In a significant legal move, Sony Music Entertainment has launched a lawsuit against the artificial intelligence startup Udio, claiming that the company has violated rights related to over 30,000 music recordings. This follows allegations that Udio has extracted music from YouTube without authorization from Sony in order to develop its AI models. Users of the Udio platform can input descriptions to generate music, and Sony argues that this technology can produce sound that closely resembles protected recordings.

In the lawsuit, Sony accused Udio of disregarding copyright laws, stating, “From its inception, Udio has disregarded the rights of copyright holders in the music sector in its quest to establish itself as a leading AI music generation service.” The music giant firmly believes that neither Udio nor any other generative AI enterprise should be permitted to pursue such ambitions at the expense of copyright holders.

The current lawsuit amplifies the scope of Sony's original accusations, claiming infringements involving works by notable artists such as Alicia Keys, Dolly Parton, and Elvis Presley. In a previous legal action in 2024, Sony sued Udio alongside other major labels, including Warner Music Group and Universal Music Group. Unlike Sony, both Warner and Universal have since negotiated agreements with Udio to license their music, giving individual artists a choice to participate.

Amid the ongoing legal disputes, Udio has defended itself, asserting its commitment to uphold its technology and expressing disinterest in replicating content present in its training data. In previous statements, the company highlighted that “Generative AI models, like ours, learn from examples,” comparing the process to how students absorb music and study musical scores.

The rise of AI technology has sparked debate in Hollywood regarding the unauthorized usage of creators' work for AI model training without sufficient compensation. While some entertainment entities, including Disney and Universal, have filed lawsuits against AI companies for copyright violations, others like Lionsgate are exploring partnerships with AI firms, such as Runway, in which they hold equity as part of a larger collaboration.

Meanwhile, both Sony and Universal Music Group are pursuing ongoing legal claims against another AI startup, Suno, indicating that the intersection of artificial intelligence and the music industry remains a contentious battleground.

Sony Music Entertainment initiates fresh legal action against AI company Udio.

A moody dusk scene on the steps of a grand courthouse, its facade dominated by a large illuminated Sony Music Entertainment logo. Storm-laden clouds cast deep shadows across cracked marble steps strewn with torn legal papers stamped “LAWSUIT.” In front, a small, beaten sign reading “Udio” lies shattered, while a lone leather briefcase spills more documents into the gathering darkness.

Sony Music Entertainment files new lawsuit against AI startup Udio - Los Angeles Times

- Sony Music sues Udio for using over 30,000 copyrighted recordings without permission.  
- Udio allegedly used artists' music, including Alicia Keys, to train their AI models.  
- Sony seeks damages up to $150,000 per work infringed from Udio's operations.  

Alicia Keys performs at Staples Center in Los Angeles in 2020. Sony said in its lawsuit that Udio has used Keys’ music to train its AI model.

This is read by an automated voice. Please report any issues or inconsistencies here .

See more from the L.A. Times in Google Search. Set us as preferred

Sony Music Entertainment on Monday filed a new lawsuit against artificial intelligence startup Udio, alleging it infringed on more than 30,000 recordings.

Sony said Udio ripped off music recordings from YouTube without Sony’s permission to train AI models. Users on Udio can type prompts describing the type of song they wish to create. Sony says Udio’s platform is capable of producing audio similar to copyrighted recordings.

Advertisement

“Since the day it launched, Udio has flouted the rights of copyright owners in the music industry as part of a mad dash to become the dominant AI music generation service,” Sony said in its lawsuit. “Neither Udio, nor any other generative AI company, can be allowed to advance toward this goal by trampling the rights of copyright owners.”

The new lawsuit increases the number of copyrighted recordings Sony is accusing Udio of infringing. The company said such works include music by artists such as Alicia Keys, Dolly Parton and Elvis Presley.

Sony in 2024 had sued Udio, alongside other businesses including Warner Music Group and Universal Music Group. Since then, Warner Music Group and Universal Music Group have made deals with Udio to license their music, with artists choosing whether to opt in.

Advertisement

Hollywood Inc. Universal Music Group settles with AI music startup Udio Universal Music Group reached a settlement with AI startup Udio, ending a legal battle in which UMG had accused Udio of taking copyrighted music to train its AI model.

“As observed in the intervening time since Plaintiffs initiated the original lawsuit against Udio, when those who develop such a service steal copyrighted sound recordings, the service’s synthetic musical outputs inevitably saturate the market with machine-generated content that will compete with, cheapen, and ultimately drown out the genuine sound recordings on which the service is built,” Sony said in its lawsuit.

Sony is seeking up to $150,000 per work infringed or the actual amount of damages or Udio’s profits made from the infringement, as well as other damages and costs.

Udio did not immediately return a request for comment.

The company said in response to the 2024 lawsuit that it stands by its technology and that it is “uninterested in reproducing content in our training set.”

Advertisement

“Generative AI models, including our music model, learn from examples,” Udio said at the time. “Just as students listen to music and study scores, our model has ‘listened’ to and learned from a large collection of recorded music.”

As AI technology continues to evolve, many Hollywood creatives have grown concerned about how their work is being used to train AI models without their permission and compensation. Meanwhile, proponents of AI say that it can help entertainment companies save money in areas such as visual effects and test bold ideas.

Some entertainment companies including Disney and Universal have sued AI companies alleging copyright infringement, whereas other companies such as “Hunger Games” studio Lionsgate have partnered with AI startup Runway. Lionsgate has equity in Runway as part of its expanded partnership.

Advertisement

Sony and UMG still have ongoing lawsuits against AI startup Suno.

Sony Music Entertainment files new lawsuit against AI startup Udio

Mecka AI, a startup focused on gathering and analyzing human motion data to enhance humanoid robots and other types of robotics, is on the brink of securing a new funding round led by Sequoia Capital, which could bring the company's valuation to around $500 million, according to sources familiar with the situation.

This upcoming funding round follows a successful raise just three months prior, where Mecka secured $60 million in a round led by Framework Ventures, with involvement from Menlo Ventures, SV Angel, and Kindred Ventures.

While the specific amount of the new funding round has yet to be disclosed, the deal's terms remain tentative and could be subject to change.

Mecka AI has not provided any comments on the matter, and Sequoia has also opted to remain silent.

Founded in 2024, Mecka AI was established by a team of four entrepreneurs, including Canadians Josh Gao and Mogen Cheng, who previously developed a fintech platform for restaurants, along with Jason Chong, who joined the company after Coinbase acquired his crypto exchange. The team's fourth member, Duy Nguyen, who handles operations, is the only non-Canadian.

The founders do not possess traditional robotics backgrounds. However, they identified a significant gap in physical-world data and realized that capturing real-time human interactions was crucial to overcoming the limitations faced by general-purpose robots, such as humanoid models.

The name Mecka is inspired by the term "mecha," a fictional concept of giant humanoid robots controlled by humans. The startup aims to revolutionize robotics by doing for this sector what companies like Scale AI have accomplished for large language models. To achieve this, Mecka compensates individuals for recording themselves performing everyday tasks—such as brewing coffee or repairing vehicles—using body sensors and smartphones.

By early June, Mecka was projecting it would achieve a $100 million annual run rate by the end of 2026, as Gao shared with Fortune during the announcement of their previous funding.

Although Mecka AI has not publicly revealed its clients, the startup’s innovative “egocentric” data collection approach is utilized by numerous robotics firms and AI laboratories, who rely on this real-world data along with other physical data acquisition methods, such as teleoperation, to develop their models.

Other companies in the same space focused on real-world data collection for robot training include XDOF, which was recently reported to be approaching a new funding round at a valuation of $1.2 billion, along with human-data platforms branching out beyond just large language models, such as Scale AI and Micro1.

Mecka AI approaches $500M valuation in Sequoia-led agreement during surge for robot training data.

A sunlit robotics lab flooded with warm morning light: a sleek humanoid training robot sits at a steel table, carefully examining a stack of transparent data canisters labeled “Vision,” “Lidar,” and “Sensor Logs.” Behind it, a large red Sequoia Capital tree logo is etched on a glass partition, and a digital ticker overhead reads “Mecka AI nears $500 M valuation.” The scene feels vibrant and optimistic, capturing the investment rush for robot training data.

Mecka AI nears $500M valuation in Sequoia-led deal amid rush for robot training data - TechCrunch

- Mecka AI is nearing a new funding round led by Sequoia Capital at $500 million.  
- The startup previously raised $60 million three months ago, with multiple venture participants.  
- Mecka collects human motion data to enhance training for humanoid and general-purpose robots.

Mecka AI, a startup that collects and analyzes human motion data to train humanoid robots and other robotics, is nearing a new round led by Sequoia Capital at a valuation of about $500 million, according to two people with knowledge of the deal.

The new financing comes just three months after Mecka announced that it raised $60 million in a round led by Framework Ventures that included participation from Menlo Ventures, SV Angel, and Kindred Ventures.

TechCrunch has not learned the precise size of the new round. The terms of the deal are not final and could still change.

Mecka AI didn’t respond to a request for comment. Sequoia declined to comment.

Mecka AI was co-founded in 2024 by four entrepreneurs, including Canadians Josh Gao and Mogen Cheng, who previously built a restaurant fintech startup, and Jason Chong, who joined Coinbase after it acquired his crypto exchange. Duy Nguyen, the only non-Canadian on the team, focuses on operations at Mecka.

The four co-founders don’t have backgrounds in robotics. But they did recognize that there was a dearth of physical-world data and realized that capturing real-world interactions was the primary bottleneck holding back general-purpose robots, including humanoids.

Mecka, which derives its name from “mecha,” a fictional giant robot controlled by humans, set out to do for robotics what Scale AI, Mercor, Surge, and other human data companies have done for LLMs. The startup pays people to record themselves performing everyday tasks — like making coffee or fixing cars — using body sensors and smartphones.

As of early June, Mecka was projecting that it would end 2026 at an annual run rate of $100 million, Gao told Fortune when the startup announced its previous fundraise.

While Mecka AI hasn’t publicly disclosed its customer list, many robotics companies and AI labs rely on real-world data captured through this “egocentric” approach, alongside other physical data collection methods like teleoperation, to build their models.

Other startups collecting real-world data for robot training include XDOF, which TechCrunch reported last week was nearing a new round at a $1.2 billion valuation, as well as human-data platforms expanding beyond LLMs, such as Scale AI and Micro1.

Mecka AI nears $500M valuation in Sequoia-led deal amid rush for robot training data

In a recent discussion about the practices of Chinese AI laboratories, Y Combinator's CEO, Garry Tan, expressed a desire for minimal regulatory interference. He suggests that American AI labs could also benefit from employing similar distillation methods currently utilized by their Chinese counterparts.

In an interview with CNBC, Tan stated, “I would do nothing,” emphasizing the need for an American distillation framework to allow domestic open-weight AI labs to learn from leading U.S. AI models. This strategy aims to enhance the range of open-weight options available in the U.S., making them distinct from those developed in China.

Distillation refers to the process where a model leverages prompts to understand the workings and reasoning of another model. This technique is widely accepted within the AI community to facilitate the training of new models.

This week, Anthropic released a report claiming that certain Chinese labs are conducting “illicit distillation attacks” by concealing their identities and using fraudulent means to gain access to models without proper authorization. Dario Amodei, CEO of Anthropic, has previously urged U.S. regulators to take action against such practices.

Interestingly, Tan holds a different perspective from many in the industry regarding these practices.

It is important to clarify that Tan does not support unethical behaviors like using stolen credentials for distillation. Instead, he advocates for a more transparent approach. He believes that it is excessive for AI labs to control how users interact with the data obtained from their models. Furthermore, he points out that these proprietary AI systems did not seek permission when they aggregated vast amounts of knowledge and content from various sources, often including copyrighted materials.

“Regulating what customers can do with API calls to closed-weight models seems overly restrictive. The government has a role in ensuring that access to intelligence trained on widely available public data leans towards being a public good rather than being tightly controlled by restrictive terms,” Tan explained to TechCrunch.

Tan, an enthusiastic AI user who has even described himself as experiencing cyber psychosis, advocates for a balance between open-weight and frontier AI labs. He believes it is essential for frontier labs to thrive financially while allowing open-weight models to promote accessibility and freedom.

He highlights a critical concern regarding the future of AI, warning against the potential monopolization of this powerful technology. “The worst-case scenario for AI is a single company dominating the landscape,” he asserted. “If one entity holds a monopoly, backed by extensive capital and top-tier researchers, it would undermine innovation and diversity in the field.”

Y Combinator's Garry Tan calls for US open-weight AI labs to also 'refine' frontier models.

A sunlit, high-tech AI lab interior: Garry Tan stands at a central workbench, holding a glass flask into which glowing streams of intricate neural-net data flow from a larger distillation apparatus. Around him, scientists in white coats examine monitors showing complex models being “distilled” into compact versions. Behind them, a bold Y Combinator logo hangs above a softly waving U.S. flag. Natural daylight floods through large windows, casting warm highlights on the glassware and screens, creating an optimistic, discovery-driven atmosphere.

Y Combinator’s Garry Tan wants US open-weight AI labs to ‘distill’ frontier models, too - TechCrunch

- Y Combinator CEO Garry Tan advocates for no regulatory interference in AI distillation techniques.  
- He supports American labs using distillation on frontier models to enhance open-weight options.  
- Tan warns against monopoly in AI, stressing the need for balance among providers.  

When it comes to Chinese AI labs using distillation techniques to extract knowledge from frontier model makers, Y Combinator CEO Garry Tan is hoping regulators stay out of it. In fact, he thinks U.S. AI labs should perhaps play the same game.

“I would do nothing,” he told CNBC in an interview earlier this week. “We could argue that there should be an American distillation regime.”

He elaborated to TechCrunch that this means he wants smaller, American open-weight AI labs to use the same kind of training techniques on American frontier AI labs, giving the U.S. a more robust set of open-weight options that aren’t Chinese.

Distillation is when a model maker extensively prompts another model in order to learn how it works and reasons. It is commonly, and legitimately, used by AI labs to help train new models.

Anthropic this week released its second report alleging that Chinese labs are engaged in “illicit distillation attacks,” hiding their identities to distill without permission and relying on fraud and stolen credentials to do so. Anthropic CEO Dario Amodei had previously publicly called on U.S. regulators to crack down on distillation.

It’s notable that the commander of Silicon Valley’s prestigious and prolific startup accelerator doesn’t agree.

To be clear, Tan isn’t advocating for American AI labs to use stolen credentials to distill. He wants them to be free to come in the front door. In fact, his argument is twofold. He feels it’s an overreach for AI labs to dictate what their customers can do with the information their models share with them.

He also notes that the proprietary AI labs didn’t ask permission when they vacuumed up as much human knowledge as they could to train their models. They famously ingested plenty of copyrighted material without the permission of those intellectual property holders.

“Controlling what users and customers do with API calls to closed weight models feels constraining, and there’s a role government can play here to normalize the fact that access to intelligence that was trained on broad public access data should itself also be more a form of a public good than something locked away behind restrictive terms of service,” he told TechCrunch when asked why American labs should be free to distill, too.

Tan, who is himself such an avid AI user that he once described himself as having cyber psychosis, wants to see a balance between open-weight AI labs and frontier labs.

“They are at the frontier and driving it forward. We want that to be fundable, and be a great business model ongoing,” he told CNBC. “You want open weight models to give people freedom and access.”

To him, the true AI doomer scenario is for all the immense power of frontier AI to wind up in the hands of a single powerful, proprietary provider. “The nightmare scenario, the doomer scenario for AI is that there’s just one company,” he said. “It has the best access to capital. It has the best AI researchers. It runs away with it and suddenly there’s one company that’s monolithic. And that would be bad.”

Y Combinator’s Garry Tan wants US open-weight AI labs to ‘distill’ frontier models, too

Google is stepping up its game in the artificial intelligence sector by onboarding new talent from Mechanize Inc., a startup based in San Francisco.

Public records indicate that the tech behemoth has successfully finalized a talent acquisition deal with Mechanize. Notably, Tamay Besiroglu, the co-founder and former CEO of Mechanize, has transitioned into a role as a research scientist at Google’s AI division, DeepMind, as highlighted in his LinkedIn profile. Additionally, more than a dozen former employees from Mechanize have joined Google, primarily contributing to the company's midtraining initiatives—crucial components for enhancing its chatbot capabilities.

Neither Google nor Besiroglu has issued any comments regarding this development.

Business Insider previously reported that Google engaged in negotiations for a deal valued at over $1.5 billion for Mechanize's technology and expertise, although the specifics of the agreement were not made public. According to LinkedIn, Guive Assadi, Mechanize's former chief of staff, now holds the title of CEO.

Recently, Google has faced challenges in launching an AI model that effectively competes with others in the market. To gain a competitive edge, the company has opted for talent deals—typically structured to sidestep antitrust concerns—rather than traditional acquisitions. For instance, last year Google acquired talent from the AI coding startup Windsurf, with ex-CEO Varun Mohan now leading Google's agentic coding initiative, Antigravity.

Mechanize's technology is designed to enhance the coding capabilities of AI models, an area where Google has encountered difficulties. Earlier this year, Mechanize announced that it secured $9.1 million in funding at a valuation of $500 million. Prior to founding Mechanize, Besiroglu co-founded Epoch AI, a company dedicated to testing AI models.

Google Finalizes Acquisition of AI Agents Startup Mechanize

A sunlit, modern tech campus courtyard with a large, colorful Google logo mounted on a sleek glass building. In the foreground, a small group of professional engineers wearing subtle “Mechanize” lapel pins shake hands with Google executives at an open lobby entrance. Warm natural light streams in, and faint mechanical gears are softly etched into a nearby glass wall to hint at AI agent integration.

Google has a new edge in the AI race - Business Insider

- Google has completed a talent deal with San Francisco startup Mechanize Inc.  
- Mechanize's co-founder Tamay Besiroglu is now a research scientist at DeepMind.  
- Google aims to enhance its AI models, particularly in coding, through this acquisition.  

Google has some fresh talent in the AI race.

The tech giant has completed its talent deal with San Francisco startup Mechanize Inc, public postings show.

Mechanize's co-founder and ex-CEO, Tamay Besiroglu, now works as a research scientist at Google's AI division, DeepMind, according to his LinkedIn profile. Over a dozen other former Mechanize employees now work at Google, per LinkedIn, mostly on the company's 'midtraining' efforts — a key part of building the best chatbot.

Google and Besiroglu declined to comment.

Google had been in talks for a $1.5 billion-plus deal for Mechanize's tech and talent, Business Insider previously reported. The final terms of the deal weren't disclosed. Mechanize's former chief of staff, Guive Assadi, now describes himself as its CEO, per LinkedIn.

Google has been struggling to release an AI model that's widely viewed as competitive lately. It has used talent deals — which are typically structured to avoid antitrust scrutiny compared to fully fledged acquisitions — to gain an edge. For example, Google snapped up AI coding startup Windsurf's talent last year, with former CEO Varun Mohan now leading Google's agentic coding program Antigravity.

Mechanize's technology helps tech companies' AI models improve at coding, a challenge Google has struggled with.

Mechanize said earlier this year that it raised $9.1 million in a funding round at a $500 million valuation. Besiroglu previously cofounded Epoch AI, which also focused on testing AI models.

Google Completed Its Talent Deal for AI Agents Startup Mechanize

Listen Labs, a startup specializing in voice AI for customer interviews, has recently engaged in negotiations for a $125 million Series C funding round, slated to be led by Menlo Ventures and valuing the company at $1.5 billion, as reported by sources familiar with the discussions. However, this financing round did not reach closure, with Listen Labs opting to withdraw from the signed term sheet—a decision that is somewhat unusual and typically viewed negatively in the venture capital sector.

The potential breakdown of the Series C round is largely attributed to ongoing acquisition discussions with Salesforce. Reports suggest that the CRM leader is exploring the possibility of acquiring Listen Labs for approximately $2 billion. Yet, these negotiations are still in preliminary stages and may not culminate in a finalized agreement.

Operating in the fast-evolving domain of AI-driven customer research, Listen Labs is gaining traction. Founded just three years ago, the company boasts an annual revenue of roughly $30 million—significantly outperforming its competitor Simile, which focuses on predicting human behavior. Simile recently announced the completion of a $200 million Series B funding round at a valuation of $2 billion, potentially raising the valuation expectations for Listen Labs among investors.

Should the acquisition talks with Salesforce fall through, industry insiders anticipate Listen Labs might re-enter the funding market, aiming for a valuation that could exceed $2 billion. While integrating Listen Labs' AI capabilities could enhance Salesforce’s customer needs predictions, it remains uncertain if Salesforce will agree to a steep revenue multiple of 67 times, as suggested by an individual with negotiating experience in similar situations.

Co-founded in 2023 by Florian Jüngermann, a former national champion in competitive programming from Germany, and Alfred Wahlforss, the founder of a staffing startup called Bemlo, Listen Labs emerged from their shared experiences while pursuing master's degrees at Harvard University. 

The innovative technology developed by Listen Labs automates the process of crafting survey questions and conducting customer interviews via audio or video. The insights derived from these interactions are compiled into detailed reports and PowerPoint presentations, akin to those generated by traditional market researchers.

Large corporations, including Microsoft, Canva, Anthropic, and Sweetgreen, utilize such research to assess customer preferences and satisfaction with their products and services. Traditional market research often incurs high costs and can involve lengthy timelines. In contrast, Listen Labs’ technology enables organizations to gather critical feedback quickly and cost-effectively, allowing for agile product iteration based on customer responses.

In addition to Listen Labs and Simile, a number of other startups are also aiming to innovate in the customer research landscape. Competitors such as Outset, Keplar, and Aaru are exploring various methodologies, with some leveraging real human interview processes while others, like Aaru and Simile, use synthetic approaches to simulate human behavior without direct interviews.

Previously valued at $500 million, Listen Labs had raised $69 million in a Series B funding round led by Ribbit Capital back in January, attracting investments from prominent backers like Sequoia, Conviction, and Pear VC.

AI research startup Listen Labs canceled a $1.5B funding round to engage in discussions with Salesforce.

A darkened boardroom bathed in warm tungsten light, a projector beam casting a large Salesforce cloud logo on the far wall. On the polished table below, a pile of shredded documents stamped “Listen Labs $1.5 B Funding Round” spills toward silhouetted executives leaning forward in tense discussion.

AI research startup Listen Labs scrubbed a $1.5B funding round for Salesforce talks - TechCrunch

- Listen Labs signed a $125 million Series C term sheet but later withdrew from it.  
- Talks with Salesforce for a potential $2 billion acquisition have not finalized.  
- The startup develops AI-driven customer research, significantly cutting traditional market research time and costs.  

Listen Labs, a market research startup that uses voice AI to conduct customer interviews, recently signed a term sheet for a $125 million Series C at a $1.5 billion valuation, with Menlo Ventures set to lead the round, according to several people with knowledge of the matter.

But that round never closed, the people said. Listen Labs walked away from the signed term sheet, a rare occurrence in the venture world and one that is generally frowned upon, according to VCs.

The financing likely collapsed because of acquisition talks with Salesforce. The CRM giant has recently held talks to buy Listen Labs for around $2 billion, Business Insider reported. The discussions are not finalized, however, and may not result in a deal, the outlet notes.

Listen Labs is one of the leading startups in the rapidly growing field of automating customer research with AI. The three-year-old startup has about $30 million in annualized revenue, about three times more than Simile, a competing startup that predicts human behavior, according to two people familiar with the companies’ financials. In late July, Simile announced that it had closed a $200 million Series B at a $2 billion valuation led by Greenoaks — likely setting a new valuation benchmark for Listen Labs, one person said.

If talks with Salesforce collapse, several VCs told TechCrunch that they expect Listen Labs to return to market and target a valuation of $2 billion or higher.

While acquiring Listen Labs could strengthen Salesforce’s AI capabilities by using the startup’s AI to help predict customer needs, the CRM giant may ultimately decide that paying a 67 times revenue multiple is too steep a valuation, according to a person with experience negotiating exits to Salesforce.

Listen Labs, Salesforce, Menlo Ventures, and Simile did not immediately respond to requests for comment.

Listen Labs was co-founded in 2023 by Florian Jüngermann, a former German national champion in competitive computer programming, and Alfred Wahlforss, who previously founded a staffing startup called Bemlo. The two met while pursuing master’s degrees at Harvard.

Listen Labs’ AI develops survey questions and interviews customers over audio or video. The resulting conversations are then packaged into reports and PowerPoint presentations, similar to those traditionally produced by human market researchers.

Fortune 500 companies rely on this type of research to gauge customer needs and satisfaction with their brands and products, but traditional market research is expensive and can take weeks to complete.

Listen Labs’ technology helps reduce the time and cost of these projects, enabling companies to quickly understand how customers are reacting to product changes and iterate on them more efficiently.

The startup’s customers include Microsoft, Canva, Anthropic, and Sweetgreen. Listen Labs and Simile aren’t the only startups using AI to disrupt the customer research market.

Besides Simile, competitors in the space include Outset, Keplar, and Aaru. While some platforms automate interviews with real humans, others startups — like Aaru and Simile — take a synthetic approach, using AI to simulate human behavior and predict responses without interviewing anyone at all.

Listen Labs was previously valued at $500 million when it announced a $69 million Series B round in late January led by Ribbit Capital, with participation from returning backers Sequoia, Conviction, and Pear VC.

AI research startup Listen Labs scrubbed a $1.5B funding round for Salesforce talks

Harvey, the innovative legal AI company that’s capturing the attention of venture capitalists, recently disclosed that it has successfully secured an additional $550 million in funding, bringing its valuation to an impressive $15.5 billion, as announced on Wednesday.

This latest funding round was jointly led by Diffusion and Lightspeed Venture Partners and comes shortly after a prior round in March where the company raised $200 million at an $11 billion valuation. Remarkably, this growth follows a December round that valued Harvey at $8 billion, marking a rapid ascent in a brief timespan.

With this recent influx of capital, Harvey's total funding has surpassed $1.55 billion, effectively doubling its valuation in less than nine months. Like other successful tech firms, Harvey has not specified the name of this funding round, though Pitchbook indicates that the company has completed at least eight funding rounds since 2023, with five occurring since 2025. Given that some of these rounds were extension rounds, this could likely be categorized as a Series F.

Notably, this funding follows Harvey's introduction of its first proprietary model, known as Harvey Tenet. This model, developed from the open-weight Kimi K3 model and fine-tuned with legal data in collaboration with inference partner Fireworks (who has also assisted Cursor), represents a significant advancement for the company. Beyond just providing models, Harvey is actively promoting the adoption and post-training of open-weight models by its clients. This positioning showcases how the legal sector can harness AI technologies effectively while diminishing reliance on leading AI developers like OpenAI and Anthropic.

Harvey's valuation climbs to $15.5B, just months after hitting $11B.

A gleaming rocket emblazoned with the Harvey logo soars into a sunrise sky, its exhaust trail glowing “$11B” at the base then morphing to a brilliant “$15.5B” at the tip, bathed in warm natural light.

Harvey hits $15.5B valuation, months after reaching $11B - TechCrunch

- Harvey raised $550 million in funding, achieving a $15.5 billion valuation.  
- The company has now raised over $1.55 billion, nearly doubling its valuation in nine months.  
- Harvey introduced its first in-house model, promoting open-weight model training for legal applications.  

In Brief

Harvey, the legal AI startup that VCs can’t get enough of, has raised another $550 million in funding, this time at a $15.5 billion valuation, the company announced on Wednesday.

The round was co-led by Diffusion and Lightspeed Venture Partners and follows the previous $200 million round at an $11 billion valuation announced in March. And that was just a few months after an $8 billion valuation from a round in December.

With this latest infusion of capital, the company has raised more than $1.55 billion total and nearly doubled its valuation in about nine months. Like Databricks, the company didn’t name the round, though it has completed at least eight priced rounds since 2023 (five of them since 2025), Pitchbook estimates. Since a couple of those were considered extension rounds, this one could be considered a Series F.

The fresh cash comes a couple of weeks after Harvey announced its first in-house model, Harvey Tenet, built from open-weight model Kimi K3 and post-trained with legal data with the help of inference provider Fireworks (who also helped Cursor train its homegrown model). In addition to offering a model, Harvey is encouraging its customers to adopt and post-train their own open-weight models. So Harvey is quickly becoming an example of how an entire industry — law — can both heavily use AI and not rely on proprietary frontier AI labs like OpenAI and Anthropic.

Harvey hits $15.5B valuation, months after reaching $11B

A pioneering chip startup has recently emerged from its quiet development phase after more than seven years of rethinking computer memory architecture. Based in San Jose, California, Kepler Computing is confident that its innovative approach could help alleviate the current global memory chip shortage, given it can scale up production effectively.

Founded in 2018 by a talented team of physicists and computer scientists, Kepler has introduced a novel architecture for high-bandwidth memory (HBM) that tackles some of the supply chain issues plaguing the semiconductor market today. Traditionally, chip manufacturers have depended on costly extreme ultraviolet lithography (EUV) to miniaturize transistors and fit more functionality into smaller spaces. However, Kepler's unique “3D stacking” technique, combined with a proprietary material, enables them to increase memory density without the use of EUV, making it compatible with existing semiconductor manufacturing facilities.

The startup has also achieved significant improvements for high-speed cache memory, which is integral to CPUs, GPUs, and XPUs, further streamlining data transfer within chip designs. While HBM consists of stacked DRAM components, Kepler’s focus on enhancing SRAM allows for quicker access to data.

In terms of financial backing, Kepler has secured $468 million in funding from prominent investors, including GlobalFoundries, Intel Capital, AMD Ventures, the UK-based Baillie Gifford, and the Gates Frontier fund established by Bill Gates. Additionally, in July, the U.S. Department of Commerce pledged up to $245 million to support Kepler’s efforts in developing advanced high-performance AI memory technologies in the United States, leveraging innovative 3D and ferroelectric technologies.

A Discreet Startup Believes It Has Solved the Memory Shortage Problem

Sunlight streams through a dusty workshop window onto a sleek workbench where a hooded figure (face obscured) carefully snaps jewel-like memory modules into a transparent motherboard puzzle. Warm golden filaments bridge empty chip slots, while scattered blueprints and soldering tools hint at a secret startup’s breakthrough hack against the global memory shortage.

A Stealth Startup Thinks It Just Hacked the Memory Shortage - WIRED

- Kepler Computing aims to address global memory-chip shortages with innovative architecture for memory.  
- The startup's 3D stacking approach enhances chip density without costly EUV technology.  
- Kepler has secured $468 million in funding, including support from major tech investors.  

An ambitious chip startup that has spent more than seven years quietly trying to redesign the architecture for computer memory has just come out of stealth mode and believes its new approach can help ease the global memory-chip shortage—provided it can produce its technology at scale.

Kepler Computing, a San Jose, California–based startup founded in 2018 by a team of physicists and computer scientists, says it has developed a new architecture for high-bandwidth memory (HBM) that directly addresses some of the chip supply bottlenecks that are constraining the computing market.

While chipmakers typically rely on expensive extreme ultraviolet lithography (EUV) to shrink the transistors on a chip, thereby packing more technology into the same amount of space, Kepler claims that its “3D stacking” approach and a proprietary new material allow it to increase density without relying on EUV at all—and it can work with existing semiconductor fabrication plants.

Kepler says it has made similar gains for the high-speed cache memory typically used in CPUs, GPUs, and XPUs. This so-called SRAM sits within the core of a chip die in order to cut down on data transfer times. HBM, by contrast, uses stacks of DRAM, which is a separate memory component of chips.

The company has raised $468 million in funding over the past few years from big-name backers. This includes GlobalFoundries, Intel Capital, AMD Ventures, the British investment fund Baillie Gifford, and Bill Gates, through his private Gates Frontier fund. In July, the US Department of Commerce committed to give Kepler up to $245 million to “develop in the US a new class of high-performance AI memory technology, enabled by innovative 3D and ferroelectric technologies.”

A Stealth Startup Thinks It Just Hacked the Memory Shortage

Israeli startup Euno has successfully secured $23 million in a Series A funding round, led by N47, with contributions from existing investor 10D and several notable tech visionaries. This includes Yinon Costica, co-founder of Wiz; Yotam Segev, Cyera's co-founder and CEO; Ofir Ehrlich, Eon’s co-founder and CEO; Rotem Weiss, founder of Tavily; and Mark Nelson, the former CEO and president of Tableau.

With this latest investment, Euno has now raised a total of $29 million. The company is addressing a critical challenge faced by businesses integrating AI—providing agents with sufficient organizational context to discern reliable information, understand its usage, and recognize permissible actions.

Euno is focused on creating a platform that equips AI agents with insights into a company's internal knowledge and data management practices. They describe their innovative technology as an “AI-native context brain” that perpetually learns how organizations create, utilize, and govern their data. This capability enables businesses to deliver the precise context that AI agents need for trustworthy action.

Currently, Euno has a team of nearly 30 employees operating in both Israel and the U.S., including a dedicated research unit specializing in organizational context. The company aims to double its workforce by the end of the year, channeling most of the new investment towards sales, marketing, and AI research.

Founded in 2023 by CEO Sarah Levy and CTO Eyal Firstenberg—both alumni of Talpiot, an elite Israeli military technology program—the company possesses significant technological acumen. Levy holds a master’s degree in physics from the Weizmann Institute of Science, previously serving as CTO at Sight Diagnostics, which focuses on AI-driven blood analysis, and has held senior tech roles in the IDF’s Unit 81. Firstenberg, who was a former section head in Unit 8200, also held the position of VP of R&D at LightCyber before its acquisition by Palo Alto Networks.

Initially, Euno set out to create its platform at a time when companies were just beginning to explore AI applications. As their focus shifted towards the necessity for AI agents to access organizational knowledge—knowledge that human employees typically absorb intuitively—they recognized that organizations aim to harness AI’s potential effectively. As Levy noted, “Every organization wants to be AI. Today the agents are very intelligent, but they struggle to determine which data is trustworthy and relevant when integrated into an organization.”

Euno emphasizes that enterprise data systems were mainly constructed with human users in mind. Traditionally, businesses have relied on data catalogs, manual semantic layers, and governance rules to aid employee comprehension and usage of information. However, AI agents operate across various platforms at an accelerated pace, becoming increasingly autonomous in their functions.

The company's platform intends to supplant much of the traditional human-generated context with a dynamic system that comprehensively understands data flow within organizations, including its meaning and utilization.

Central to Euno's strategy is the research aimed at deriving institutional knowledge from operational indicators, rather than solely relying on manual documentation. They have discovered that the crucial insights needed to identify trustworthy data often stem from patterns within ever-evolving metadata graphs.

Euno leverages these patterns to reconstruct organizational knowledge and utilizes feedback from AI agents’ interaction with this context to enhance their understanding continuously. The aim is to enable companies to transition AI projects from pilot stages to full production in weeks, rather than enduring months or even years to make their data “AI-ready.”

Moreover, the system incorporates governance within its contextual framework, determining the specific information that an individual agent can access based on its tasks and roles while maintaining organizational boundaries during operations.

Levy states, “We build a brain that connects to the models, explaining what they can and cannot do, and where confidential information resides. We implement policies to empower businesses.”

Euno is already collaborating with Fortune 500 companies across various sectors, including market intelligence platform AlphaSense and telecommunications leader Zayo Group.

Founders of Wiz, Cyera, and Eon support Israeli AI startup Euno with $23 million investment round.

Interior of a sunlit modern Tel Aviv startup office: a huge “Euno” logo painted on a white brick wall, an Israeli flag hanging beside it. On a sleek wooden table, three polished golden plaques embossed with “Wiz,” “Cyera,” and “Eon” flank an illuminated sculptural “$23 M.” Warm morning light streams through floor-to-ceiling windows, casting dynamic shadows across minimalist tech sketches and hinting at the infusion of investor capital into Euno’s AI vision.

Wiz, Cyera and Eon founders back Israeli AI startup Euno in $23 million round - calcalistech.com

- Israeli startup Euno raised $23 million to develop AI agents' organizational context understanding.  
- Euno's platform enables AI agents to access relevant data safely and efficiently.  
- The company plans to double its workforce, focusing on sales, marketing, and AI research.

Israeli startup Euno has raised $23 million in a Series A round led by N47, with existing investor 10D and a group of prominent technology founders participating, including Yinon Costica, co-founder of Wiz; Yotam Segev, co-founder and CEO of Cyera; Ofir Ehrlich, co-founder and CEO of Eon; Rotem Weiss, founder of Tavily; and Mark Nelson, former CEO and president of Tableau.

The funding brings Euno’s total raised to $29 million as it seeks to solve a growing problem for enterprises adopting AI, giving agents enough organizational context to know what information to trust, how to use it and what they are allowed to do with it.

1 View gallery Euno founders. ( Zohar Shitrit )

Euno develops a platform designed to help AI agents understand an organization’s internal knowledge and navigate its data. The company describes its technology as an AI-native “context brain” that continuously learns how organizations build, use and govern their data, allowing companies to give agents the specific context they need to act reliably.

Nearly 30 employees work for the company in Israel and the U.S., including a research group focused on organizational context. Euno expects to double its workforce by the end of the year, with most of the new funding going toward sales, marketing and AI research.

The company was founded in 2023 by Sarah Levy (CEO) and Eyal Firstenberg (CTO). Both are graduates of Talpiot, the Israeli military’s elite technology program. Levy holds a master’s degree in physics from the Weizmann Institute of Science and was previously CTO of Sight Diagnostics, which developed AI-powered blood testing technology. She also held senior technology positions in the IDF’s Unit 81.

Firstenberg previously served as a section head in Unit 8200 and later became VP of R&D at LightCyber, which was acquired by Palo Alto Networks.

Euno began developing its platform as companies were only starting to experiment with AI inside their organizations. The company later shifted its focus toward AI agents and the problem of giving them access to the organizational knowledge that human employees often absorb naturally.

“Every organization wants to be AI. Today the agents are very intelligent, but when they are in the organization they don’t know what data to trust, what data is relevant and what data to use,” Levy told Calcalist. “They are stupid when it comes to the organization’s knowledge.”

Euno argues that enterprise data infrastructure was largely designed for humans. Companies have traditionally relied on data catalogs, manually maintained semantic layers and governance rules to help employees understand and use information. AI agents, however, can move across multiple systems and domains, consume and create information at a much faster pace and increasingly operate with a degree of autonomy.

The company’s platform is designed to replace much of that human-maintained context with a continuously updated system that understands how data flows through an organization, what it means and how it is used.

At the core of Euno’s approach is research into whether institutional knowledge can be inferred from operational signals rather than manually documented. The company says it has found that much of the information needed to determine which data should be used and trusted is encoded in patterns across continuously changing metadata graphs.

Euno analyzes those patterns to reconstruct organizational knowledge and then uses feedback from how AI agents interact with the context to refine its understanding. The goal is to allow companies to move AI initiatives from pilot to production within weeks rather than spending months or even a year making their data “AI-ready.”

The system also incorporates governance into the context layer. It is designed to determine what information an individual agent should be able to access based on its task and role, while enforcing organizational boundaries as the agent operates.

“We build a brain that connects to the models and explains to them what they are allowed to do and how. It also tells them where they are not allowed to touch and what information is confidential,” Levy said. “We use policies to enable business.”

The company is already working with Fortune 500 organizations across industries, including market intelligence platform AlphaSense and telecommunications company Zayo Group.

Wiz, Cyera and Eon founders back Israeli AI startup Euno in $23 million round

Sapien, a two-year-old artificial intelligence startup, has successfully secured a fresh round of funding, spearheaded by Ali Partovi from Neo, and is now valued at $180 million. The company has evolved from simply offering financial planning software to developing a more extensive platform that analyzes the impact of operational decisions on a business’s financial performance. Notable clients include Bayer, Carlex, Cooper Standard, Blink Charging, and Westgate Resorts.

The changing nature of Sapien's platform use is highlighted by a recent case study on automotive supplier Carlex. In this case, Sapien enhanced an existing profitability analysis, revealing that factors previously believed to contribute positively to the company’s EBITDA by $10 million were actually leading to a $2 million decline. Additionally, the platform uncovered a previously unnoticed opportunity worth $1.5 million through patterns in customer channels, all accomplished in approximately 20 minutes.

Jason Waltz, the VP of finance for Carlex's Aftermarket Division, remarked on the efficiency of this analysis, noting that a process that could have taken his team two weeks was completed in a fraction of the time with Sapien’s technology.

Founded in 2024 by Ron Nachum, Pranav Ravella, and Arya Grayeli, Sapien initially centered on financial planning and analysis. However, the company's vision has expanded to include a focus on understanding how operational choices influence financial results. Nachum describes Sapien as an integrated system that links financial data from a company’s statements with the underlying business activities, enabling finance and operations teams to delve into the factors affecting revenue, margins, and cash flow.

This new approach goes beyond merely providing dashboards or automating reports; Sapien aims to leverage AI to perform the investigative tasks that finance and operations teams typically manage. Carlex has since broadened its use of the platform to areas such as pricing, inventory management, customer orders, and supply chain operations. Other firms like Cooper Standard and Blink Charging have also reported similar efficiencies, with analyses being streamlined from hours or days down to mere minutes. 

Nachum acknowledges that a significant hurdle for AI adoption in financial sectors is not the capability of the technology itself, but rather building trust in its outputs. He notes that, while large companies generate substantial data, they often struggle with effective analysis. Establishing trust in the results is essential, and once finance teams feel secure using the platform for their analysis, its use can expand to other functional areas such as sales and accounting.

This evolution signifies Sapien's ambition to connect financial outcomes with the operational decisions that create them. Nachum is committed to steering clear of creating an "Excel copilot" and instead aspires to design a system that can analyze business operations, identify valuable trends, and assist teams in transforming these insights into replicable processes.

The overarching vision for Sapien is to harness AI not merely for accelerating existing financial workflows but to discover operational strategies that can dramatically enhance a company's financial performance.

In leadership changes, Jennifer Fulk has recently been appointed CFO of Kura Oncology, effective September 8. Fulk comes with over 15 years of experience from Eli Lilly and has held various senior financial roles in different organizations. Meanwhile, Mira Mircheva has opted to step down as EVP and CFO of Bally's Corporation effective September 4, with George Papanier stepping in as interim CFO while the search for a permanent replacement begins.

In employment updates, U.S. employers added 162,000 jobs in August, a significant rise compared to an average of 31,000 monthly over the previous year, while the unemployment rate remained steady at 4.1%. Job growth was particularly strong in the restaurant and bar sector, which contributed 59,000 new jobs, while local government education added 42,000 positions. However, the information sector saw a decline, losing 23,000 jobs, primarily due to ongoing cuts in tech-related industries. Wage growth remained steady at 3.1% year-over-year, with average hourly earnings reaching $37.75, and the Bureau of Labor Statistics revised job numbers for June and July upward by 55,000.

As the market anticipates a potential Federal Reserve rate hike, current employment trends may influence these expectations.

AI startup Sapien secures funding at a $180M valuation to assist businesses in identifying the true drivers of profitability.

A bright, sunlit modern boardroom where executives in smart business attire study a large transparent touchscreen display filled with rising profit graphs and glowing data nodes. A sleek golden neural-network sphere labeled “Insight” floats at the center, with the bold Sapien logo prominently embossed on the screen. Through floor-to-ceiling windows, a vibrant cityscape basks in morning light, hinting at new growth.

Exclusive: AI startup Sapien raises at $180M valuation to help companies find what’s really driving profit - Fortune

- Sapien raised funding at a $180 million valuation, expanding beyond financial planning software.  
- Analysis previously taking weeks is now done in minutes, revealing significant financial opportunities.  
- The company emphasizes building trust in AI for finance to enhance operational decision-making.  

Good morning. Sapien has raised a new funding round led by Neo’s Ali Partovi at a $180 million valuation, as the two-year-old AI startup pushes beyond financial planning software into a broader system for analyzing how operational decisions affect a company’s bottom line. Its customers now include Bayer, Carlex, Cooper Standard, Blink Charging and Westgate Resorts.

The shift is reflected in how customers are using the platform. According to a forthcoming case study I was able to review on automotive supplier Carlex, Sapien rebuilt an existing profitability analysis and found that factors the company had identified as contributing $10 million in positive EBITDA were actually producing a $2 million drag.

Sapien later found another $1.5 million opportunity in a customer-channel pattern the team had not been looking for. The analysis took about 20 minutes.

“It took 20 minutes,” said Jason Waltz, business unit VP of finance for Carlex’s Aftermarket Division. “It would have probably taken us two weeks, and we probably wouldn’t have gotten to that level.”

That kind of analysis is increasingly the focus for Sapien, which founders Ron Nachum, Pranav Ravella, and Arya Grayeli launched in 2024 with an $8.7 million seed round led by General Catalyst.

The company has also grown quickly, with headcount increasing fivefold over the past year, bringing together AI researchers and engineers from Meta, Google, and Palantir with executives and operators from McKinsey & Company, Blackstone, Barclays, and Plaid.

Nachum, Ravella, and Grayeli met in high school before attending Harvard, Stanford, and the University of Texas at Austin, respectively.

From FP&A to operating decisions

When I first spoke with Nachum, who serves as CEO, two years ago, Sapien was focused on FP&A. The company has since moved toward a broader question: not just what happened to a company’s financial results, but what operational decisions caused them.

Nachum describes Sapien as a financial and operational system that connects the numbers in a company’s financial statements with the underlying business. The idea is to help finance and operating teams investigate changes in revenue, margins and cash flow, identify the drivers behind them and then act on those findings.

That is a different proposition from adding another dashboard or automating another report. Sapien is betting that AI can do more of the investigative work traditionally handled by finance and operating teams.

Carlex has since expanded its use of Sapien into pricing, inventory, customer orders, OEM quoting, supply chain and operations, Nachum said.

Cooper Standard and Blink Charging are using the platform in similar ways, according to the case studies, with analyses that previously took hours or days being reduced to minutes and teams using the results to identify potential savings and margin improvements.

The trust factor

For Nachum, the biggest obstacle to AI adoption in finance isn’t necessarily whether the technology can produce an answer.

“Every large company is data rich and analysis poor,” he said. The harder part, he argues, is getting finance teams to trust the answer—whether the underlying numbers are correct, the analysis is reliable and the company’s data remains secure.

That makes trust an important part of Sapien’s approach. Once teams become comfortable using the platform for financial analysis, Nachum said, its use can spread into functions such as supply chain, sales, and accounting.

The expansion also reflects Sapien’s broader shift away from being an FP&A tool. Rather than treating finance as a standalone function, the company is trying to connect financial outcomes to the operational decisions that produce them.

One of Nachum’s core bets has been to avoid turning the company into what he calls an “Excel copilot.” The goal, instead, is to build a system that can investigate a business, find important patterns and eventually help teams turn those findings into repeatable processes.

For Sapien, that is ultimately the bigger opportunity: using AI not simply to make existing financial work faster, but to find the operational decisions that can materially change a company’s financial performance.

Sheryl Estrada

Sheryl.Estrada@fortune.com

Leaderboard

Jennifer Fulk was appointed CFO of Kura Oncology, Inc. (Nasdaq: KURA), a biopharmaceutical company, effective Sept. 8. For more than 15 years at Eli Lilly and Company, she held increasingly senior roles spanning commercial finance, research and development, investor relations, treasury, and global operations. Her positions included CFO of Lilly's U.S. Bio-Medicines business. Fulk has also served in a series of senior finance and operating leadership roles, including as CFO of Talkspace, chief operating officer and CFO of 120Water, and, most recently, CFO of Soleno Therapeutics.

Mira Mircheva has resigned from her role as EVP and CFO at Bally's Corporation, effective Sept. 4, due to personal reasons, according to an SEC filing. Her last day will be Sept. 30. George Papanier, Bally's president, will serve as interim CFO. The company is searching for a permanent CFO.

Big Deal

U.S. employers added 162,000 jobs in August, the Bureau of Labor Statistics reported, beating the average monthly pace of just 31,000 over the prior year, while unemployment held steady at 4.1%.

Restaurants and bars led hiring with 59,000 new jobs, and local government education added 42,000, largely reversing a dip the month before. Manufacturing extended its climb, up 16,000 and now 58,000 above its December low.

The information sector was the exception, shedding 23,000 jobs as tech infrastructure, data processing, publishing, and broadcasting firms kept cutting. Wage growth held at 3.1% year-over-year, with average hourly earnings reaching $37.75, and the BLS revised June and July's job counts up by a combined 55,000.

Going deeper

"As Wall Street shifts expectations towards a Fed rate hike, the White House turns up the pressure on Warsh’s central bank" is a Fortune article by Eleanor Pringle.

Surprisingly healthy employment data has tipped expectations for a rate hike at the Federal Open Market Committee’s meeting higher this week, with interest rate traders now placing the likelihood at 58.4%, Pringle writes. Read more here.

Overheard

"In a world of infinite information, trust is the scarcest resource."

—Nirav Tolia, CEO, president, chair, and co-founder of Nextdoor, writes in a Fortune opinion piece.

AI startup Sapien raises at $180M valuation to help companies find what’s really driving profit

In the dynamic landscape of Silicon Valley, there’s little that excites investors more than an impressive growth in annual recurring revenue (ARR). However, amid the ongoing AI surge, venture capitalists are starting to raise questions about the true meaning of this figure.

Greg Isenberg, CEO of Late Checkout, sparked discussion with a candid remark that went viral, revealing concerns about the misinformation surrounding ARR from VC-backed AI firms. He noted, "The amount of VC-backed AI companies lying about their ARR publicly is absolutely unsettling," highlighting a sentiment that has been commonly speculated but seldom articulated in tech discussions.

Originally popularized during the software boom, ARR was meant to provide a snapshot of a company's expected yearly earnings from customers who pay regularly, serving as a quick reference for investors evaluating a startup's worth. However, insiders who spoke to Business Insider suggest that the term has become increasingly ambiguous, conflating various forms of revenue such as subscribed services, future contracts, hardware sales, and even one-off revenue spikes.

In essence, true recurring revenue is often misrepresented. Shruti Gandhi, a general partner at Array Ventures, referred to the trend as "ARR inflation," expressing that there currently exists a lack of accountability. She has pointed out instances of founders inflating their figures, with the common rebuttal being that such behavior is ubiquitous.

Alexander Niehenke, a partner at Scale Venture Partners, notes that this trend indicates a psychology reminiscent of the peak of the 2021 market, where risk tolerance among investors was elevated. "It feels like we're at the tail end of 2021 again, given some of the behavior I'm seeing in the venture ecosystem," he remarked.

Founders are well aware of the enthusiasm VCs have for ARR. Niehenke compares their tactics to "catnip," exploiting investor excitement for their own advantage. This behavior was exemplified when Roy Lee, CEO of Cluely, admitted to inaccurately reporting his company’s ARR to a TechCrunch journalist—a rare moment of honesty in an otherwise unregulated environment for startups, where revenue claims often go unchecked.

In the wake of these revelations, more startups are turning to run rate as a metric, which extrapolates a single month’s revenue to project annual earnings without implying that the revenue is consistent. This approach emerged as the nature of AI revenue became less predictable. Unlike traditional software-as-a-service (SaaS), where predictable contracts ensure stable ARR, many AI companies base their pricing on variable usage—commonly calculated in tokens—which leads to significant fluctuations in both revenue and costs.

Pocket, known for its AI recording device, recently announced it surpassed a $100 million run rate. Founder Akshay Narisetti prefers this metric since it reflects the unpredictable nature of token-based revenues. He emphasizes avoiding the pitfalls of annualizing revenue spikes from singular high-performing months.

Despite the issues surrounding run rate, it offers a clearer alternative during these uncertain times. For example, major AI entities like OpenAI refer to their subscription sales as ARR but label their new advertising revenue as an annualized run rate. Meanwhile, Anthropic has generally categorized its sales under the run-rate banner.

In contrast, Linear—a unique SaaS startup—recently affirmed its achievement of $100 million in ARR based on the stability of its revenue model, which includes multiyear contracts. Cofounder and CEO Karri Saarinen defended the use of ARR by pointing to Linear's longevity and reliable forecasting capabilities, though he acknowledged that this is not the case for many newer AI-focused startups, which often find themselves in an increasingly foggy revenue landscape.

VCs Scrutinize AI Firms' High Revenue Assertions

Dark boardroom scene under a single tungsten spotlight: a circle of sharply dressed venture capitalists lean over a sleek glass table strewn with printed charts and spreadsheets. One holds a magnifying glass over a giant, fractured revenue graph glowing in warm amber—its bars bulging then cracking. Behind them, a shadowy silhouette of a generic “AI” emblem flickers on the wall. Tense postures, furrowed brows and long shadows convey deep skepticism.

'There's no accountability;' AI startups keep claiming huge revenue. VCs say the math is getting 'murky.' - businessinsider.com

- Venture capitalists are increasingly skeptical of AI companies' claims about annual recurring revenue.  
- Many startups exaggerate revenue metrics, leading to what is termed "ARR inflation."  
- AI's unpredictable revenue model makes traditional ARR less reliable than in software companies.

Few metrics thrill Silicon Valley investors more than surging annual recurring revenue (ARR). In the AI boom, however, VCs are increasingly skeptical of what that number actually represents.

"The amount of VC-backed AI companies lying about their ARR publicly is absolutely unsettling," Greg Isenberg, CEO of Late Checkout, wrote in a post that went viral for saying something out loud that has been whispered about, but rarely said publicly in tech circles.

ARR was a metric that came into vogue during the software era, intended to estimate the yearly sales a company can expect from paying customers, a quick way for investors to judge a startup's valuation.

Now, venture capitalists and founders who spoke with Business Insider said ARR has become so loosely used that it can blur the line between contracted subscriptions, future deals, token consumption, hardware sales, and a single strong month of revenue.

In short, recurring revenue is often not recurring.

"It's ARR inflation," said Shruti Gandhi, a general partner at Array Ventures. "There's no accountability right now."

Gandhi says she has called out founders for exaggerating revenue. Their defense: "Everybody else is doing it."

To Alexander Niehenke, a partner at Scale Venture Partners, the current trend of padding ARR reflects a top-of-the-bubble mentality when founders have all the leverage.

"It feels like we're at the tail end of 2021 again, given some of the behavior I'm seeing in the venture ecosystem," he said. "Anytime you get deeper into a bull cycle, everybody has more tolerance for risk."

Explore BI Games Take a smarter break in your day - and see how far you get. Play now Want more Business Insider in your news feed? Add BI in Google so our reporting is easier to find when you’re searching for what matters. Add to Preferred Sources

Founders know how excited VCs get about ARR, according to Niehenke. "The smart and brilliant entrepreneurs use it like catnip on us," he said.

Cluely CEO Roy Lee raised eyebrows earlier this year when he admitted to misstating his startup's ARR to a TechCrunch reporter. But that was very much the exception. With startups under much less regulatory scrutiny than public companies, revenue claims are rarely verified.

it’s genuinely crazy tech has so isolated itself from regulatory forces that you can just admit to securities violations on the internet — Will Manidis (@WillManidis) March 5, 2026

More startups are using run rate, but is it better?

Before AI, Silicon Valley was in love with software-as-a-service companies, where ARR was much more dependable. SaaS companies like Salesforce or Workday often sell annual or multiyear deals, making it fairly straightforward to tally the value of subscriptions likely to recur.

"When you have ARR, you have a signed contract, and you know what you've signed up for for the year," said Matt Murphy, a partner at Menlo Ventures.

AI has proven less predictable. Many AI companies no longer sell a simple per-seat license that costs the same every month. They charge customers based on how much they use, often measured in tokens. Someone may use a product aggressively, then pull back, or switch to a cheaper model. The startup's own costs vary widely, too, because it may have to pay a model provider or cloud company every time a customer uses its product.

AI startups have increasingly shifted to describing run rate, which extrapolates a month of revenue over a full year and does not purport to be recurring, according to Murphy.

"Venture does this from time to time, which is that there's a new metric that shows up," he said.

Pocket, which makes an AI recording device, is one of those companies, having recently said it crossed the $100 million run rate threshold. Founder Akshay Narisetti said he avoids ARR because tokens make revenue too unpredictable to expect it to be recurring.

Run rate can come with its own problem in that a gangbusters month does not necessarily translate to a year, but Narisetti said he avoids doing that.

"We never annualize one large month with a spike," Narisetti said.

The biggest AI labs have used both metrics. OpenAI has described subscription sales as ARR while calling its newer advertising business an annualized revenue run rate. Anthropic has generally described its overall sales as run-rate revenue.

Linear, which is an increasingly rare breed of SaaS startups, recently said it crossed $100 million in ARR.

Cofounder and CEO Karri Saarinen said he stands by the metric because Linear sells multiyear contracts and has been in business since 2019.

"Since we've been operating longer, we can better forecast the actual ARR," he said, adding that for newer AI startups, it is a different story.

"It's gotten really murky," he said.

Why VCs Are Questioning AI Companies' Big Revenue Claims

According to a report from Bloomberg News on Monday, Anthropic has chosen to abandon its plans to acquire the artificial intelligence startup known as Decart AI. This decision comes from sources who are knowledgeable about the situation.

Anthropic abandons efforts to acquire $6 billion-valued AI startup Decart

A dimly lit corporate boardroom at night, overhead tungsten spotlight casting long shadows over an empty negotiation table. On the table lies a torn contract stamped with the Decart logo. On the far wall, a large illuminated Anthropic logo flickers and dims as if powering down. The scene is moody and atmospheric, dark grays and warm amber highlights emphasizing a deal collapsed.

Anthropic drops pursuit of AI startup Decart - The Economic Times

- Anthropic has opted not to acquire AI startup Decart AI, reports indicate.  
- Sources reveal the decision follows internal evaluations of the potential acquisition.  
- The move reflects Anthropic's strategic focus and current investment priorities.

Synopsis

Anthropic âhas decided not âto pursue an acquisition of artificial âintelligence startup Decart AI, Bloomberg News reported on Monday, citing people familiar with the matter.

Sony Music Entertainment initiates fresh legal action against AI company Udio.

More from AI Startups & Investments

Mecka AI approaches $500M valuation in Sequoia-led agreement during surge for robot training data.

Y Combinator's Garry Tan calls for US open-weight AI labs to also 'refine' frontier models.

Google Finalizes Acquisition of AI Agents Startup Mechanize

AI research startup Listen Labs canceled a $1.5B funding round to engage in discussions with Salesforce.