According to a recent Bloomberg report, Stripe has successfully completed its acquisition of OpenRouter.

OpenRouter specializes in helping clients select from a variety of AI models tailored to their unique requirements and financial constraints. Earlier this year, the company announced a substantial $113 million in Series B funding, achieving a valuation of approximately $1.3 billion. Notable investors in this round include Sequoia, Andreessen Horowitz, Menlo Ventures, and Alphabet's Capital G.

CEO Alex Atallah previously likened OpenRouter to a "Stripe for AI," emphasizing its role in providing a centralized platform for accessing diverse AI systems, thus avoiding vendor lock-in. The startup proudly reported having 8 million users worldwide and offering access to over 400 innovative AI models.

Last month, The Wall Street Journal revealed that Stripe and OpenRouter were engaged in acquisition negotiations. Bloomberg now states that these discussions have culminated in a deal valued at over $7 billion.

In response to inquiries about the acquisition, a spokesperson for Stripe informed TechCrunch that the company does not comment on rumors or speculation.

Stripe is set to purchase AI gateway company OpenRouter for over $7 billion, according to reports.

A modern data center flooded with warm natural daylight: at its center, a giant white vault door swings open bearing the Stripe logo. From inside, glowing golden fiber-optic streams branded “$7B+” flow outward and seamlessly merge into a sleek row of rack-mounted routers labeled “OpenRouter,” symbolizing the high-value takeover.

Stripe will reportedly acquire AI gateway startup OpenRouter for $7B+ - TechCrunch

- Stripe has finalized an acquisition deal for OpenRouter, valued over $7 billion.  
- OpenRouter offers access to various AI models based on customer needs and budgets.  
- The startup has 8 million users and more than 400 models available.  

In Brief

Stripe has finalized a deal to acquire OpenRouter, according to a new report in Bloomberg.

OpenRouter helps customers to select different AI models to perform different tasks, depending on their specific needs and budget. The company announced in May that it had raised a $113 million Series B, at a reported $1.3 billion valuation. (Investors include Sequoia, Andreessen Horowitz, Menlo Ventures, and Alphabet’s Capital G.)

At the time, OpenRouter CEO Alex Atallah described the company as the equivalent of Stripe for AI, because it provides customers with a single access point for different systems and prevents lock-in. The startup also claimed to have 8 million global users and to provide access to more than 400 models.

The Wall Street Journal reported last month that Stripe and OpenRouter were in acquisition talks. Now, Bloomberg said those discussions have led to a deal price of more than $7 billion.

A Stripe spokesperson told TechCrunch that the company does not comment on rumors or speculation.

Stripe will reportedly acquire AI gateway startup OpenRouter for $7B+

In a move to broaden its educational offerings, Harvard Business School is incorporating artificial intelligence avatars to deliver personalized feedback to its students.

These innovative avatars have been developed by a startup known as HeyGen and are integrated into the HBS Foundry bootcamp, an eight-week program priced at $699, designed specifically for aspiring entrepreneurs. Although participants enjoy live weekly sessions with instructors, it is the AI avatars that step in to provide evaluations during practice pitches and board meetings.

New York Times reporter Sarah Kessler experienced this firsthand, pitching an AI-driven recreation of Jeff Bussgang, co-founder of Flybridge Capital. While both Bussgang and his digital counterpart were not particularly taken with her idea of creating “Uber for bananas,” Kessler noted that the avatar maintained a eerily static smile throughout her presentation.

Katharina Rings, the project director, initially thought of the AI feature as a simple chatbot. However, following feedback from a trial run, students expressed a desire for a more interactive experience.

Interestingly, while many college students have voiced skepticism about artificial intelligence, those in the Foundry program indicated their appreciation for the avatars. Bussgang himself remarked that while his digital doppelgänger may come across as somewhat “creepy,” he noted, “My students love it.”

Harvard's $699 startup bootcamp features AI avatars of its teachers.

A sunlit, airy Harvard startup bootcamp studio: adult entrepreneurs sit at sleek desks with open laptops, taking notes as lifelike AI avatars of instructors appear on slim desk-mounted screens, gesturing over digital whiteboards. A large Harvard crest hangs on the bright front wall, warm natural light flooding the room to evoke innovation and tradition united.

Harvard’s $699 startup bootcamp offers AI avatars of its instructors - TechCrunch

- Harvard Business School uses AI avatars for feedback in its entrepreneurial bootcamp.  
- Participants found the AI avatars more engaging than traditional chatbot interactions.  
- Despite mixed feelings, students appreciate the virtual avatars' role in practice sessions.  

In Brief

As Harvard Business School seeks to expand its reach, it’s leaning on AI avatars to provide individual feedback.

These avatars were created by a startup called HeyGen and are included in the eight-week, $699 HBS Foundry bootcamp for entrepreneurs. The program offers live sessions with instructors every week, but the AI avatars are the ones providing feedback during practice pitches and board meetings.

New York Times reporter Sarah Kessler actually tried this out herself by pitching an AI-generated copy of Flybridge Capital co-founder Jeff Bussgang. Apparently, both the real Bussgang and his simulacra were unimpressed by her plan to build “Uber for bananas,” but Kessler said the virtual version offered a noticeably frozen smile during her pitch.

Project director Katharina Rings said she initially envisioned the AI component as something closer to a chatbot. However, after HBS released a trial version, students said they wanted a more guided experience.

And while some college students haven’t been shy about expressing their negative feelings towards AI, Foundry participants told Kessler they like the avatars. As for Bussgang, he acknowledged his digital copy is a little “creepy,” but he said, “My students love it.”

Harvard’s $699 startup bootcamp offers AI avatars of its instructors

Michael Polansky is remarkably unassuming for someone operating in a corner of the world known for outsize egos.

Seated at a leafy patio outside a popular bakery in Mill Valley, an affluent town about 15 miles north of San Francisco, Polansky — bespectacled, his fresh face framed by dark hair shot through with gray — has the look and the friendly demeanor of a young professor.

He is, in fact, both the founder of a buzzy AI and biology startup called Outer Biosciences and the creative, business, and romantic partner to Stefani Germanotta — better known as Lady Gaga. It’s an unlikely double life. On the one hand, the couple invariably moves in the world that comes with celebrity; on the other, he’s running an outfit that has spent years figuring out how to keep living human tissue alive outside the body — for over a month, so far — without anyone outside the company knowing.

He didn’t see any of it coming. Polansky grew up in Minnesota and went on to Harvard, where he studied applied mathematics and computer science, graduating in 2006. Afterward, he logged three years at the hedge fund Bridgewater Associates — “a very unique place,” Polansky says over coffee, and one where he had “a really good experience,” even if it wasn’t a place he was going to “wake up and be excited about . . .every day.”

His pilgrimage from Bridgewater to Silicon Valley ran through Minnesota. By coincidence, Sean Parker’s assistant at the time had been Polansky’s neighbor growing up. At a 2007 wedding in their home state, she mentioned that Parker was looking for someone to work with him. Polansky already knew who Parker was and was looking to move west. The two had dinner in New York, “hit it off immediately,” and, as Polansky tells it, he quit Bridgewater the next day and moved to San Francisco.

He first landed a role as a principal at Founders Fund when the firm was run by its four original partners — Peter Thiel, Sean Parker, Luke Nosek and Ken Howery – with Polansky and another young principal at the time, Brian Singerman, sharing an office. “It was a really, really great experience,” Polansky says.

When Parker left Founders Fund after becoming liquid in his Facebook stock and wanted to build out his own family office, he brought Polansky with him. Polansky ran that office — handling Parker’s business, investment, and philanthropic interests, including helping stand up the Parker Institute for Cancer Immunotherapy (where Polansky remains executive director) until COVID, when his life “moved in a different direction.”

That change had a lot to do with Germanotta. Polansky met her in late 2019 at one of Parker’s birthday parties. Charmingly, the meeting came at the urging of her mother, Cynthia Germanotta, president of the Born This Way Foundation, whom Polansky had come to know through his philanthropic work.

“She had said, you know, for months and months, ‘I want to set you up with my daughter,’” Polansky recalls. “I was like, I think you’re making fun of me.” She wasn’t. When their own mothers later met, he says, laughing, “it all made perfect sense.” (His mother and Germanotta’s mother are now close friends.)

Their relationship is a full partnership, professionally and personally. Notably, for Gaga’s most recent world tour, which began in July of last year and ended in April, the couple managed a massive operation across three 747s, something Polansky compares to running “a 200-person startup that travels around the world every day.”

It also includes Haus Labs, the cosmetics brand Germanotta initially built “on her kitchen floor,” Polansky says, instead of simply licensing her name to an existing company. That business, based in El Segundo, California, an operation with roughly 70 employees, is reportedly thriving.

Germanotta also sits on the board of Outer Biosciences, and the two companies collaborate at the margins. For example, Outer Biosciences’ chief scientist, Kyung-Jin Jang, sits on Haus Labs’ scientific advisory board, and the companies have run some joint projects.

Says Polansky, beaming as he talks about her from our sun-dappled table: “People really haven’t gotten to know a certain side of her publicly … She’s such a brilliant businessperson.”

From cancer to tissue in a dish

While Polansky is also a businessperson, he’s not a scientist. He got into life sciences “accidentally,” he says, through more than a decade spent alongside Parker in cancer immunotherapy, a field that was “very fringe” when they entered it.

Outer Biosciences, which Polansky founded in 2022 and where he is CEO, grew out of frustration that the pace of innovation in biology and chemistry has never matched software, in large part because there’s no ethical way to run experiments directly on people. Meanwhile, the proxies scientists rely on instead — animal models, simplified cell cultures, lab-grown organoids — are poor stand-ins for how a real human organ behaves.

So Outer Biosciences took a different approach. Instead of engineering a synthetic organ, the company sources human skin that would otherwise be discarded after surgery – mostly plastic surgery – through what it describes as vetted non-profit and commercial biobanks and brokers operating under “institutional review board oversight and documented donor consent,” principally the National Disease Research Interchange and the Cooperative Human Tissue Network. (Both outfits receive federal funding from the NIH and the National Cancer Institute without being federally operated.)

Polansky is careful to note that there is no single government tissue network that qualifies buyers. He says Outer Biosciences pays fees to these suppliers on a cost-recovery basis rather than purchasing tissue outright. He also says the company spent roughly two years on building that pipeline and handling the protocols required to receive that tissue “within hours” of surgery, while it’s still living.

Asked about the donors’ privacy, he says that every sample arrives already de-identified — stripped upstream by the supplying organizations of names, contact information and other direct identifiers. A proprietary support system developed by his team then feeds the tissue nutrients and removes metabolic waste, extending its viable life well past the industry norm.

That norm, by the way, is a matter of days. That’s enough time to test for acute toxicity, but not for slower biological processes like collagen remodeling, pigmentation change, or barrier repair that take weeks to unfold. (Dermatologists routinely tell patients to expect changes over a matter of weeks for this same reason.)

Outer Biosciences’ system meanwhile keeps tissue alive for up to a month, says Polansky, who says it retains its “day-zero architecture and preserves its day-zero epidermal, stromal and immune-associated molecular programs.” In plainer English, that means that 30-day-old tissue cared for by the company looks a lot like day-one tissue, but it isn’t fully indistinguishable from it.

Sunburn is one of the clearer examples of what that extra time buys Outer Biosciences. Polansky says that its researchers can induce UVB damage in living tissue, then track the stress, inflammatory, and recovery-related responses that follow over the following weeks as an information-gathering exercise. He says the team isn’t “healing” the skin but rather watching an injury happen and then watching the biology that follows it over time.

Perhaps anticipating pushback from the scientific community, Polansky is careful about how he frames the company’s achievements when this reporter asks questions about rival technologies. The startup’s value, he says, isn’t any one piece of what it’s doing but how the pieces fit together: human tissue that can be kept alive for weeks, a diverse donor pool that Outer Biosciences’ team can subject to controlled experimental conditions, and repeated molecular measurements taken along the way.

And it’s all fed into one, closed self-enforcing system. An AI model predicts which untested chemicals are likely to have a beneficial effect on a specific skin function. Those chemicals get run through the living-tissue system. Then the results, whether the prediction was right or wrong, get fed back into the model, improving the next round of guesses.

It’s a giant improvement from where things started, says Polansky. Early on, the company relied on a “brute force” approach, mining scientific literature and partnering with the National Cancer Institute on natural compounds from extreme environments. That phase produced a couple of leads over about 18 months, but with AI layered in, the company is now generating a new candidate roughly every six weeks, with six leads currently active in its pipeline and several dozen additional “hits” logged.

What makes that pace truly astonishing, Polansky says as the crowd around us thins out, is the size of the current universe of skin-active ingredients.

It’s almost impossible to know the exact number, but it’s small. “Active ingredient” means something different informally than it does formally. While the FDA maintains rules covering 13 categories of over-the-counter skin drugs (think sunscreen, antifungals), across all of them, only about 120 to 130 active ingredients are approved. Add cosmetic ingredients backed by actual research, says Polansky, and that number is closer to 200.

This, of course, presents opportunities.

Outer Biosciences is discovering cosmetic ingredients, not drugs, so there’s no FDA approval to seek out. Instead, the route runs through two steps: first, getting the ingredient a standardized industry name; then safety testing under guidelines set by the OECD, a Paris-based international body whose member countries agree to accept each other’s properly run studies — meaning a study done correctly in one participating country is accepted in the other 40-plus.

A partner outfit then commercializes the whole thing. Indeed, rather than build its own consumer brand, Outer Biosciences right now plans to license or sell its finished ingredients to beauty or pharma outfits that will then formulate these into actual products (a serum, a cream) and bring them to market under their own brands. Already, four of Outer Biosciences’ six current leads look likely to reach commercialization, Polansky says.

In the meantime, the company is generating money from collaborative research partnerships, including a pharmaceutical partner that’s studying why certain cancer drugs cause severe skin rashes, and consumer beauty brands that are testing whether Outer Biosciences’ data holds up against their own product-development and marketing needs.

If Polansky is raising more money for the company currently, he won’t say. To date, the company has raised roughly $23 million, with early backers including Wing Ventures, Initialized, and Polansky’s own investing firm, Hawktail, among others. The company employs 19 people, with all but Polansky based just outside Cambridge, Massachusetts.

Asked why he’s chosen now to talk about the company after years of near-total silence — he says he barely discussed it even with close friends — Polansky points to the data the team is beginning to amass and the confidence that has given them. “Trying to do this in private is hard,” he says, as the wait staff start flipping chairs onto tabletops, signaling it’s closing time at the bakery. “We kind of want to start working in public now,” he adds, shrugging.

Outer Biosciences is not the only company chasing this idea, even if its use of real tissue, rather than synthetic, is distinctive.

Vivodyne — a Philadelphia-based competitor building lab-grown human organ tissue paired with predictive AI to replace animal drug testing — just this month announced it has raised close to $80 million to date, including a $38 million seed round and a $40 million Series A, both led by Khosla Ventures.

Other rivals are pursuing various flavors of organ-on-a-chip and microphysiological systems for preclinical testing.

Polansky doesn’t seem especially preoccupied with any of them — less, it seems, out of arrogance than because he seems to have his hands full. Besides, there’s plenty of room for everyone in this moment. Unlike AI companies training on scraping the internet, there is no “biology internet” to scrape.

And Outer Biosciences has two other reasons to focus on its own knitting. First, the data it generates doesn’t exist anywhere else, which, conceivably, makes the company’s position more defensible, if a lot slower, to build than “traditional” software-based AI startups. It’s also cheaper to run, with modest compute demands compared with training a large language model. In fact, all of the company’s AI work currently runs on-premise, not in the cloud, because “we don’t want the data in the cloud,” Polansky says.

Whether over time, Outer Biosciences becomes a standalone commercial-ingredients business, licenses its discoveries, or eventually reorganizes around a single breakout compound, Polansky says he hasn’t settled on – and the team doesn’t need to. The more important goal, he says, is a predictive model that’s accurate enough that the company can spot promising directions in skin biology without having to run every experiment physically first, opening up a rate of discovery in dermatology that doesn’t currently exist.

For now, instead, the work of turning a promising compound into a real product — the formulation, the manufacturing scale-up, the supply chain, the safety testing — is still done manually by the same scientists who discover the compounds in the first place. Building out a product-development team, with people who’ve done this kind of work before, is next on the roadmap.

“I think it’s going to be fun,” he says, “to have people know that this is what we’ve been doing.”

Michael Polansky is training an AI model on skin that’s still alive

Generate a cinematic news thumbnail for title:Michael Polansky is training an AI model on skin that’s still alive. No text in image.

Michael Polansky is training an AI model on skin that’s still alive - TechCrunch

Rillet co-founder and CEO Nicolas Kopp seems justifiably confident as we talk over Zoom a day after his company announced a $100 million raise at a $1 billion valuation. The U.S. has a shortage of accountants right now, which is driving growth of his AI-native accounting platform so much, he raised that cash in 48 hours without even trying.

Rillet emerged from stealth two years ago. Since then, it has raised $200 million from top investors like Iconiq, Andreessen Horowitz, and Sequoia. It’s also amassed 600 customers, most of whom are looking to ditch legacy accounting systems like Oracle and NetSuite, Kopp says.

A few weeks ago, Rillet held a board meeting and shared with investors its growth since its $70 million Series B last summer. Annualized revenue rate had doubled in the last quarter alone; the startup added new clients, many of them public companies, and an alliance with EY to introduce AI tools to the auditing giant.

His customers aren’t piloting Rillet either, he said — they’re yanking out ERP and accounting software from competitors like Intuit, NetSuite, or Oracle.

After that board meeting, text messages were fired, calls were made, and 48 hours later, Rillet was a unicorn. The company wasn’t even looking to raise, Kopp said.

Seth Pierrepont, the general partner at Iconiq who led the round, said that the deal came together fast but “it wasn’t a cold start,” he described.



“Rillet had already proven it could win against the incumbents that have owned this category for decades,” Pierrepont told TechCrunch. Iconiq also co-led the company’s Series B, and with this latest round, Pierrepont joins the Rillet board. “A year of watching the team deliver on that made doubling down and leading the Series C an easy call.”

Julien Bek, Sequoia’s lead investor on the deal, also said that, though 48 hours might look rushed from the outside, from their perspective, re-investing in Rillet was a “very easy decision,” after the company’s growth in the past year.

“Rillet’s initial wedge is accounting, but ultimately they are reinventing the entire finance function,” Bek told TechCrunch, adding that agentic finance could become “one of the largest application software opportunities of the AI era.” Sequoia led Rillet’s Series A last summer.

“When the opportunity came together,” Bek continued. “We already had all the context we needed.”

Rillet is one of many AI-native startups now giving legacy players a run for relevance. Earlier this year, software stocks on the public market dipped as investors worried about how emerging AI tools would affect them. Kopp thinks there’s some truth to that.

“AI is going to come hard at these legacy players,” he said, because it is giving customers compelling alternatives.

Rillet, for example, was built for AI agents, not humans, letting humans work alongside the AI agents on corporate bookkeeping. Rillet clients range from laundromats to a major sports franchise. Some 50% of Rillet customers come from Intuit, 30% from NetSuite and Sage Intacct, and 20% from Oracle, SAP, Workday, and Microsoft products, he said.

Security is critical when working with sensitive client data, Kopp said. Rillet includes model routing, so customers can redirect requests to the foundational model of their choice (like OpenAI or Anthropic), and Rillet’s harness prevents these models from training on their data, he says.

Rillet product imagery Image Credits:Rillet

There’s also no cross-training — meaning one customer’s data remains proprietary. The agents also have memory, so they can remember and store historical actions they can then use for their own process and improvement.

About three months ago, Rillet released a governance feature letting accountants see and audit every decision the AI agent has made — including what numbers the agents pull and how they calculated them. Creating this was harder than it looks, Kopp said, because the team had to compress agent data into a format humans could understand.

Kopp said this feature was only possible to build recently because AI agents have gotten so powerful so quickly. They can, for example, now support multi-step workflows over longer periods of time. Because of that, auditing what they are doing has become even more important for clients.

“We barely scratched the surface of potential and opportunity that this technology has,” he said.

Right now, regulations for public companies require that every transaction made by an AI agent be approved by another human. He thinks regulators and top names are watching how the accounting industry evolves around this new technology. He’s hopeful that new rules and regulations will evolve that align more with where everything is headed.

“It’s a very normal process,” he said. “Similar to when the cloud came, of just getting everybody familiar with what’s going on and how it helps the profession.”

Kopp also doesn’t think mass job displacement from AI is coming anytime soon, especially in accounting. (Stanford released a report a few weeks ago that found no widespread job displacement yet.) He insists that Rillet isn’t a human replacement, not even for junior accountants. They can use Rillet to help automate and assist with some of the profession’s grunt work.

He also pointed out the expected shortage of accountants in the U.S. The number of those graduating with an accounting degree has been declining since at least 2010. In a recent report, the Controllers Council Organization found that 61% of finance leaders have struggled to find finance, accounting, and CPA talent in the past year. The pullback is not entirely shocking: accountants’ hours are long, the pipeline to the top is arduous, the pay often doesn’t match the workload, and the work doesn’t appeal to everyone.

At the same time, the Bureau of Labor Statistics has projected that accounting-related needs are expected to grow by at least 5%, adding 72,800 jobs by 2034. It also doesn’t expect AI to reduce the demand for accountants, even as the technology becomes more widespread. “The automation of routine tasks, such as data entry, will instead make accountants’ advisory and analytical duties more prominent,” the BLS said.

“I just don’t see people losing their job anytime soon,” Kopp said. “These people have started their professions to help businesses make better financial decisions,” he added. “We can fully enable them to do that.”

This piece was updated.

How AI accounting startup Rillet raised $100M and became a unicorn in 48 hours

Generate a cinematic news thumbnail for title:How AI accounting startup Rillet raised $100M and became a unicorn in 48 hours. No text in image.

How AI accounting startup Rillet raised $100M and became a unicorn in 48 hours - TechCrunch

The near-bottomless demand for unique AI training data from top labs and corporations is driving a massive boom for a cohort of data-labeling startups.

One of these fast-growing businesses is Micro1, a four-year-old startup that expanded its gross annual run rate from $100 million to $500 million over the past eight months, according to a person familiar with the company. Like its peers that hire domain experts such as doctors, lawyers, and scientists on a contract basis, Micro1 retains roughly 60% to 70% of that figure, putting its net annual run rate between $150 million and $200 million.

While Micro1 still lags competitors like Mercor (which hit $2 billion in gross annualized revenue this summer) and Handshake (which reached $1 billion earlier this year), the startup’s revenue growth shows that there is more than enough demand to support multiple players supplying AI training data.

The rapid growth is bound to continue, with some researchers hypothesizing that future AI spending on data could rival spending on compute.

That outlook bodes well for Micro1, which is seeing its contract sizes grow at an accelerated pace and expects its margins to expand over time. The startup is increasingly generating synthetic data without human involvement, such as by creating automated descriptions of video content. Additionally, some of the data it generates can be sold to multiple customers, driving gross margins for this “off-the-shelf” data as high as 80% to 90%, a person familiar with the startup’s finances told TechCrunch.

Selling the same datasets to multiple clients has sparked recent controversy, with critics arguing that distributing off-the-shelf data to Chinese AI developers helps make their models as powerful as top U.S. models.

Micro1’s founder, Ali Ansari, said last month on X that unlike some of its competitors, the startup doesn’t sell its data to Chinese model makers. “Some human data companies work with foreign adversaries. [A]nd the results show today in Kimi K3. We believe it’s shameful to claim American AI dominance desires while selling millions worth of data to countries that we are in adversarial competition with.”

Like Mercor, Micro1 began as an AI recruiting startup. But after noticing that data-labeling clients were using his AI platform to vet and recruit engineers for annotation, Ansari decided to pivot and enter the data-labeling business, too.

Ansari previously told TechCrunch that in addition to having its experts evaluate model outputs — a concept known as reinforcement learning gyms — the company is building a robotics pre-training dataset by having hundreds of generalists record everyday object interactions in their homes.

Micro1 raised its Series A at a $500 million valuation last September, and TechCrunch understands that the startup may have recently raised another round at a significantly higher valuation.

Micro1 didn’t respond to a request for comment.

AI data startup Micro1 reaches $500M gross run rate amid AI training boom

Generate a cinematic news thumbnail for title:AI data startup Micro1 reaches $500M gross run rate amid AI training boom. No text in image.

AI data startup Micro1 reaches $500M gross run rate amid AI training boom - TechCrunch

A new player in the AI landscape is vying for access to Spirit Airlines' data. Ali Ansari, the 25-year-old founder and CEO of Micro1, a startup focused on enhancing AI model training, revealed to Business Insider that his company has proposed a significantly higher bid for Spirit's data, following Google’s successful $10 million offer.

On Wednesday, Micro1 formally submitted a $12.5 million bid to Spirit Airlines' legal representatives, as confirmed by a source familiar with the development. As part of its bankruptcy proceedings after filing last year, Spirit Airlines is liquidating its assets, which includes valuable corporate data.

Experts in bankruptcy proceedings stated that while it's atypical for courts to entertain late submissions such as Micro1's, it is not entirely outside the realm of possibility, especially if the new offer can provide greater financial returns for creditors.

Ansari emphasized the importance of company data in training AI systems, noting its potential for enhancing model performance with more realistic training inputs. He characterized Google's offer as "actually quite low" considering Spirit Airlines' extensive operational history, adding that Micro1 compensates other data providers up to $2 million for their information.

Operating out of San Francisco, Micro1 employs gig workers to refine AI models for its clientele and has created digital environments where AI agents can master tasks, such as booking travel tickets. Founded in 2022, the startup boasts around 150 employees and has sought valuation offers of up to $2.5 billion, as noted by Forbes last December.

The competition for Spirit Airlines' data isn't limited to Micro1; another AI training company, Mercor, previously entered the fray, resulting in a bidding contest that saw Google’s initial bid of $5 million escalate to $10 million. Mercor ultimately placed a final bid of $7.5 million but fell short against Google's offer.

Neither Google, Spirit Airlines, nor Mercor have commented on Micro1's recent bid. The creditors’ committee for Spirit Airlines also did not respond to inquiries regarding the matter.

The potential for Micro1 to disrupt Google's agreement remains uncertain. Legal analysts suggest that bankruptcy judges sometimes consider later bids if they promise greater returns for creditors but stress that maintaining the auction process's integrity is crucial.

"A well-publicized and organized auction is typically not overturned," stated Nancy Rapoport, a bankruptcy law professor at the University of Nevada, Las Vegas, adding that the initial higher bidder was given the chance to participate in the auction. Meanwhile, Lindsey Simon, an associate law professor at Emory University, mentioned that the final decision often depends on the presiding judge.

As of now, the federal bankruptcy court in New York has yet to ratify the auction results. A hearing regarding the data sale has been scheduled for September 9, and Ansari indicated that Micro1 is preparing to provide comprehensive details about its offer to the court.

AI startup Micro1 plans to contest Google's successful bid for Spirit Airlines data by submitting a higher proposal.

A sunlit corporate auction suite with floor-to-ceiling windows overlooking a Spirit Airlines jet on the tarmac, its logo bold on the tail. On a polished mahogany table sit two oversized illuminated bid paddles: one bearing Google’s logo, now dimmed, and beside it a brighter paddle stamped with the Micro1 logo—capturing the tense David-versus-Goliath moment of a higher offer.

AI startup Micro1 wants to challenge Google's winning bid for Spirit Airlines data with a higher offer - Business Insider

- Micro1 offered $12.5 million for Spirit Airlines' data, surpassing Google's $10 million bid.  
- Spirit Airlines is selling data as part of its bankruptcy proceedings from last year.  
- Bankruptcy court may consider late bids if they benefit creditors or shareholders significantly.  

Another AI company wants to get its hands on Spirit Airlines' data.

Ali Ansari, 25, the founder and CEO of Micro1, a startup that helps train AI models, told Business Insider that his firm is making a "materially higher" offer for Spirit Airlines' data after Google successfully bid $10 million for it this week.

Micro1 sent a $12.5 million offer to Spirit Airlines' legal team on Wednesday, a person familiar with the matter said.

Spirit Airlines is selling off its assets in bankruptcy court — including troves of corporate data — after going under last year.

Bankruptcy experts told Business Insider it would be unusual — but not out of the question — for the court to consider Micro1's offer after it missed the deadline to bid in the original auction.

Related video Microsoft is betting big on AI. Company insiders have serious doubts.

Ansari said Micro1's interest reflects the value of company data for training AI models, as it contains the kind of messy inputs that help AI models learn to operate in real-world environments.

"Realistic data is very valuable," he said, calling Google's winning bid "actually quite low" given Spirit Airlines' decades of operation. Micro1 pays other companies up to $2 million for their data.

Based in San Francisco, Micro1 pays gig workers to improve AI models for its clients, and runs digital 'environments' where AI agents can get better at tasks like booking train tickets. Founded in 2022, it has about 150 employees and has fielded offers at a $2.5 billion valuation, Forbes reported in December.

Micro1 CEO Ali Ansari. Micro1

AI training startups are scrambling to acquire data from companies to supercharge their clients' chatbots. Mercor, another AI training startup, also bid for the Spirit Airlines data, sparking a bidding war with Google, which initially bit $5 million. Mercor's final bid of $7.5 million lost to Google's $10 million.

Google, Spirit Airlines, and Mercor declined to comment on Micro1's last-minute move. A representative for Spirit Airlines' creditors committee didn't respond to requests for comment.

Whether Micro1 can actually disrupt Google's deal isn't clear.

Bankruptcy law professors told Business Insider that judges sometimes entertain late bids if they could generate significantly more money for creditors and shareholders, or offer other terms that those parties find more favorable.

At the same time, courts also want to preserve the integrity of the original auction process.

"A duly noticed, well-run auction generally won't get undone," Nancy Rapoport, a law professor at the University of Nevada, Las Vegas, said. "The higher bidder had the opportunity to bid during the auction."

Lindsey Simon, an associate law professor at Emory University, said it often comes down to the judge overseeing the case.

"The bankruptcy code isn't clear on this," she said.

The federal bankruptcy court in New York has not yet approved the auction results.

On Tuesday, it delayed a hearing on the auction after a group of former Spirit Airlines' flight attendants objected over concerns about what would happen to their data.

Google previously told Business Insider that any data it receives will be "rigorously scrubbed" of any personally identifiable information by a third party.

A hearing for the data sale will be held on September 9. Ansari said Micro1 is working on submitting details about its offer to the court.

AI startup Micro1 wants to challenge Google's winning bid for Spirit Airlines data with a higher offer

Andreessen Horowitz (a16z) has garnered attention for its American dynamism thesis, but holding a U.S. passport isn't essential for entrepreneurial success. Gabriel Vasquez, a partner specializing in AI applications and the firm's global investment strategy, highlighted that “44% of our investments in the Apps Fund One and Two feature founders from outside the U.S.”

Alongside general partner Angela Strange, Vasquez spearheads a16z’s Borderless Founder network, an initiative aimed at empowering immigrant and international entrepreneurs. The duo asserts that exceptional companies can emerge from anywhere, but they’ve come to believe that foreign founders may hold a competitive edge over their American counterparts.

Historically, the U.S. has been a magnet for international innovators, but a16z recognizes that the current landscape, particularly in the AI sector, favors those with connections in both their home countries and Silicon Valley. “There is now an advantage to having one foot in your home country and one foot in Silicon Valley,” Vasquez and Strange noted in a pre-release post shared with TechCrunch.

In a conversation with TechCrunch, Vasquez detailed how this emerging advantage has prompted a16z to expend “more than one million air miles” pursuing international investment opportunities, rather than simply waiting for teams to relocate to the U.S.

This shift didn’t happen in isolation. Especially for enterprise startups, previous patterns of international buyer engagement showed limited activity and a reluctance to spend. Vasquez pointed out that “buyers from countries outside the U.S. were not moving rapidly, and their willingness to pay was very low,” highlighting how the landscape has evolved.

Over the past three to five years, many startups have begun to secure substantial clients globally, even in markets that are typically less recognized for tech advancements.

When a16z began receiving interest from seed-stage startups collaborating with Fortune 500 companies across the globe, Vasquez initially suspected it might be an anomaly. However, he quickly realized it was a broader trend. Numerous international AI startups are now landing significant clients early in their development—an occurrence that was rare just a few years back.

In Europe, large corporations previously engaged with startups primarily through open innovation programs or accelerators, seldom committing significant funds. This made converting these startups into paying customers a considerable challenge, prompting France to launch initiatives like “I Choose French Tech.” Yet, according to Vasquez, AI has catalyzed a substantial shift in this paradigm.

AI technology has led established companies to understand the necessity of incorporating third-party solutions to stay competitive—even in regions like Latin America, where labor costs used to deter software development. “Software never really took off in the region due to the competition from cheap labor,” Vasquez explained. However, with the rise of AI agents that are more accurate and perpetually accessible, they have found a role in customer service and other fields.

The demand for AI solutions at the enterprise level is rapidly increasing worldwide, yet Vasquez noted that many American startups currently lack the agility to meet these needs from the outset, which often results in them prioritizing domestic clients instead.

This gap presents an opportunity for local founders to develop AI solutions for businesses in their home countries, while maintaining connections to the U.S. “The notion of headquarters is evolving; it’s becoming more flexible,” Vasquez remarked, observing that even governmental views are shifting in this direction.

A relevant example is Poland, which recently invested in a16z portfolio company ElevenLabs. Although the voice AI startup only maintains a subsidiary in Poland, it boasts clients like InPost and LOT Polish Airlines, and is co-founded by Polish entrepreneurs Piotr Dąbkowski and Mateusz Staniszewski—indicative of the strong presence of Polish talent in the AI sector, which is also notable at OpenAI.

Vasquez believes that access to talent serves as a crucial differentiator in today's market. “Recruiting in Silicon Valley is incredibly challenging since you’re up against companies like Anthropic and OpenAI, which have secured massive funding for top talent.” In contrast, international founders frequently tap into deep talent pools located outside the U.S.

These talent hubs are why Vasquez finds himself frequently traveling, often to Stockholm, to scout for the next potential unicorn. Some of Europe’s leading AI scaleups have emerged from academic institutions, catching the attention of venture capitalists. In this context, it’s worth noting the observation from Dealroom founder Yoram Wijngaarde: “European talent remains one of America’s biggest startup advantages.”

AI provides an edge to foreign founders, according to a16z.

A bright, sunlit startup incubator where a diverse group of adult founders lean over a glass table, collaborating on sleek laptops displaying AI-generated designs. Warm morning light streams through floor-to-ceiling windows. A large a16z logo banner hangs on the wall, and behind them a digital world map on a screen glows with network lines tracing global connections.

For a16z, AI gives foreign founders an advantage - TechCrunch

- A16z's Borderless Founder network supports international founders, reflecting a shift in investment focus.  
- The rise of AI has enhanced global startup engagement with major companies outside the U.S.  
- Recruiting talent locally is increasingly vital as competition in Silicon Valley intensifies.  

Andreessen Horowitz has made headlines for its American dynamism thesis, but a U.S. passport is no requirement. According to Gabriel Vasquez, a partner focusing on AI apps and the firm’s global investment strategy, “44% of our investments in the Apps Fund One and Two have an international founder.”

Together with general partner Angela Strange, Vasquez is the driving force behind a16z’s Borderless Founder network, an initiative to support immigrant and international founders. Their premise is that great companies could be born anywhere, but they now believe foreign entrepreneurs actually have an edge over their American peers.

The U.S. has always attracted foreign founders; but in the current AI-dominated cycle, a16z acknowledges that their roots are more helpful than they used to be. “There is now an advantage to having one foot in your home country, and one foot in Silicon Valley,” Strange and Vasquez wrote in a post to which TechCrunch had early access.

Talking to TechCrunch exclusively, Vasquez broke down how this advantage has become significant enough for a16z to “spend more than one million air miles” in order to go after international dealflow — rather than expecting each and every team to move to the U.S.

This expectation didn’t come out of nowhere, especially for enterprise startups, where there’s been a marked shift in buying patterns from those based outside the U.S. “The buyers from countries outside the U.S. were not moving rapidly, and their willingness to pay was very low,” Vasquez said of the old pattern.

But this has changed dramatically “in the last three to five years,” with startups now able to secure major companies worldwide as customers, even in less cutting-edge markets.

When a16z started hearing from international seed-stage startups working with Fortune 500 companies, “we thought this might be an exception, but it was clearly a trend,” Vasquez said. Examples now abound of international AI startups signing major clients early on — but Vasquez is also right that it was rarely the case even a few years ago.

In Europe, in particular, it used to be that corporations would only engage with startups through open innovation programs and accelerators, but rarely with their main checkbook. Turning them into paid customers was hard enough that France launched a dedicated initiative, “I Choose French Tech.” But according to Vasquez, AI has been the driver in making this shift a reality.

That’s because AI made legacy players realize they would have to buy third-party solutions to remain competitive — even in areas where human labor is still affordable, like Latin America, Vasquez said. “Software never really picked up in the region, because you were competing with cheap labor.” But now that AI agents are “always on” and increasingly accurate, they have become commonplace in customer service and more.

Others could go after these opportunities, too. “There’s so much appetite at the enterprise level for companies all around the world to consume AI, but American [startups] don’t have the speed yet to go serve the entire market from day zero, so they obviously prioritize U.S. companies,” Vasquez said.

This leaves the door open for a local founder to sell an AI solution to companies in their country, while splitting their time between home and the U.S. “The concept of headquarters is something that is changing a lot; it’s just more fluid,” Vasquez said, noting that even governments seem to take a similar view.

Case in point: Poland recently took a stake in a16z portfolio company ElevenLabs. The voice AI startup only has a subsidiary in the country, and not its headquarters, but its customers include InPost and LOT Polish Airlines, and its founders are Piotr Dąbkowski and Mateusz Staniszewski — two examples of Polish talent in AI, also strongly present at OpenAI.

From Vasquez’s point of view, access to talent is a key differentiator. “Right now, recruiting in Silicon Valley is the hardest thing, because you’re competing with Anthropic and OpenAI that have raised billions and billions of dollars for the best talent.” But international founders are often tapping into the talent pool that exists in many places outside of the U.S.

Talent clusters are also the reason why Vasquez has been spending a lot of time on a plane — often in Stockholm — in search of the next unicorn. Some of the top AI scaleups are European academic spinouts, and VCs are going after them. In a sense, the story goes the other way around; to quote Dealroom founder Yoram Wijngaarde: “European talent remains one of America’s biggest startup advantages.”

For a16z, AI gives foreign founders an advantage

Parting ways with Google after nearly three decades was undoubtedly an emotional experience for Jeff Dean. During a recent talk at Stanford University, the renowned computer scientist, who has now embarked on establishing his AI startup, Discovery Loop, appeared visibly moved as he shared his decision to leave the tech giant. Released in a video on Tuesday, his remarks captured a poignant moment when he expressed pride in Google's endeavors towards enhancing the Gemini models while visibly struggling with his emotions.

Dean’s appearance on August 7 was his first public engagement since departing Google, where he garnered respect for spearheading numerous key projects. In his address, he emphasized that working in a smaller startup environment—with just four members at Discovery Loop—presents new avenues for expedited innovation. He stated that advancements in cloud computing enable emerging companies to leverage infrastructure from giants like Google, thus eliminating the need for them to invest extensively in building their own systems.

He remarked, "That allows a small group to secure substantial funding and efficiently utilize these platforms," highlighting the advantages of modern technology for startups. Dean also noted that a compact team enables them to minimize distractions typically encountered in larger organizations, fostering a more focused approach to scientific and engineering advancements.

The landscape for startup companies is evolving rapidly, especially as AI technologies increasingly automate operations that previously required teams of employees. Dean's Discovery Loop is positioned to innovate in this space by streamlining the procedure through which scientists and engineers formulate, execute, and assess experiments.

During his talk, Dean acknowledged that he and his cofounders might opt for decisions driven by a vision for societal benefits, rather than strictly for financial gain. He elaborated on their mission to create systems addressing some of the 14 grand challenges identified by leading global engineers and scientists, which aim to enhance life as we know it. This initiative includes diverse goals such as understanding brain functions and preventing nuclear terrorism.

On August 5, a press release detailed that Discovery Loop's initial funding is spearheaded by Radical Ventures and Khosla Ventures, with contributions from Lightspeed, Kleiner Perkins, and Doerr Capital. Notably, Alphabet is also a founding investor and cloud partner, as confirmed by Google’s CEO, Sundar Pichai. Reports from Business Insider last week indicated that Dean is negotiating to secure $1 billion at a valuation of approximately $10 billion.

Despite the exhilarating potential of launching a new venture, Dean candidly described the accompanying challenges. "It’s a little nerve-wracking," he admitted, balancing his anxiety with the exhilarating thrill that comes with embarking on this new journey.

Jeff Dean explains his departure from Google after 27 years to create a small AI startup.

A lone engineer in casual attire strides away from a sleek glass Google headquarters bearing a large, colorful Google logo. He holds rolled-up blueprints stamped “AI Startup” in one hand, while across a sunlit plaza a modest workshop glows with a simple “AI Startup” sign above its door. Warm, natural morning light casts long, hopeful shadows, hinting at a bold new beginning.

Why Jeff Dean says he left Google after 27 years to build a small AI startup - Business Insider

- Jeff Dean left Google after 27 years to start AI company Discovery Loop.  
- His startup aims to automate scientific processes and tackle global engineering challenges.  
- Dean emphasizes the advantages of small teams in accelerating innovation and focus.  

After 27 years, it must have been hard for Jeff Dean to let go.

Dean, who quit Google to build his own AI startup called Discovery Loop, briefly choked up while describing his decision to leave the search giant after more than a quarter-century during a talk at Stanford University earlier this month.

"They have a plan for making the Gemini models awesome," Dean says in a video of the event released on Tuesday, flashing a thumbs up while appearing to choke back tears. The talk, held on August 7, was his first public appearance since leaving Google.

Though Dean built a sterling reputation at Google, leading numerous high-profile projects, he said during the talk that being part of a small startup — four people in Discovery Loop's case — was now not only possible but an optimized vehicle for accelerated innovation.

Pointing toward the growing availability of cloud compute, Dean said that smaller startups can now "rely on" Google and other cloud providers to "build out the infrastructure" necessary to power AI work.

Related video What are the real-life consequences of AI?

"That makes it possible for a small set of people to raise a fair bit of capital and be able to use those platforms without having to build them themselves," he said.

Working on a smaller team, Dean said, will also allow his startup to cut through "slight distractions" that would ordinarily hamper progress in large organizations.

"We're super excited as a small company about working in a really focused way on science and engineering automation," he said.

More small companies have been popping up recently as AI agents prove capable of automating processes that once required the work of multiple employees. Modern-day startups are scaling more quickly, and thriving with only a handful of employees.

Dean's Discovery Loop, a public benefit company, aims to automate the process by which scientists and engineers propose an experiment, implement and run it, and evaluate the results.

Dean said at the talk that he and his cofounders "might make decisions that are not in the company's financial interest but are in the broader societal good."

Specifically, Discovery Loop wants to build systems that can tackle parts of the 14 challenges laid out by some of the world's most accomplished scientists and engineers to improve life as we know it. Formerly known as the "National Academy of Engineering Grand Challenges," the 14 topics range from "Reverse Engineer the Brain" to "Prevent Nuclear Terror."

A Discovery Loop press release from August 5 said that the company's initial funding round is being led by Radical Ventures and Khosla Ventures, with participation from Lightspeed, Kleiner Perkins, and Doerr Capital. Alphabet is a founding investor and cloud partner, according to Google CEO Sundar Pichai .

Business Insider reported last week that Dean has been in talks to raise $1 billion at about a $10 billion valuation.

Nevertheless, Dean said creating a new startup still comes with its own anxieties.

"It's a little nerve-wracking," he said. "But at the same time it's also exciting."

Why Jeff Dean says he left Google after 27 years to build a small AI startup

In a recent conversation with Bloomberg, Fei-Fei Li, a renowned Stanford professor celebrated as the "Godmother of AI" and the creator of ImageNet, shared insights about her role in the AI startup world, revealing a more nurturing side often associated with being a "tiger mom." This side has emerged through her experiences as the co-founder of World Labs, a startup she established in 2024 that focuses on developing "world models" to enhance AI's ability to comprehend and navigate 3D environments. 

World Labs recently gained significant attention by securing $1 billion in funding from well-known investors such as AMD, Autodesk, and Nvidia. Li highlighted her position within the company, noting that she is the most senior member, surrounded by talented, yet younger co-founders, engineers, and researchers. 

"I genuinely feel like a parental figure here," she remarked. "There are moments when I adopt the 'tiger mom' approach—running this operation with high standards." This metaphor encapsulates her commitment to excellence, drawing from a cultural reference popularized by Amy Chua in her 2011 book, "Battle Hymn of the Tiger Mother," which described a strict parenting style focused on discipline and achievement.

Throughout her journey, Li has embraced the role of a mentor, often sharing anecdotes about AI's profound influence on students and emerging talent. In a podcast from 2025, she relayed her excitement over receiving job applications from young professionals in AI and engineering but expressed concern that many of them become overly fixated on minor details in their career paths. 

"I often find myself in a mentoring role during conversations with job candidates," she explained. "My aim is not solely recruitment; I want to encourage them to concentrate on what truly matters. They should consider their passions, the alignment with the company's mission, and the potential impact they can make."

Despite her seniority and experience, Li emphasized that being a founder or leader in the tech industry does not confer omniscience. "As a CEO and entrepreneur, our fundamental goal is to contribute positively to society and empower individuals," she stated. "Having a 'God complex' isn't productive, and it’s not a mindset I wish to adopt."

Fei-Fei Li describes her management style at her AI startup as being like a 'tiger mom'

A bright, modern startup conference room flooded with warm natural sunlight: Fei-Fei Li stands at the head of a long wooden table, arms folded in confident “tiger mom” authority. Behind her, a large stylized tiger-head logo of her AI startup is mounted on a textured wall. Around the table, professional team members jot notes beside open laptops displaying AI code and network diagrams. Golden morning light streams through floor-to-ceiling windows, casting dynamic shadows that underscore the driven, nurturing energy of her leadership.

Fei-Fei Li Says She Runs Her AI Startup Like a 'Tiger Mom' - Business Insider

- Fei-Fei Li takes on a parental role at her AI startup, World Labs.  
- She describes herself as a "tiger mom," maintaining high standards for her team.  
- Li emphasizes mentoring young talent to focus on passion and meaningful impact.  

Fei-Fei Li says running an AI startup has brought out her "tiger mom" side.

In an interview with Bloomberg released on Wednesday, the Stanford professor, known for creating ImageNet and often called the "Godmother of AI," said that her seniority at the startup has put her in a bit of a parental role.

Li cofounded World Labs in 2024. The startup is developing "world models" designed to help AI understand and interact with 3D environments. Earlier this year, the company announced $1 billion in new funding from investors including AMD, Autodesk, and Nvidia.

"I'm by far the most senior in the company," Li said. "From my co-founders to my engineers, and researchers, and scientists, they're so talented, but they're also young."

"I do feel like a mom," she added. "And oftentimes I feel like a tiger mom. I do run this shop as, I would say, high standard."

Related video LinkedIn marketing leader Jessica Jensen said flexibility will help companies attract and retain top talent

The term "tiger mom" generally describes a highly demanding and strict parent who pushes their children to work hard, be disciplined, and strive for high achievement. It was popularized by American author and Yale Law School professor Amy Chua in her 2011 memoir, "Battle Hymn of the Tiger Mother," which detailed her strict approach to raising her daughters, including demanding top grades and intensive music practice.

Li previously said she finds herself taking on the role of a mentor with younger people — and she's shared some anecdotes about AI's impact on students and early careers.

In a 2025 podcast, she said it was a privilege to receive job applications from young AI and engineering talent, but she found that they overfocused on the small details of their career decisions.

"Sometimes I do want to encourage young people to focus on what's important because I find myself constantly in mentoring mode when I talk to job candidates," she said. "Not necessarily recruiting, but just in mentoring mode."

She said she encourages them instead to think about where their passion lies, whether they believe in a company's mission and team, and what kind of impact they could make.

Still, Li said that being a founder or senior leader is not an invitation to think you know more than everybody else.

"As a CEO, as an entrepreneur, our founding premise is we are here to contribute to society and to empower people," she said. "I don't think having a God complex is constructive, and I wouldn't want to do that myself."

Fei-Fei Li says she runs her AI startup like a 'tiger mom'

AI hardware often has a track record of underwhelming outcomes, but Pocket aims to turn that narrative around.

The startup specializing in AI notetaker hardware claims to have achieved an impressive $100 million run rate just eight months after its debut. With many AI devices failing in the market, Pocket believes that its slim recorder, which attaches to smartphones to transcribe discussions, is poised for longevity and relevance.

"It's only going to go up from here," stated Akshay Narisetti, cofounder and CEO, in an interview with Business Insider. "We’re seeing substantial revenue growth from subscriptions."

While the term "run rate" is frequently used by startups to indicate growth, it’s essential to note that it reflects projected revenue based on current performance rather than actual revenue for the prior year.

According to Narisetti, Pocket is experiencing roughly 50% monthly growth with aspirations to reach a $1 billion run rate within the next two years. The startup operates with a lean team of just 12 employees and has already sold over 200,000 devices.

A participant in this year’s Y Combinator accelerator program, Pocket’s cofounders—Narisetti and Gabriel Dymowski—met during their time at The Residency, a hacker house in San Francisco supported by Sam Altman.

Initially, they created an app focused on meeting notes. Yet, they discovered that users like busy professionals such as doctors and salespeople preferred a hardware option for easy recording of consecutive meetings.

"Instead of fumbling with their phones, they could simply press a button," Narisetti explained.

After experimenting with various formats, including glasses and pendants, similar to the $700 device launched by Humane AI in 2024—which faced criticism and was subsequently discontinued—Pocket settled on a credit-card-sized gadget priced starting at $129. Users can access a free basic subscription, while premium offerings, which feature advanced AI functionalities, start at $16.59 monthly.

Amid their rise, Pocket and other AI notetakers have sparked debates in Silicon Valley about the ethics of recording conversations without consent.

Narisetti contends that having a visible recording device is better than discreetly capturing audio via smartphones, stressing that the Pocket must be placed visibly on a desk and requires a button press to activate recording. He noted that wearables with always-on listening capabilities are often frowned upon.

Competition remains fierce, with tools from Google Meet, Zoom, and Microsoft Teams increasingly utilizing AI for note-taking, alongside startups like Granola targeting users seeking automated summaries. Additionally, OpenAI is expected to release a much-anticipated AI hardware device next year, designed by renowned former Apple designer Johnny Ive.

According to Narisetti, Pocket's edge is in facilitating offline conversations where no singular platform governs the experience.

Most of Pocket's growth has stemmed from direct sales to consumers, referrals, and organic word-of-mouth, with Amazon accounting for a mere 5% to 10% of its total revenue. Approximately 85% of its sales are still domestic, but the company is planning to venture into retail and international markets while expanding its offerings to include personal note-taking and dictation solutions.

Recent data from Crunchbase reveals that global venture capital investment in physical AI surged to $47.4 billion over 521 deals in the first half of 2026—a remarkable fourfold increase from the $12 billion raised during the latter half of 2025.

In June, Pocket secured $11 million in venture funding, led by Accel, as reported by Pitchbook. Other notable investors include Y Combinator and Guillermo Rauch, CEO and cofounder of Vercel. While Narisetti refrained from disclosing a new valuation, he emphasized the company's approach to fundraising, focusing on current operational performance rather than speculative future growth.

"We prefer to not inflate valuations," he remarked.

Pocket's widely-used AI notetaker, introduced 8 months prior, is experiencing rising revenue, but will it sustain?

A bright modern workspace bathed in warm morning light: on a sleek desk sits a laptop displaying an AI notetaker app with gold revenue bars and a soaring arrow. Behind it, a large Pocket logo is mounted on the wall. An adult professional stands beside the desk, hand on chin, peering thoughtfully at the screen. Through a floor-to-ceiling window, a sunlit city skyline hints at rapid growth—and the question of what comes next.

Pocket's popular AI notetaker launched 8 months ago. Revenue is soaring, but can it last? - Business Insider

- Pocket has achieved a $100 million run rate within eight months of its launch.  
- The startup's device records conversations and targets professionals needing seamless note-taking.  
- Pocket plans expansion into retail and international markets, broadening its product offerings.

AI hardware has had more misses than hits. Pocket wants to change that.

The AI notetaker hardware startup says it has reached a $100 million run rate eight months after launching. While the tech graveyard is already littered with failed AI devices, Pocket is betting its thin recorder, which snaps onto the back of smartphones and transcribes conversations, is not just a fad.

"It's only going to go up from here," cofounder and CEO Akshay Narisetti told Business Insider. "We have so much money coming in from subscriptions."

Run rate is a popular way for young startups to describe growth, but it is not the same as actually generating $100 million in revenue over the previous 12 months. Instead, it takes a company's current revenue pace and projects it over a full year.

Narisetti said Pocket is growing roughly 50% a month and believes it could reach a $1 billion run within two years. He said the company has only 12 employees and has sold more than 200,000 devices.

Pocket was part of this year's Y Combinator accelerator class. Its two cofounders, Narisetti and Gabriel Dymowski, first met at The Residency, a San Francisco hacker house backed by Sam Altman.

They initially built an app for taking meeting notes, but later found that users, such as busy doctors and salespeople, preferred dedicated hardware so they could seamlessly record back-to-back appointments and meetings.

"Every single time they tried to unlock their phone, they could rather press a simple button," Narisetti said.

After trying glasses and pendants, some similar to the $700 pin that Humane AI released in 2024 and discontinued after poor reviews and sales last year, Pocket landed on a credit card-sized device costing as little as $129. A basic subscription is free, while premium plans offering more advanced AI models start at $16.59 a month.

Criticism in Silicon Valley

Pocket and other AI notetakers have become so popular that they have led to criticism that people no longer ask for permission to record every conversation in Silicon Valley

Narisetti argued that a visible device is preferable to a covert recording on a phone because the Pocket has to be placed on a desk and requires pressing a button to start recording. He contrasted that with always-on wearable devices, which he said people "inherently hated" because they are constantly listening.

Pocket faces plenty of competition. Google Meet, Zoom, Microsoft Teams, and other major platforms increasingly offer AI-powered note-taking, while startups such as Granola compete for users seeking automated meeting notes. OpenAI is planning to release a much-anticipated device as early as next year, designed by former Apple legend Johnny Ive.

Narisetti believes Pocket's advantage lies in offline conversations, where no single platform automatically controls the experience.

A basic subscription is free, while premium plans offering more advanced AI models start at $16.59 a month. Pocket

Pocket says most growth has come from direct-to-consumer sales, referrals, and word of mouth. Amazon represents less than 5% to 10% of revenue, Narisetti said, and roughly 85% of its business still comes from the US. The company plans to expand into retail and international markets and move beyond conversation notes into personal note-taking and dictation.

In the first half of 2026, global venture funding in physical AI totaled $47.4 billion across 521 deals, according to new Crunchbase data. That's a fourfold increase compared to the second half of 2025, when physical AI startups raised $12 billion across 470 deals.

Pocket last raised $11 million of venture funding in a deal led by Accel in June, according to Pitchbook data. Other backers include Y Combinator and Guillermo Rauch, cofounder and CEO of Vercel. Narisetti declined to put a new valuation on the company, but said Pocket prefers raising based on where the business is today rather than pricing in years of expected AI-driven growth.

"We don't like to grow into valuations," he said.

Pocket's popular AI notetaker launched 8 months ago. Revenue is soaring, but can it last?

Vivodyne, a biotechnology startup, claims there's a significant data deficit within the AI-driven drug discovery sector, and it has developed a solution to address this issue.

The company's innovative modular robotic labs, known as HIVE, can cultivate 20 types of human tissues and autonomously administer doses and conduct monitoring. This process generates causal biological data that is currently lacking, as most existing data primarily relies on animal testing or studies involving single cells and proteins, rather than living tissues.

Andrei Georgescu, the CEO and co-founder of Vivodyne, raises a pertinent question: “Without human testing, what can these AI models achieve?” He emphasizes that they might only succeed in “curing cancer in mice.”

Dario Amodei, CEO of Anthropic, echoed this sentiment over the weekend, noting that claims of AI curing cancer have become more of a cliché than a viable expectation, stating, “what will actually work is curing cancer."

Interestingly, Amodei himself had previously suggested that AI could play a role in cancer cures, while Sam Altman has often referenced this potential as a driving reason for OpenAI’s pursuit of artificial general intelligence and expansive data centers. Similarly, Demis Hassabis of Google DeepMind mentioned last year that AI might eradicate all diseases within a decade.

However, the actual results from AI-driven drug discovery have been modest. Although a few AI-designed drugs have reached human trials—a single candidate even entered Phase III testing—the obstacles many of these initiatives face are not necessarily ones that current AI technologies can overcome.

Despite AlphaFold's significant contributions to understanding the fundamental elements of life, it has yet to lead to the creation of a new drug. Isomorphic Labs, founded with the intention of building upon AlphaFold's technology, has reported delays, pushing back its anticipated trials initially scheduled for 2025 to late this year. The company has stressed that successful drug discovery needs “highly accurate predictive models covering a wide range of biochemical properties and interactions.”

Georgescu advocates for a “sanity check” in the industry, expressing concern that prevailing models lack the necessary data to comprehensively understand human biology. This challenge is a broader issue for the pharmaceutical industry, where 90% of drugs that prove effective in animal models fail to gain regulatory approval for human use.

What sets Vivodyne apart is its foundation; the company emerged from the University of Pennsylvania in 2021, following Georgescu's completion of a PhD in bioengineering there. Vivodyne asserts that its human tissues closely mimic the functionality of genuine human organs—its liver cells boast a 94% predictive accuracy in toxicity tests compared to human trials, its airway tissue demonstrates a 96% match with real human tissue behavior, and its bone marrow shows a perfect 100% concordance in tests involving 20 different chemotherapy agents.

Recently, Vivodyne, which has successfully raised nearly $80 million across two funding rounds led by Khosla Ventures, inaugurated what it describes as the world's largest "human data center" just outside San Francisco. Georgescu states that his team is already achieving double the throughput of all animal trials conducted across the U.S.

Vivodyne aims to streamline the drug development process by gaining a better understanding of which candidates are likely to succeed prior to the costly clinical trial phase, which typically requires tens of millions of dollars. While it is not disclosing its corporate partners, Vivodyne claims to collaborate with several major pharmaceutical companies to tackle a challenge that Georgescu likens to automotive crash testing—automakers have confidence in their vehicles meeting safety standards before actual testing, unlike drug developers who often lack such assurance going into clinical trials, where most candidates do not achieve FDA approval.

Looking beyond immediate concerns, Georgescu envisions that these autonomous biology labs will play a crucial role in generating the causal data needed to train new AI models focused on human biology. He refers to a recent study published in Nature Methods, which suggests that existing cellular data does not follow clear scaling laws when training generative AI models.

Georgescu highlights that most training currently involves static snapshots of cells, meaning models are unaware of the processes that lead to certain cellular states. “They learn merely that ‘this is cell state A’ and ‘this is cell state B,’ but not that ‘cell state B results from inflaming cell state A,’” he explained.

In contrast, Vivodyne’s HIVE systems are conducting hundreds of thousands of live experiments where diseased tissues are subjected to various stimuli. Georgescu anticipates that this approach will yield the reinforcement learning necessary for AI models to accurately interpret human biology, facilitating more significant advancements in healthcare.

He believes this foundational understanding will be essential not only to address current medical challenges but also for a future where complex diseases will require therapies that target multiple biological pathways simultaneously. Georgescu emphasizes, “To explore combination therapies, the search space expands dramatically—it cannot be approached experimentally. We must pinpoint the effects we want and determine the causes required to achieve them. Establishing causality in human biology is fundamental to all of this.”

This startup claims it understands what is needed for AI to effectively treat cancer.

A dimly lit biotech lab in warm amber tones, with fragments of a broken AI robotic arm strewn beside glass tumor‐cell models on a bench. In the background, an adult scientist—back to us—studies a chalkboard covered in cancer-cell schematics and algorithmic notes. On the adjacent wall, a softly illuminated blueprint pinned under a single spotlight hints at the bold new roadmap this startup claims will finally beat cancer.

AI isn’t close to curing cancer. This startup says it knows what it will take. - TechCrunch

- Vivodyne's HIVE labs grow human tissues to generate causal biological data for AI models.  
- The company aims to improve drug discovery success rates before expensive clinical trials.  
- Georgescu emphasizes the need for causal data to enhance AI understanding of human biology.  

A biotech startup called Vivodyne says the AI drug-discovery industry has a data problem, and that it has built a machine to fix it.

HIVE, modular robotic labs built by the company, can grow 20 kinds of human tissue, then autonomously dose and monitor them, generating the kind of causal biological data that today’s AI models are missing — data that today mostly comes from animal testing, or studies of single cells or proteins, not living tissue.

“Absent human testing, what are these [AI] models going to do?” asks Andrei Georgescu, Vivodyne’s CEO and co-founder. “They’re going to cure cancer in mice.”

Even Anthropic CEO Dario Amodei wrote over the weekend that claims that AI will cure cancer have become more cliché than credible — “the thing that will work is actually curing cancer,” as he put it.

To be fair, the idea that AI will cure cancer is something Amodei himself has tossed out in previous essays; Sam Altman has repeatedly cited curing cancer as a justification for OpenAI’s push toward AGI and ever-larger compute buildouts; and Google DeepMind’s Demis Hassabis said last year that AI could potentially cure all disease within a decade.

The actual results remain tepid. A handful of AI-designed drugs have proceeded into human trials — one as far as Phase III, widespread human testing — but the reality is that the roadblocks aren’t necessarily ones that AI can solve today.

Nobel-prize-winning AlphaFold was a big advance for understanding the building blocks of life, but it has yet to actually produce a new drug. Isomorphic Labs, founded to build on AlphaFold, is expecting its first trials, originally planned for 2025, by the end of this year. In February, the company wrote that true drug discovery will require “highly accurate predictive models, across an expansive range of biochemical properties and interactions.”

Georgescu says the space needs “a sanity check” — that existing models don’t have the data to capture the complexity of human biology. It’s a challenge already facing the pharmaceutical industry, where 90% of drugs that are effective enough in animal testing to enter clinical trials don’t receive regulatory approval for humans.

Vivodyne’s plan is different. Vivodyne was spun out of the University of Pennsylvania in 2021, after Georgescu received a PhD in bioengineering there. The company says its tissues closely match the behavior of real human organs — that its liver cells have 94% predictive accuracy compared to human trials that test for toxicity, its airway tissue matches the behavior of real human tissue 96% of the time, and its bone marrow has achieved 100% concordance in tests of 20 different chemotherapy drugs.

Last week, the company, which has raised just under $80 million across two rounds led by Khosla Ventures, opened what it calls the world’s largest “human data center” just outside of San Francisco, and Georgescu says his team is already achieving twice the throughput of all the animal trials being held in the U.S.

The biolab of the future? Image Credits:Tim Fernholz / TechCrunch / TechCrunch/Tim Fernholz

The idea is to accelerate the path of drug candidates by having a better idea of what will work before going through the expense of a clinical trial, which typically costs tens of millions of dollars. Though it won’t name its partners publicly, Vivodyne says it is working with multiple major pharma companies to solve a problem that Georgescu compares to automotive crash tests: An automaker is typically confident its car will pass NHTSA requirements before testing it, but drugmakers rarely have that same confidence going into a clinical trial, where the vast majority of drugs fail to win FDA approval.

But there is a larger vision: Georgescu sees his autonomous biology labs as key to generating the kind of causal data that can be used to train new models on human biology. He points to studies like this one, published in Nature Methods last month, that find no clear data scaling laws when training generative AI models on existing cellular data.

“All the training is done on static snapshots of these cells, and the models are not conditioned at all by the how a cell got to that state,” Georgescu told TechCrunch. “In other words, the model learns ‘this is cell state A,’ ‘this is cell state B,’ but never ‘cell state B is the effect of inflaming cell state A.’”

Vivodyne’s HIVE machines, however, are tracking hundreds of thousands of ongoing experiments where diseased tissue is exposed to some stimulus, which Georgescu expects to provide the kind of reinforcement learning that will produce AI models that understand human biology enough to make more meaningful progress in healthcare.

Georgescu believes that will be key not just for today’s medicine challenges, but also for a future where complex diseases require drugs that, unlike the majority of those available today, target multiple pathways.

“If we want combination therapies, the space that has to be searched explodes — it can’t be an experimental approach,” he told TechCrunch. “You have to say, ‘I want this effect to happen, so what cause should I invoke?’ Establishing causality in human biology is the basis of all of this.”

Stripe is set to purchase AI gateway company OpenRouter for over $7 billion, according to reports.

More from AI Startups & Investments

Harvard's $699 startup bootcamp features AI avatars of its teachers.

Michael Polansky is training an AI model on skin that’s still alive

How AI accounting startup Rillet raised $100M and became a unicorn in 48 hours

AI data startup Micro1 reaches $500M gross run rate amid AI training boom