to be able to say it on Facebook and Instagram without fear of censorship,” Kaplan said Tuesday.
Meta’s Oversight Board, which provides an independent check of the company’s content moderation, lauded the company’s changes on Tuesday.
“The Oversight Board welcomes the news that Meta will revise its approach to fact-checking, with the goal of finding a scalable solution to enhance trust, free speech and user voice on its platforms,” the board told CNBC in a statement, adding that “specifically in the United States, rightly or wrongly, Meta’s previous approach has been perceived as politically biased by many of its users.”
Prominent Republican lawmakers have previously criticized Meta and other technology companies for allegations regarding the censorship of conservative voices on their respective platforms. For instance, House Judiciary Chair Jim Jordan, R-Ohio, subpoenaed Zuckerberg and other tech CEOs in 2023 as part of a probe to “understand how and to what extent the Executive Branch coerced and colluded with companies and other intermediaries to censor speech.”
Zuckerberg has had a rocky relationship with Trump over the years, with the president-elect more recently describing Facebook as an “enemy of the people” in a March interview with CNBC. Meta levied a two-year suspension on Trump’s Facebook and Instagram accounts in 2021 shortly after the company determined that the former president’s actions following the Jan. 6 insurrection in Washington, D.C., could potentially incite more violence.
In 2023, Trump was able to regain access to his Facebook and Instagram accounts, but he also faced some restrictions and potential penalties if he were to violate the company’s community guidelines. Meta eventually removed Trump’s account-related restrictions in July during the lead-up to the 2024 U.S. presidential election.
The company has taken additional steps to appease the incoming administration in recent months. On Monday, Meta announced Dana White, CEO of the Ultimate Fighting Championship and a longtime friend of Trump, is joining its board.
Following Trump’s presidential victory in November, Zuckerberg joined a number of other big technology executives who visited the president-elect at the Mar-a-Lago resort in Palm Beach, Florida, and in December, Meta confirmed a $1 million donation to Trump’s inaugural fund.
DLocal is one of Latin America’s most prominent payment players. It specializes in cross-border payments for emerging markets such as Brazil, Mexico, Colombia and its home country, Uruguay.
Sopa Images | Lightrocket | Getty Images
LONDON — Uruguayan payments firm dLocal has secured a U.K. payment institution license, adding to the company’s growing portfolio of regulatory authorizations as it furthers global expansion.
The emerging markets-focused fintech told CNBC it had acquired an authorized payment institution license from the Financial Conduct Authority, which is Britain’s financial services regulator. That would allow it to start onboarding U.K. merchants for the first time.
DLocal will onboard U.K. merchants through a local entity, Larstal Limited. The subsidiary, which trades in the U.K. as AstroPay, was previously unable to onboard clients locally because of restrictions placed on it by the FCA. DLocal said the restrictions were the result of the U.K.’s exit from the EU.
Pedro Arnt, dLocal’s CEO, told CNBC he expects the business to stand out from domestic payment tech rivals, such as Worldpay and Checkout.com, given its focus on emerging markets in places like Latin America, Africa and Asia.
“The differentiating factor for us when we think of our U.K. base of merchants is that the geographies where we serve them, and those are the only geographies we work,” Arnt said in an interview. He added that dLocal is also targeting global merchants that have a U.K. presence.
“The U.K. has become a hub for many global companies — even the American companies, some Asian companies — for their emerging market expansion, primarily in Africa, and in some cases LatAm,” Arnt told CNBC.
UK expansion plans
Established in 2016, dLocal is one of Latin America’s most prominent payment players. It specializes in cross-border payments for emerging markets such as Brazil, Mexico, Colombia and its home country Uruguay.
With a payment license now under its belt, dLocal is looking to boost its U.K. footprint, with plans to increase headcount and grow business.
Arnt said dLocal has already been expanding its U.K. footprint, with a number of its senior executives — like Chief Operating Officer Carlos Menendez and Chief Revenue Officer John O’Brien — based in London. Globally, dLocal currently has over 1,000 employees.
Arnt said a major benefit the U.K. payment license will bring dLocal is recognition as a “licensed partner” that companies in the developed world can trust to handle payments in emerging markets with complex regulatory needs. DLocal now holds over 30 licenses and registrations worldwide.
Still, dLocal will come up against some fierce competition. Britain already has an established fintech ecosystem with numerous well-capitalized players in the world of payments operating there, including PayPal, Stripe, Adyen, Checkout.com, Mollie and Revolut — to name a few.
‘Not for sale’
DLocal went public on the Nasdaq in 2021, notching a $9 billion valuation at the time. It’s seen its market capitalization decline since then. As of Tuesday, the business was worth $3.4 billion. Still, the stock has risen about 40% in the past six months.
Last month, Reuters reported dLocal was in the process of exploring a potential sale. When asked about buyout speculation by CNBC, Arnt said he didn’t want to comment on rumors, but clarified that dLocal isn’t currently for sale.
All in all, Arnt said, being a public company comes with a level of transparency and oversight that he sees as “positive commercially” for it. At times, he added, “rumors will emerge that someone’s interested in the asset — but I wouldn’t assume there’s too much to that.”
“While there would be a fiduciary duty to shareholders to entertain takeovers, Arnt said that for now, “the company is not for sale.”
Jensen Huang, Nvidia’s founder, president and CEO, speaks about the future of artificial intelligence and its effect on energy consumption and production at the Bipartisan Policy Center in Washington, D.C., on Sept. 27, 2024.
Chip Somodevilla | Getty Images
Nvidia revealed new chips for desktop and laptop PCs on Monday that use the same Blackwell architecture underpinning the company’s fastest AI processors for servers and data centers.
The chips, called GeForce RTX 50-series, will come pre-installed in computers ranging from about $550 to $2,000, the company said. Laptops with the chips will start shipping in March.
Nvidia unveiled the processors at CES in Las Vegas, where CEO Jensen Huang delivered a keynote address on Monday.
“Can you imagine, you have this incredible graphics card, Blackwell, I’m going to shrink it and put it in there,” Huang said, holding up a laptop.
Nvidia, which has soared past $3.5 trillion in market cap by selling AI chips to giant cloud vendors and other tech companies, was until the last few years known for selling graphics processing units (GPUs) to power video games. Nvidia’s first chip in late 1999 was designed to draw triangles and polygons quickly for 3D games.
“Of course, back then, we were a gaming company, and these GPUs were created to accelerate games,” Justin Walker, senior director of product at Nvidia, said on a press call.
Wall Street is less enthused about Nvidia’s gaming business these days given the explosion in AI and ever-increasing demand for more processing power. In the quarter that ended in October, Nvidia’s gaming sales accounted for under 10% of total revenue, compared to 88% from data center chips.
Nvidia has the vast majority of the AI GPU market for data centers, outpacing rivals Advanced Micro Devices and Intel.
But CES, formerly known as the Consumer Electronics Show, is all about consumer products, and the new chips announced Monday are primarily intended for gaming.
Nvidia says the RTX 50-series chips will support a feature called DLSS 4 that uses AI to boost gaming frame rates. They also can display character faces with more details, and will generally provide users with better graphics and higher resolution.
Nvidia’s gaming business is growing, with revenue increasing 15% from a year earlier in the latest quarter. But data center sales have at least doubled for six straight quarters, topping $30 billion in the most recent period.
Nvidia says that technical improvements made for its massive AI business will trickle down into its graphics cards for games.
“While we are now an AI company as well as a gaming company, our gaming side still benefits tremendously from the fact that we are an AI company,” Walker said.
The Blackwell GPU architecture and core design that the 50-series chips use debuted in the company’s AI accelerators, which were announced in March and started shipping later last year. Nvidia said they were designed and optimized to run neural networks used by OpenAI’s ChatGPT and Google’s Gemini.
The new chips for PCs and laptops will come in a number of different configurations. The company says the most expensive and powerful of the chips, the RTX 5090, will be sold individually for $1,999 and is twice as fast as its predecessor, the RTX 4090. Nvidia says it has 92 billion transistors, versus 208 billion transistors on the company’s B200 GPU for servers.
Nvidia says the chips will be optimized to run AI models and do computer graphics, not just run the latest games. The chips will be powerful enough for some game makers to integrate generative AI into their characters in games like “PUBG: Battlegrounds.”
The new processors will also will be powerful enough to to run large language models and image generation models from companies including Meta, Mistral and Stability AI, Nvidia said.