• Cooley is joining a sudden rush of law firms creating their own AI technology, with help from Legora.
• Cooley Go Lab is an online portal built to help startups with routine contract review and drafting.
• The portal will be available exclusively to founders in Y Combinator's summer cohort to start.

Law firms know more clients are asking chatbots for advice before ever calling a lawyer. So legal giant Cooley is building technology that it hopes founders will use instead.

Cooley plans to give select startups access to Cooley Go Lab, an online portal where founders can upload files and ask questions about their documents, Matt Bartus, global cochair of Cooley's emerging companies and venture capital practice, told Business Insider.

To build it, Cooley teamed up with Legora, a fast-growing legal technology startup that sells to law firms and corporate legal departments. Last year, Legora entered a new line of business with what it calls "portals" — white-labeled workspaces where law firms and their clients can work together on legal matters.

Cooley Go Lab is aimed at catching a common startup problem early, Bartus said. Founders often handle routine contracts themselves to avoid outside counsel's hourly rates. That can leave startups with a trail of messy agreements that their first in-house lawyer has to unwind later.

Legora founder Max Junestrand knows the problem well. When he started the company at age 23, he said he used an early version of ChatGPT to rewrite contracts. Junestrand, a software engineer, not a lawyer, said he let some early contracts include an unlimited liability clause — a provision that can leave a company exposed to damages far beyond the value of the deal.

"When our general counsel started, she freaked out," he said.

Cooley is now trying to give the next crop of founders a way to use artificial intelligence to move faster, but with a law firm's guardrails around it.

Cooley Go Lab will have a limited rollout to start. It will be available first to startups in Y Combinator's summer cohort.

Much has changed for startups since Legora's turn in the famed startup accelerator. Teams can write code and release technology faster with coding agents. They are signing customers and growing revenue earlier, and the hottest companies seem to be raising funding nonstop. But moving faster also means legal work that once came later in a startup's life is being pulled closer to the beginning.

At Y Combinator, partner Gustaf Alströmer is seeing that shift play out in real time. In the last batch, a record 14 startups reached $1 million in annual recurring revenue — the amount of revenue a company expects to collect over a year. Alströmer said giving founders access to tools like Cooley Go Lab could help them keep that pace without creating contract-slop.

The portal includes features that review documents like nondisclosure and contractor agreements and flag issues for founders to consider. The tool also draws on Cooley Go, the firm's central hub of standard startup forms, templates, and guidance.

If you can't beat 'em, join 'em

Law firms like Cooley are facing a new reality. The better the frontier models get at legal work, the more founders and in-house lawyers may decide they can handle situations themselves rather than send them to outside counsel. Anthropic is trying to make that easier with new tools built for contract review and drafting.

Law firms are now figuring out how to stay relevant as more legal work becomes software-enabled. Some are building their own tech. Kirkland & Ellis has tapped Palantir to help it build tools to manage parts of the firm's private funds practice, while Freshfields is working with Anthropic on software that could eventually be sold to other law firms. Harvey, a leader in legal technology, says it's getting into training custom models for law firms.

Taken together, the moves point to a new attitude spreading across Big Law: If clients are going to use artificial intelligence anyway, law firms want to shape how they use it.

Bartus doesn't seem especially concerned about losing business to in-house legal departments. Cooley has been on a hot streak. Profits rose 6.7% to $922 million last fiscal year. The firm also scored a fair use win for Meta in a copyright case involving its model Llama last year, and it advised Jony Ive's hardware startup Io in a $6.5 billion sale to OpenAI.

Bartus is confident that companies will continue to depend on law firms for the important work. Cooley Go Lab, he said, is meant to help founders handle routine legal work more cleanly, not replace lawyers.

Cooley says the tool is not protected by attorney-client privilege, so founders will need to be careful about what they upload because those materials could be turned over in litigation.

"If you want actual legal advice," he said, "you need to talk to a lawyer."

No auge da euforia de 2020 e 2021, o mercado de venture capital parecia viver em um universo paralelo: valuations estratosféricos e rodadas gigantescas, frutos do excesso de liquidez da época. Era um tempo que, como disse Rodrigo Baer, naquela época à frente da estratégia de early stage do Softbank, “tinha muita gente fazendo bobagem”. […]

O post Menos bobagem, mais ambição global: a nova fase das startups brasileiras, segundo Rodrigo Baer, da 14B apareceu primeiro em NeoFeed.

• Billions of dollars are riding on the promise that artificial intelligence can absorb legal work.
• Crosby, a tech-driven law firm, built a benchmark to measure how well models negotiate contracts.
• Redline Bench is meant to help lawyers answer whether they can trust the technology's work.

Legal technology wants its vibe-coding moment. But first, it has to prove the tools can think like a lawyer.

Taking up the task is Crosby, a startup-meets-law-firm that sells basic legal services to companies, including Cursor and Rogo. On Wednesday, it released the Redline Bench, a tool built to measure how well artificial intelligence models perform real-world legal tasks, starting with contract review.

Software engineers have spent the past few years watching these systems get shockingly good at writing code and debugging errors. Now legal tech companies are chasing a similar prize: artificial intelligence that can review contracts, spot risks, and haggle terms faster and cheaper than lawyers.

But law has a problem that coding does not, says Ryan Daniels, a former in-house lawyer turned Crosby founder. "It's really hard to define 'good' or 'bad,'" he said.

Models can write code that either runs or breaks. Legal work is a murkier target. A sales contract can be edited, or "redlined," in lots of defensible ways, Daniels explains. A change that one lawyer sees as prudent, another might call too aggressive.

That ambiguity has become a headache for companies racing to automate legal work, from the scrappy neofirms to the model labs themselves. Anthropic has spent the past few months courting in-house lawyers with tools built for them. That push has been closely watched by investors. Earlier this year, Anthropic's new legal plugin stirred a sell-off in legal tech stocks.

Benchmarks are one of the main ways companies track progress. The labs building frontier models use them as stress tests, measuring whether a new system is better at tasks than the last one.

Coding has hundreds of benchmarks for evaluating models. But the legal industry still lacks a shared way to answer the question: Is the AI's work any good?

Crosby has been working on a new yardstick. The company pulled its engineers and lawyers into a tactical unit called Crosby Intelligence to build agents for Crosby's law firm and a benchmark to grade them against. That team includes engineer Sharan Ramjee, who worked on transformer models to sniff out fraud at Stripe, and Ross Weiser, a lawyer who joined from elite law firm Sullivan & Cromwell.

Crosby also partnered with Micro1, a company that helps model-makers recruit expert workers, to find more lawyers who could help define what counts as good legal work.

To build the benchmark, senior lawyers simulated software deals and marked the contract changes they considered most important at each stage of the negotiation. Those changes were turned into weighted criteria.

When Crosby runs a new test, it gives models the same contracts and asks them to make their own edits. Then a panel of three judges compares these redlines with the lawyer-built rubric. The judges vote pass or fail on each item, and the final score shows how often the models made the kinds of edits that lawyers considered important.

Redline Bench will be made public so any lab can put its models through Crosby's paces. Crosby also plans to regularly release reports tracking how major models compare.

The first release of the Redline Bench put ChatGPT 5.5 at the top of the heap, with a score of 50.5%, meaning the model's redlines matched half of the edits that lawyers prioritized. Gemini 3.5 Flash followed at 45.1%, and Claude Opus 4.8 scored 44.4%.

Crosby was able to test Anthropic's highly capable new model, Fable 5, only once before Anthropic pulled it off the shelves. The results were promising, with a score of 47.3%. When access is restored, Crosby will run the benchmark again and update it.

Crosby isn't the only company trying to measure how the models stack up. Harvey, one of the best-funded legal startups, has released benchmarks for case law research and contract review.

Anthropic and OpenAI also build their own benchmarks to measure performance on real-world tasks. But Daniels said those results can be hard to trust. Over time, the labs eventually tune their systems to perform well on their own tests, he said.

The stakes are bigger than a scoreboard. Billions of investment dollars are riding on the promise that artificial intelligence can lower legal bills and absorb work that used to pile up on the general counsel's desk.

Lawyers will only use the tools if they trust them. Crosby wants to give them a reason to.

• UK-based business owner Sean McDonnell relies on AI for his web design business and SaaS website.
• The White House ordered Anthropic to cut foreign access to Fable 5 while McDonell was mid-task.
• McDonnell emphasizes importance of backup plans due to AI tool disruptions like the Fable incident.

This as-told-to essay is based on a conversation with Sean McDonnell, 43, who lives in England. McDonnell is the founder of the web design company Kaizen and the SaaS website Consigns. The following has been edited for length and clarity.

Developing my website would not have been possible without AI.

I started my web design business earlier this year, which also led me to create a website that provides software to help companies track their waste. I run both of these ventures with my partner, and we enlist contractors for some operations and software development.

We're a small team, and AI tools are a big help. Last week, I saw a few posts online showing the amazing things that Anthropic's new Fable model can do.

I was keen to try this new technology, but didn't get much of a chance to use it. A few hours in, I was mid-task when the US government forced Anthropic to cut off foreign access to Fable with little to no notice.

The rug got pulled from under me pretty quickly, but because I was well-prepared, it didn't have a hugely disruptive impact on my business. It's a reminder that you can't rely too heavily on AI as a founder, and you should always have a backup plan in case of unforeseen circumstances.

I was keen to give Fable a try, but it was short-lived

I like using OpenAI's Codex for repetitive, code-intensive work, and Claude for tasks that help design the product's aesthetics. AI has been able to completely change the architecture of our codebase in a day, whereas a task like that would've taken a developer weeks to do manually.

After seeing so much about it online, I wanted to use Fable to conduct a full review of our product for safety and security flags. The model was in the middle of making some key changes to our codebase when it got shut off instantly with a notice saying, "Claude Fable 5 is currently unavailable."

I didn't realize until the next day that this had happened because the US government ordered Anthropic to block foreign access to the model. It's been a bit of a bummer, and I feel bad for the people at Anthropic for making a brilliant product and having the rug pulled from under them, too. I'm also quite annoyed we didn't get to do more with Fable. I think it could've propelled us so much further.

Being prepared helped us avoid a huge disruption

This isn't the first time we've had issues with using Claude. In the past, when we used Opus 4.6, it would stop mid-task because it hit the token limit so quickly. We didn't realize how token-heavy the tool was, and it left our codebase in a bit of a mess.

Because we'd learned this lesson with 4.6, we made sure we were more prepared for unforeseen circumstances with using AI.

When we started our product review last week, I asked Fable to create a guide that both Claude or other AI models could follow. This enabled us to pass the remaining tasks to other agents when we lost access to Fable. We passed some to Codex and others to Claude 4.8. If we hadn't been prepared this way, the Fable issue could've resulted in lots of work being out the window.

Fable getting pulled didn't have a major impact because we were ready for it, but it ruined our momentum. We're working on a deadline, and every minute counts, so delays like this can be quite disruptive.

Always have a plan B

This Anthropic incident has solidified my conviction that you can't depend completely on AI.

If the government were to shut off AI access completely, our business wouldn't end, because we've already built out our platform, but we are quite dependent on AI. A situation like that would likely increase our costs, partly because we'd have to switch to the old-school method of hiring developers.

In today's AI era, it's important to always have a plan B. Don't just rely on one AI tool. It's good to understand the strengths of different models.

Make sure you're documenting things as you go by keeping records that exist outside your AI tool. If Claude knows all about our code base, but it gets pulled tomorrow, would I be able to give that over to a developer? At this stage, I think I could, because I've been documenting everything as I go. It's a fail-safe.

A spokesperson from The White House told Business Insider, "The Trump administration is collaborating with AI industry leaders to balance cutting-edge innovation with national security concerns that affect both the United States and our allies."

Anthropic did not immediately respond to requests for comment.

Do you have a similar story to share? If so, you can reach out to one of the reporters at aapplegate@businessinsider.com and ccheong@businessinsider.com.

• Josh Giegel joined SpaceX in 2009 and worked there for 3 years. He says the equity he received has been "liberating."
• Giegel's SpaceX equity has allowed him to put a down payment on a house and help pay off his wife's student loans.
• "The equity also allows me to take a lower salary at my startup," Gambit, he said, and that means he can hire more people.

This as-told-to essay is based on a conversation with Josh Giegel, the 41-year-old cofounder of the AI startup Gambit, who lives in Los Angeles. It's been edited for length and clarity.

I was in grad school at Stanford, finishing my master's and wanting to do a Ph.D.

I had worked at NASA the previous summer, and one of the women I worked with was also a Stanford graduate, and was like: "You're going to be so bored at NASA. Why don't you check out this small space company in Los Angeles called SpaceX?"

I applied and interviewed in the two weeks between flight three and flight four of Falcon 1. I interviewed with Elon; he was still interviewing pretty much everyone at the time. I remember going back to my advisor and saying, "There's nothing I'd rather do on the planet than what he just described."

My Master's ended at the end of 2008, and I began in 2009.

I was on what's called the propulsion analysis team, which was four or five people. Our responsibility was: How do you design the first reusable rocket engine? A very small group of us was responsible for the initial stuff that was on Falcon 9.

I started there when I was 23, and I left when I was 27. It was a little bit of naive immaturity. I knew I wanted to start a company one day, and SpaceX was growing like crazy. I wanted to be on a founding team. I still love the company; I almost went back two or three years later before I ended up starting a company of my own.

The IPO is pretty cool. I'm on a bunch of text threads with guys who were there around the same time, and a couple of them are still there. It's cool to see just how big it became.

When I got there, and they gave the offer, there was an equity component. I remember the HR woman who was going over it with me saying, "We think some day, in 10 or 15 years, this might be worth $250,000-300,000." I distinctly remember her saying, "It might get you a nice down payment on a house in Los Angeles."

We all laugh about it now. But, at the time, the saying was: the fastest way to become a millionaire in space is to start as a billionaire.

Buybacks have been really regular for the last 10 years. Every now and then, we'd take a little bit out. For example, we paid off my wife's student loans a number of years ago. We put down a down payment on a house.

I joke: We did actually get a down payment on a house! She wasn't lying when she said that. It's a house that, on our normal salaries at startups, we wouldn't have been able to afford without that additional windfall.

We also love traveling. We've got a seven-year-old and a one-year-old. We're going to go on slightly more adventurous trips because of it.

My wife is also thinking of doing a larger career change that would come with a decent salary reduction, which she probably wouldn't have been able to do without something like SpaceX.

Professionally, I've always been risky. If the majority of your net worth is tied up in a rocket company, you must be a risk-tolerant individual.

Gambit is a VC-backed company. We've raised about $15 million to date, and there are a couple more investment rounds that are coming. The IPO puts you in a position where folks with a substantial amount of equity could be interested in becoming investors.

At least ten of the people I worked with intimately have started their own company. There was a band that I played in with five SpaceX people; four of us started our own companies. I played guitar.

That whole ecosystem can fund its own endeavors and each other. The quantum of capital that they can put in is not like your typical family and friends round. That's typically $20,000, $50,000, maybe $100,000. Here, that could be on the order of $1 million, maybe $2 million per check.

You also become a bit of a mercenary, asking, "I don't need a paycheck from what I'm going to go do, so what am I going to go do?" It's liberating.

The equity also allows me to take a lower salary at my startup, so that I can go out and hire more people to make my company more successful.

• Founders and workplace experts said that post-pandemic AI startups have a different work dynamic.
• AI startup employees often voluntarily come to the office and work longer hours without an RTO.
• Founders said that in-person work fosters a high-trust environment that spurs innovation.

"What is an RTO?"

That was Together AI CEO Vipul Ved Prakash's response when asked by Business Insider whether he had ever had to send a return-to-work (RTO) memo to push employees back to the office of the cloud compute startup.

"People generally like to come in," said Prakash. "We've never enforced it."

Prakash's response illustrates a stark cultural difference between AI startups formed after the COVID lockdowns and long-established corporations, with people voluntarily coming to the office, sometimes on weekends.

Nicholas Bloom, an economics professor at Stanford University, told Business Insider that the age demographics and personal stake many startup employees have in their companies created a work mode that is "almost entirely in-person" and "100% work focused."

"For a single 23-year-old with equity worth $20 million, it makes sense to work in the office for 100 hours a week," said Bloom. "They don't work from home, they home from work."

The tight-knit culture of AI startups

Arvind Jain, founder and CEO of Glean, an enterprise AI for productivity, said he "was not eager" to bring his team members back to the office because finding an office is a hassle, but everyone wanted to be in person and return to their original mode of working when the company first started right before the pandemic in 2019.

"We just simply didn't know how to work from home because everybody was in this one small room," said Jain of the early days of the pandemic lockdown. "We used to be sitting next to each other, brainstorming what to build, and so we found that very, very hard."

Over time, said Jain, he learned to enjoy remote work and got to spend time with his family, but the team genuinely wanted to be together again.

"That's the difference — there's this startup spirit, and it's only 10, 15 people, and we want to be with each other," said Jain. "They love each other, they bond with each other, we used to play games together, and we have very fond pre-pandemic memories as a close-knit group."

Jain said that as Glean grew more rapidly in recent years, it has since moved into a larger office space and dedicated Thursday as its work-from-home day.

Spiros Xanthos, founder and CEO of Resolve AI, an enterprise technology startup that builds multi-agent AI systems, said the company has a "very strong culture" of in-person work and has never had to ask anyone to be in the office.

"We have a fairly big office now, and we have breakfast, lunch, and dinner," said Xanthos. "Most people have lunch in the office together with their colleagues, and many people stay to have dinner in the office."

Xanthos said that since founding the company in early 2024, "cohesion and culture and friendship" among employees has been critical for the company, and that he often brings colleagues based in New York to the Bay Area for offsite retreats so the team could get to know each other better.

"People will actively avoid working remotely at this point," Xanthos added. "Especially for some of the younger folks who didn't have many years of experience, but maybe worked remotely before this, many of them tell me it's day and night — the fact that they have so many friends at work now that they can trust."

AI's innovative nature demands in-person interactions

Richard Florida, an urban studies theorist and professor at the University of Toronto, said the AI wave has unique characteristics compared to other startup booms, which may generate greater in-person demand.

"Innovators have to be close to end users because end users are a part of the innovation system," said Florida, of why it's easier to work in person in the AI industry.

"If you're an AI company, the technology itself is interesting, and you can invent it, but what you really learn is by interaction with the end user, by interacting with your customers and clients," Florida added.

Xanthos said the demand for in-person attendance ultimately boils down to the nature of an innovative industry.

"As a company, we're solving very, very hard problems, and to solve these problems, you operate at the frontier," said Xanthos. "And this means that you need to experiment a lot, try a lot of things that might fail."

"That in turn requires a very high trust in an environment of psychological safety where people feel that they have the ability to innovate bottom up," Xanthos added, "Where they don't need to be told what to do, where there is communication velocity and bandwidth."

So the next time you speak to an AI startup founder, don't ask how their RTO is going — they're probably too busy trying to squeeze everyone into the office.

Agentes que otimizam investimentos de acordo com os objetivos do usuário, trocam de fornecedor quando os preços de mercado mudam ou fazem as compras do mês com base no que falta na geladeira — tudo isso de forma automática. Essa realidade, que até pouco tempo parecia distante, está cada vez mais próxima com os avanços […]

O post Fintech brasileira leva Pix à era da inteligência artificial apareceu primeiro em NeoFeed.

• Business Insider is compiling its third annual list of VCs funding social startups.
• We want to hear from you about who the standout VCs are in 2026 — and where they're placing bets.
• Please submit your ideas through this form (or below) by June 4.

With AI taking over feeds and the biggest social media platforms feeling less social than ever, there's an opportunity for a breakthrough startup.

Business Insider wants to know: Who is backing those companies?

Some venture capitalists are on the hunt for the "next big thing" in consumer technology — even as VC interest in the broader consumer category is in flux.

There are new social networking startups vying to be the next generation's Facebook, Tumblr, or TikTok. The space is also brimming with new companies trying to disrupt dating and day-to-day life with AI assistants that help you navigate your in-real-life relationships and plans.

We're publishing our third annual list of investors funding innovative startups building technology that's advancing the internet, and meeting the needs of people who want to feel more connected (whether that's online or IRL).

In 2025, the investors Business Insider highlighted were placing bets in categories like IRL social, vibe coding, and platforms for niche interests or communities.

This year, we want to know which leading investors are backing social and consumer-facing internet startups, and where they are placing their bets.

Please submit your ideas through this form (or below) by June 4:

A startup brasileira Sinatra AI, que usa inteligência artificial para reduzir perdas no e-commerce, movimentou R$ 10 milhões em uma rodada seed. O aporte foi liderado pela BluStone e teve participação de Caravela Capital, GR8 Ventures e Plug and Play. Além da injeção de capital, parte dos recursos da rodada foi destinada à compra da […]

O post O que a BluStone viu na Sinatra AI, startup que caça erros no e‑commerce apareceu primeiro em NeoFeed.

Nova York – Se o inverno das startups foi o período marcado pela escassez de capital, o momento atual é o início de um novo ciclo de investimentos, marcado por empresas que amadureceram, aprenderam a gerir caixa, buscar rentabilidade e manter o crescimento mesmo com escassez de recursos. “Hoje temos muitas empresas maduras, em estágio […]

O post Brasil sai do “inverno das startups” com empresas mais fortes e capital mais seletivo apareceu primeiro em NeoFeed.

• Tokenmaxxing is all the rage in Big Tech. For startups, the trend is opening up debate.
• Some founders told Business Insider that they spent big on tokens; others used capped subscription plans.
• One founder called tokenmaxxing "extremely stupid." Another said: "You've got to spend money to make money."

Kavitta Ghai wants her startup's engineers to spend more tokens.

The 29-year-old cofounder of Nectir started setting minimum quotas for Claude Code use. First it was at least $100 in tokens a week, then $200. Now, the expectation is that her engineers each spend a couple thousand in AI tokens a month.

The strategy has been successful, Ghai said. Some of Nectir's senior engineers were previously skeptical of AI coding tools; now, they call it their "army of coders," she said.

But she doesn't think Nectir is "tokenmaxxing," the buzzword du jour for techies racing to spend as much as they can. "We don't really play into the Silicon Valley trends," Ghai said. "We live in our own world, and we're competing against ourselves."

Across Big Tech, engineers are racing to spend as many tokens as possible. A token is a measure of AI compute. The more tokens burned, the more the engineer employs AI tools. Employees at Meta reportedly competed on a token leaderboard before it was taken down.

What of the little guy? Startups are an edge case: relatively tiny teams that want to be on the cutting edge of tech but might not have the same money to spend. Some startup leaders told Business Insider that big token bills helped them succeed. Others scoffed at the idea, preferring to stick to the lower-cost subscriptions.

The startups spending big on tokens

Aron Solberg doesn't want the competition of a token leaderboard — but he does want the mindset behind it.

The 44-year-old cofounder of Risotto sees token spending as a "force multiplier" for a small team. The company uses OpenAI and Anthropic's models, and said it spends $4,000-5,000 per month on tokens. Six months ago, Risotto says he spent one-tenth of that sum.

"It's trending up a lot," Solberg said.

"There's an old adage that rings true," he said, whether it was for hiring new employees or spending liberally on tokens: "You've got to spend money to make money."

Quang Hoang is similarly spending big. He wrote in an email that his startup, Vybe, has an "unlimited credit policy" and was thinking about minimum quotas.

Investors are also incentivizing spending — and might foot the bill.

Hoang tells founders he invests in to allocate "at least their salary amount to tokens." (Nvidia CEO Jensen Huang made headlines last month for saying he would be "deeply alarmed" if one of his $500,000 engineers did not consume at least $250,000 of tokens.)

Accelerators like Y Combinator offer free token credits to their participants. "At YC, we let our engineers let it rip," CEO Garry Tan wrote on X. Those credits help some founders to spend big. These founders aren't tokenmaxxers, but do believe that there are productivity benefits.

Traverse cofounder Lance Yan believed in Tan's message: "We usually just let it rip." The 19-year-old said he uses the best models with the maximum effort, not worrying about the costs. Between his Claude Max subscription and the credits that offered by YC, he can spend big without hitting a limit.

He's not a fan of rationing tokens. "That's stupid," he said. "You're just harming your own startup."

26-year-old Boris Skurikhin said that the YC credits helped his startup Docket get off the ground. He's mostly run through them now, except for the models he uses less frequently.

Skurikhin said he noticed a 10x increase in productivity in his own work when he used the tools. "It is expensive to build with tokens," he said, but "not as expensive as having another engineer."

Many of these startups are in the AI game, after all. Nectir's Ghai said that token spending instilled "AI literacy" — something that's especially important, given their product.

"The team itself needs to be the best versed at it first, before we try to go sell it to anyone else," she said.

The startups saying no to tokenmaxxing

Rishabh Sambare wishes he could spend more on tokens.

The 23-year-old cofounder of Gale prefers to build with Zed, an AI IDE similar to Cursor, but can't stomach the company's usage-based pricing. The subscription deals from OpenAI and Anthropic are so deeply subsidized that he uses them instead.

"It sucks, because I hate their products," he said, calling Zed "more polished and less buggy between releases."

Sambare is Gale's only engineer, though the company often has 2-3 interns. He hasn't hit a rate limit, but one of his interns has. They got him a second subscription, he said; it was still far cheaper.

These subscriptions — sending $100 to $200 to Anthropic for its "Max" tiers or $100 to OpenAI for its "Pro" plan in exchange for a stable of discounted tokens — were popular among the founders I spoke to. Hassan Ismail, the 24-year-old founder of Argos Research, said the Claude Max subscription was a "no-brainer," and that all five team members have a $200 a month subscription.

Others were more philosophically opposed to the trend. Weave's Brennan Lupyrypa didn't mince his words: "It's extremely stupid for any company to be tokenmaxxing."

Weave is still spending big on tokens because it doesn't want to "kneecap" its engineers, its 25-year-old founding engineer said. The company set up a notification for when an engineer hit $500 in token spending a month; Lupyrypa said most hit it within two weeks.

But Weave doesn't incentivize the spending itself, which Lupyrypa said was the wrong proxy. He predicted the downfall of tokenmaxxing within the next three months. "CFOs won't be happy," he said.

Still, some tokenmaxxers hold strong. I asked Risotto's Solberg about these token-hesitant founders. He said that they likely hadn't found their product-market fit yet.

"It makes complete sense to spend a lot of money on tokens, because you know that the growth is coming soon after," Solberg said. "If you're a venture-backed business, that's what you signed up for."

• The banker says he has received multiple offers from employees since posting the deal this week.
• The 13-acre Mill Valley estate features sweeping views of San Francisco, an infinity-edge pool, and a spa.
• The offer comes as Anthropic's valuation on secondary markets reached $1 trillion, and shares are scarce.

A tech banker really, really wants Anthropic shares.

The hunt for shares in Anthropic has become so frenzied in recent weeks that Storm Duncan is offering up his $4.8 million Marin County estate in exchange for stock.

"If you're going fishing, you've got to put a worm on the hook," said Storm Duncan, the founder and managing partner of Ignatious, a tech boutique investment bank, in an interview with Business Insider. "What's my other option? Not being in it?"

The offer comes as Anthropic's valuation on secondary markets soared to $1 trillion, driven by investors who have been wowed by its torrid revenue growth and momentum around its AI-powered coding assistant, Claude Code, Business Insider reported this week.

Duncan, who lives primarily in Jackson Hole, Wyo., also owns other properties, but he decided to list this one because he thought it would be especially attractive to Anthropic employees.

Duncan's 13-acre, fully furnished Mill Valley estate features sweeping views of San Francisco, an infinity-edge pool, and a spa.

"It's a 20-minute commute to the Anthropic offices in the city," he said. "No one from Anthropic probably wants my Miami or Jackson Hole place."

By offering the property, Duncan hopes to get on the radar of employees who have legitimate shares to sell and own a goldmine of Anthropic stock they can't sell until after the company goes public.

Duncan says he has received multiple offers since posting the deal this week. "Some of them are [Anthropic] employees, and some of them just happen to have invested early," he said. "I believe they're serious, but it's a complex transaction."

"There's probably a decent number of people who are sitting in a one-bedroom apartment in San Francisco even though they're earning $400,000 a year and are worth a $100 million," he said. "But they can't access that because their stock is so illiquid, so this gives them an opportunity to diversify."

It's not the first time there's been an unconventional way to secure shares in pre-IPO tech companies. In 2005, artist David Choe chose Facebook stock over $60,000 in cash to paint murals at Facebook's first office. That choice led to an estimated windfall of about $200 million once Facebook went public in 2012. In the dot-com era, some real estate owners asked startups for company stock in exchange for leasing space in San Francisco.

Some on X have dismissed Duncan's offer as a publicity stunt or a sure sign of the top of a bubble. Others have made cracks about the only thing being more precious than Anthropic shares is Bay Area real estate.

Duncan insists the offer is real and he is not seeking attention. As for why he does not simply buy shares in the company, he says a small investor like him would never be able to secure stock directly.

"Anthropic can't spend time with people like me," Duncan said. "They're looking for people who can write $100 million in a single check." (The company did not respond to a request for comment.)

The alternative is to buy shares from early employees or investors on secondary markets, but Duncan says those deals are often increasingly dubious.

He said the scarcity of shares on the secondary market has made sellers offer deals that can be rife with high fees and opaque ownership structures.

Duncan already owns shares in Anthropic that he acquired in its 2024 funding round, when it was much easier to obtain shares. He says he was recently convinced he wanted to double down after being wowed by the results of his firm's implementation of Claude Code.

"It's probably going to triple our throughput and reduce our costs by 50%," he said. "As I started to implement the platform at my own firm, I said I would like to have more exposure to this."

Os sinais de que a operação da NeuralMed inspirava sérios cuidados vieram em junho de 2025. Na época, a startup chegou ao pior momento da sua história, iniciada em 2018, ao computar uma receita recorrente mensal abaixo de R$ 100 mil e uma carteira de apenas oito clientes. Esses indicadores não condiziam com a tese […]

O post NeuralMed renova pitch e renasce com novo nome apareceu primeiro em NeoFeed.

A UnblockPay, fintech de pagamentos globais e infraestrutura de tesouraria via stablecoins, anunciou na quarta-feira, 25 de março, a conclusão de uma rodada seed na qual levantou US$ 4,5 milhões (R$ 23,6 milhões) e foi avaliada em US$ 20 milhões (R$ 105 milhões). A rodada foi liderada pela Prelude, fundo de venture capital de Londres, […]

O post Gestoras internacionais “compram” tese de fintech para avançar em pagamentos via stablecoins no País apareceu primeiro em NeoFeed.

A partir de 2019, o Softbank se tornou um dos investidores mais ativos da América Latina. Com cheques grandes e uma tese agressiva de crescimento, o grupo japonês ajudou a impulsionar uma geração de unicórnios no Brasil e na região, investindo em empresas como Wellhub, Rappi, QuintoAndar, Kavak, Unico, MadeiraMadeira, Creditas, Ualá, Petlove, Merama, Omie, […]

O post IA, caixa e paciência: como o Softbank reescreveu seu “manual” para investir na América Latina apareceu primeiro em NeoFeed.

• San José airport starts testing an AI robot called José to assist travelers.
• The pilot test launched on Tuesday amid travel chaos at many US airports.
• Some TSA workers have stopped coming into work due to a government shutdown.

One of Silicon Valley's main airports just made its newest hire, a robot named "José."

San José Mineta International Airport is turning to artificial intelligence to ease the strain of modern air travel, debuting "José," a humanoid robot, as some US airports grapple with staffing shortages and widespread delays.

Developed by Silicon Valley startup IntBot, José is designed to greet passengers, answer questions, and provide real-time updates while autonomously navigating busy terminals.

The robot will be stationed in SJC's Terminal B as part of a four-month pilot, "singlehandedly running his own gate," according to an email previewing the test that referred to José as the airport's "newest hire."

Airport officials said the launch highlights San José's role as a testing ground for emerging technologies to improve customer service.

"By piloting IntBot, we're exploring how artificial intelligence can enhance the passenger journey while reinforcing SJC's role as the gateway to Silicon Valley," said SJC Director of Aviation Mookie Patel.

The timing is notable. Airports across the US have been hit by long security lines and travel chaos, driven in part by many Transportation Security Administration workers not reporting to work during a partial government shutdown. With TSA agents going unpaid at the height of the spring break season, some airports have struggled to maintain normal operations.

José the robot represents a broader push to automate parts of the airport experience, from passenger assistance to information delivery.

SJC officials said the pilot will help evaluate how multimodal AI, combining vision, audio, and language, performs in real-world environments.

The future of air travel may include a robotic helping hand — and it can't come fast enough for weary vacationers stuck in long lines.

Sign up for BI's Tech Memo newsletter here. Reach out to me via email at abarr@businessinsider.com.

• At law firms, the billable hour is the standard way to charge clients. But timekeeping is a pain.
• The startup PointOne says it's using AI to help lawyers auto-complete timesheets and bill more time.
• PointOne raised $16 million in a funding round led by the venture capital firm 8VC.

Tracking hours is part of how lawyers get paid. It's also the bane of the profession.

A startup called PointOne wants to eliminate the most tedious part of a lawyer's job. It says its AI-powered platform passively tracks a lawyer's computer activity and uses it to complete timesheets.

The company grew revenue tenfold since July, says PointOne cofounder Katon Luaces, after signing up dozens of law firm customers, ranging from a global 1,200-lawyer outfit to solo practitioners.

Investors are taking notice. After making a small earlier investment, the Joe Londsale-founded venture firm 8VC is leading a $16 million Series A round for PointOne, Luaces tells Business Insider. Existing investors Bessemer Venture Partners, General Catalyst, and Y Combinator also participated.

Founders are flooding into legal tech, betting they can turn large language models into products law firms will trust — and competing for attention in an estimated $1 trillion industry.

Jack Moshkovich, an 8VC partner, said the market is crowded with companies trying to help lawyers do work faster. That leaves more whitespace, he said, on the operational side of the business.

Luaces isn't a lawyer. In 2019, he was a computer science major and a Google intern as the company's researchers were laying the groundwork for modern large language models.

He saw legal work as a natural target for the technology because so much of it is repetitive and text-heavy. By 2023, he and his roommate, Jeremy Ben-Meir, along with a third cofounder, Adrian Parlow, started sketching out an idea for a legal startup. (Parlow left PointOne last year and joined legal-tech giant Legora.)

When Luaces asked lawyers which part of the job they hated most, he kept hearing the same answer: timekeeping. At most law firms, the billable hour is the standard way to charge clients. Lawyers log the work they do for each client — often in six-minute increments — then tally those hours and bill accordingly. Many still track their hours in a spreadsheet or by hand on a legal pad.

PointOne's platform runs in the background as lawyers move between apps, then fills in time entries with the client, matter, a description of the work, and standardized legal codes.

Security and confidentiality are essential for law firms. Clients trust them with trade secrets and other closely held information, leaving little room for error from any software vendor.

When asked how lawyers feel about software watching them work, Luaces said their dislike of timekeeping helps overcome any discomfort. PointOne says it encrypts stored sensitive data, does not train models on firm data, and gives firms the option to use models in a private Azure environment.

For lawyers, "this is like magic beans," Luaces said.

Time savings aren't the point

Law firms are still working out how to use artificial intelligence to work faster without hurting their economics. Software that saves time can also reduce the number of hours a firm can bill.

PointOne, however, is not pitching itself as a way to save lawyers' time. Instead, it says it can help firms capture time that would otherwise go unbilled.

Some share of legal work never makes it into timesheets. Junior lawyers may undercount how long a task took, either because they're still learning or because they're embarrassed. More often, Luaces said, lawyers skip billing for small tasks because logging them takes almost as long as the work itself.

A lawyer might spend four minutes writing a client email. "I can either spend the next four minutes creating the time entry for it, or I can do more work," Luaces said. "Nine out of 10 times, everyone chooses to do more work."

He says the company's software can increase revenue by capturing billable time that would otherwise be lost.

PointOne isn't the only company making such promises. Its biggest competitor, Laurel, provides professional services firms with analytics about their operations, including time. It's raised over $150 million in funding since 2016, compared to PointOne's $20 million total.

PointOne wants to position itself for a broader shift in how legal work gets priced. Corporate clients are pushing back on soaring legal bills, and as artificial intelligence threatens to trim billable hours, firms are under pressure to test alternatives to hourly billing, including fixed fees for certain matters. Luaces said PointOne's data can help firms better understand the labor behind a matter, which in turn can help them price that work more precisely.

Have a tip? Contact this reporter via email at mrussell@businessinsider.com or Signal at @MeliaRussell.01. Use a personal email address and a non-work device; here's our guide to sharing information securely.

• Buyers have been willing to pay a premium of up to 40% to buy Anduril shares.
• The steep markup reflects a two-tiered class for accessing stock in the hottest startups.
• Data from Caplight highlights a supply imbalance, with buyer demand surging to 97% while sellers' demand is at 3%.

Anduril hasn't even finalized its next funding round, and investors are already eager to pay up like it's a sold-out concert. As marquee venture firms Thrive Capital and Andreessen Horowitz line up to back the defense tech startup at a reported $60 billion valuation, others shut out of the deal are scrambling to buy shares on secondary markets at steep premiums.

"Demand is so significant that buyers who have FOMO are willing to pay huge premiums for access," said Kelly Rodriques, CEO of Forge Global, a private marketplace exchange for shares of private companies like Anduril. "The company hates when this happens, but it happens."

The frenzy around investing in Anduril reflects the growing divide in private markets: access to the hottest startups is split between the VC firms that get in at a certain price and everyone else, forced to pay up on the sidelines. Anthropic has also seen a premium for secondary shares, though not as significant, said Rodriques.

Those investors shut out of the company's fundraising round are forced to buy via secondary markets, with existing stock in the company being sold by current or former employees or early investors. The rush for shares reminds Rodriques of buying tickets to see Taylor Swift on Stubhub when her concert sells out in minutes.

"It's scalping," he said.

Interested buyers have been willing to pay a premium of up to 40% above the $60 billion valuation to buy Anduril shares, according to Rodriques and Greg Martin, managing director and co-founder at Rainmaker Securities, another private marketplace exchange. The deals are not yet finalized because a willing seller and the company's blessing are still required.

"The magnitude of the premium is unusual," said Martin. "Usually we see premiums in the 5% to 15% range."

Anduril declined to comment for this story. Cofounders Palmer Luckey and Matt Grimm have loudly railed against unauthorized sales of the company's shares, publicly calling out some firms as "frauds."

"If I were an investor looking at this 'opportunity,' I'd run for the hills," Grimm posted in December. "Secondary markets are rife with fraud and bad actors, and it pains me to see these bottom feeders profiting off Anduril's growth while fleecing retail investors through unreasonable or opaque fee structures."

The founders have tightly controlled Andruil's stock, requiring would-be sellers to offer the company a first right of refusal to buy back those shares or assign the sale to a buyer of Andruil's choosing. The limited supply is a major reason shares have been among the hardest to obtain for any startup since last year, driving investor "frenzy."

Data from Caplight highlights a massive supply imbalance in the secondary market for Anduril stock, with buyer demand surging to 97% of total volume compared to just 3% from willing sellers—a stark shift from a 69-to-31 split in February.

If demand for Anduril shares is so high, the obvious question is: Why doesn't the company raise its share price to avoid leaving money on the table?

To explain, Rodriques went back to the analogy of a Taylor Swift concert or Nike shoes. Just because some people are willing to pay more does not mean the company wants to set its prices so high.

"It's the same reason Nike doesn't sell sneakers for $2000 if there's a secondary market for a hard-to-get sneaker," Rodriques said. "It's not in their best interest to charge their customers $2000 for a pair of shoes."

Similarly, Anduril would prefer to raise capital from its chosen VCs.

"The company has gotten to a $60 billion valuation by doing a very detailed and thorough job of working with some of the best investors in the world," he said.

• Miro's CEO says the company is plowing cash into AI subscriptions to help employees level up.
• "Our L&D budget is unlimited tooling," Andrey Khusid said.
• AI adoption is accelerating, and with it come questions about the technology's ROI.

Plenty of companies are still debating whether costly AI subscriptions are worth it. Miro has gone the other way.

Andrey Khusid, cofounder of Miro, the maker of a popular online whiteboard platform, says the company gives employees essentially unlimited access to the latest AI tools as a way to speed up how quickly they learn and work.

That approach is possible, he said, because Miro has been profitable since 2016. The company has raised $476 million to date, and Khusid suggested it does not expect to need more capital.

Khusid framed the spending as a core part of more traditional workplace training. "Our L&D budget is unlimited tooling," he said.

Rather than asking employees to learn on their own time or pay out of pocket, he said, Miro wants that experimentation to happen inside the company, as a shared effort. He later added that there should still be a clear business case for buying any tool.

Miro's strategy is part of a wider shift in tech, where AI adoption is moving from optional to expected. A new study from engineering intelligence platform Jellyfish, based on data from more than 700 companies, found that 64% now produce a majority of their code with AI assistance. Tech giants like Google are pushing employees to use AI tools more aggressively, and Microsoft has begun tying AI usage to performance evaluations. As a result, AI fluency is quickly becoming a core workplace skill rather than a nice-to-have.

Still, Khusid says many executives ask the wrong question about AI ROI. Rather than judging the tools on individual productivity gains or subscription costs, he said Miro is trying to focus on whether the company is moving faster overall.

The company tracks projects through what he described as a "discover, define, deliver" process and measures how long it takes to move from one stage to the next. The goal is to compress that timeline as much as possible.

"The most important metric from my perspective is velocity of innovation," Khusid said. "If you don't innovate fast enough, you're out of the game."

Khusid said he doesn't think the way companies use AI today is necessarily the end state. He said it will take at least until the end of this year, or even next year, to see what a workplace shaped by these tools really looks like. At that point, Miro will take a harder look at which tools are worth the price tag.

For now, he said, Miro is already seeing time savings across engineering, product, and design. That's not always the case, though. Better tools speed code generation, he said, but code reviews can still bog down projects.

"Humans have to read it," Khusid said. At least for now.

Have a tip? Contact this reporter via email at mrussell@businessinsider.com or Signal at @MeliaRussell.01. Use a personal email address and a non-work device; here's our guide to sharing information securely.

• Eight salespeople from ServiceNow have jumped ship to rival startup Serval in recent months.
• Two of the salespeople cited fears about AI as their reason for leaving ServiceNow.
• ServiceNow's stock has tumbled 40% in the last six months in the so-called SaaS-pocalypse.

Eight salespeople from ServiceNow and its newly acquired subsidiary, Moveworks, have jumped ship to rival startup Serval in recent months, Business Insider has learned.

ServiceNow is a cloud computing software company whose stock has tumbled 40% in the last six months in the so-called SaaS-pocalypse, as investors fear AI could decimate the profit margins of software giants. In an effort to stay one step ahead in the AI race, ServiceNow closed an all-cash $2.85 billion acquisition of Moveworks in December to create an "AI-native front door."

The same month, Serval closed a $75 million Series B funding round led by Sequoia Capital, valuing the rapidly growing AI-powered IT support startup at $1 billion. Sequoia was also an early investor in ServiceNow.

Eight employees represent a fraction of ServiceNow's 29,000-person workforce. Still, the exodus shows how difficult it can be for tech companies with falling valuations to retain talent when buzzy, well-funded AI companies come calling. A ServiceNow spokesman declined to comment.

The highest-level departure is Brad Patterson, who had been a ServiceNow sales VP for nearly two years.

"AI is really making serious moves," Patterson said in an interview. "In a similar way, market sentiment is responding; I think people are responding in the same way."

Every incumbent tech company is facing a similar talent drain, according to Jules Levy, ServiceNow's former head of enterprise generative & enterprise AI, who is also among those joining Serval.

"I don't think this is unique to ServiceNow," he said in an interview. "Many folks within those incumbents are looking to jump to AI-native platforms that will be able to move really fast and take advantage of this technological wave."

"I think everyone's trying to figure out what comes next," he added.

Serval is not specifically targeting ServiceNow employees, but when one employee leaves, it can have a ripple effect, according to Tatiana Birgisson, Serval's chief operating officer.

"If you hire really good people, other really good people in their network want to follow them," she said, citing Chris Comes, who became a Serval VP of Sales in November after more than three years at MoveWorks. "Multiple people got excited about seeing the announcement of his going to Servo."

Startups have usually offered less job security and lower cash compensation than public tech companies, but Birgisson says tech downsizing has made it easier to recruit candidates.

Block CEO Jack Dorsey cut roughly 40% of his workforce in February, citing the rise of AI. Meta could reportedly lay off a fifth of its staff amid skyrocketing AI costs.

"Big Tech no longer feels as safe as it once did," Birgisson said. "We are starting to see more candidates who, 5-10 years ago, would not have considered working at a startup, but are now more open to it because there isn't the clear divide between 'secure' and 'risky' jobs."

• Travis Kalanick launches Atoms, a new venture aiming to build a "wheelbase for robots."
• Atoms wants to build a platform for specialized industrial robots, not humanoid designs.
• Atoms is acquiring Pronto, an autonomous vehicle startup founded by Anthony Levandowski.

The former Uber CEO is venturing into robotics.

Travis Kalanick announced that Atoms is out of stealth mode and expanding beyond food delivery infrastructure into industries such as food service, mining, and transportation.

The ex-Uber CEO published a 1,600-plus manifesto of his company on Friday.

"When I told my friends, family, and colleagues about my plans for what was next, they were really excited that I was 'coming back,'" Kalanick wrote on the website for the new venture.

"The thing is, I never left."

Kalanick did not immediately respond to a request for a comment.

In an interview on "TBPN" on Friday, Kalanick told show hosts John Coogan and Jordi Hays that he will be folding his ghost-kitchen startup CloudKitchens into the new venture, a detail that is also mentioned on Atoms' website.

Atoms' webpage says the company plans to build a "wheelbase for robots," a platform designed to power specialized machines rather than humanoid robots.

"At Atoms we make gainfully employed robots — specialized robots with productive jobs that bring abundance to their owners and society at large," Kalanick wrote on the website.

Kalanick said on "TBPN" that the company will focus on practical industrial systems instead of humanlike designs, and that the venture was just renamed as "Atoms" from "City Storage Systems" today.

"We've been in stealth mode for eight years," said Kalanick. "Employees were not allowed to put the name of the company on their LinkedIn. We have thousands of employees."

"Humanoids have their place, but there's a lot of room for specialized robots that do things in an efficient, sort of industrial-scale kind of way, which is sort of where we play," he added.

According to Kalanick, Atoms is close to acquiring Pronto, an autonomous-vehicle startup focused on industrial and mining sites, founded by his former Uber colleague, Anthony Levandowski.

Uber didn't immediately respond to a request for comment.

Kalanick's partnership with Levandowski would be the reunion of two of the most infamous Uber alums.

Levandowski did not immediately respond to a request for comment.

Kalanick co-founded Uber in 2009 and was its CEO until 2017, when he resigned after facing immense investor pressure stemming from reports of a toxic work culture and multiple clashes with regulators.

Levandowski, an alum of Google's self-driving project that is now Waymo, was brought into Uber in 2016 after the ride-hailing giant bought his self-driving trucking outfit, Otto. In less than a year, he was fired from the company after Google sued Uber, accusing Levandowski of stealing trade secrets from the self-driving project. Uber settled with Waymo, but Levandowski was convicted in a separate criminal case in 2020 of one count of trade secret theft.

Levandowski was later pardoned by President Donald Trump before he even began serving an 18-month prison sentence.

• Young tech startup founders are having a moment in the AI era.
• From teenagers to 20-somethings, these founders are raising millions.
• Take a look at the pitch decks some of these founders shared with Business Insider.

Tech is no stranger to young founders.

Steve Jobs was 21 when he cofounded Apple in 1976. Mark Zuckerberg was 19 when Facebook launched. Whitney Wolfe Herd was 25 when she unveiled Bumble.

Many of today's startup founders are still young and scrappy. And in the age of AI, they're even more empowered to barrel ahead.

Some are following the footsteps of tech titans before them and dropping out of college. Others are opting out of the undergraduate experience altogether, with a few ditching high school to pursue careers in tech.

Arlan Rakhmetzhanov, founder of AI coding startup Nozomio, told Business Insider that he dropped out of high school in Kazakhstan after getting accepted into the competitive startup accelerator program, Y Combinator (YC). At the age of 18, he raised $6.2 million for Nozomio.

Rakhmetzhanov isn't the only teenager finding success in AI. There's also Toby Brown, a UK teen who raised $1 million for his AI project. There's also Zach Yadegari, the teenage cofounder of Cal AI, a nutrition app.

College-aged founders are also building companies and raising capital, such as the Yale students behind Series AI, a new social networking startup.

The median age for YC participants is now 24 years old, compared to 30 in 2022, YC's Pete Koomen told The New York Times in August.

Business Insider has interviewed the founders of 12 startups who are 25 years old or younger and have raised millions in funding since 2024 about the pitch decks they used to impress investors.

Read 12 pitch decks founders who are 25 years old or younger used to raise millions:

Note: Founders were 25 or younger when Business Insider published the following articles.

Series A

• Posh, an events platform, raised a $22 million Series A when its founders were 23. Read the 12-page pitch deck it used.
• Clyx, a startup trying to connect people via in-person events, raised a $14 million Series A when its founder was 25. Read 22 pages of its deck.

Seed

• Ditto, an AI dating startup founded by UC Berkeley dropouts, raised $9.2 million when the founders were 23 and 24. Read its 12-page pitch deck.
• Lyra, an AI video call startup, raised a $6 million seed out of YC when its founder was 23. Read the 8-slide pitch deck it used.
• Nexad, an AI adtech startup, raised a $6 million seed after wrapping up A16z's Speedrun accelerator. Nexad's CEO was 25. Read the 10-page pitch deck.
• Orange Slice, a YC-backed sales tech platform, raised $5.3 million when its founders were 23. Read the 7-page pitch deck.
• Golpo, a generative AI video startup, raised a $4.1 million seed out of YC when its founders — who are also brothers — were 19 and 20. Read its 7-page pitch deck.
• Bluejay, an AI agent startup, raised a $4 million seed coming out of YC when its founders were 23. Read its 9-page pitch deck.
• Novoflow, an agentic AI startup building tools for medical clinics, raised $3.1 million when its founders were 18 and 19. Read its pitch deck.
• CodeFour, an AI police tech startup, was founded by two 19-year-old MIT dropouts and raised $2.7 million coming out of YC. Read the pitch deck.
• Cerca, a dating app that connects people with mutual friends, raised a $1.6 million seed when its CEO was 23. Read the 10-slide deck.

Pre-seed

• Series, an AI social networking startup, raised a $3.1 million pre-seed when its founders were 21.

This story has been updated with additional examples.

A SpaceX, empresa de transporte espacial e de satélites fundada por Elon Musk, está lançando uma venda secundária de ações que poderia avaliar a companhia em US$ 800 bilhões, segundo pessoas familiarizadas com o assunto, relatou o jornal americano Wall Street Journal.

Se confirmado, a nova transação faria da SpaceX, mais uma vez, a startup mais valiosa do mundo, superando o recorde anterior de US$ 500 bilhões estabelecido pela OpenAI, dona do ChatGPT, em outubro.

Leia também:

• Startup de IA de Musk busca levantar US$ 15 bilhões e pode chegar a valor recorde de US$ 230 bilhões
• A nova corrida espacial se tornou uma competição entre bilionários: Jeff Bezos contra Elon Musk

A última oferta pública da empresa, discutida pelo conselho de diretores na quinta-feira (4) no hub Starbase da SpaceX no Texas, poderia mudar dependendo do interesse de vendedores e compradores internos, disseram as fontes, que pediram anonimato por se tratar de informações não públicas.

Sob um cenário preliminar, a SpaceX poderia incluir um preço por ação de cerca de US$ 300, o que avaliaria a empresa em aproximadamente US$ 560 bilhões, segundo duas das fontes. Os detalhes da transação ainda podem mudar antes do fechamento, podendo alcançar avaliações ainda mais altas, disse uma terceira pessoa.

Um representante da SpaceX não respondeu imediatamente a um pedido de comentário.

O último valor representaria um aumento substancial em relação aos US$ 212 por ação estabelecidos em julho, quando a empresa levantou recursos e vendeu ações com uma avaliação de US$ 400 bilhões.

O Wall Street Journal e o Financial Times, citando fontes não identificadas, relataram anteriormente que o negócio avaliaria a SpaceX em impressionantes US$ 800 bilhões, embora a Bloomberg não tenha conseguido confirmar a informação de forma independente.

A notícia sobre o valor da SpaceX fez as ações da EchoStar Corp., empresa de TV por satélite e telecomunicações, subirem até 18%. No mês passado, a EchoStar havia concordado em vender licenças de espectro à SpaceX por US$ 2,6 bilhões, somando-se a um acordo anterior de venda de cerca de US$ 17 bilhões em espectro sem fio para a empresa de Musk.

A SpaceX, lançadora de foguetes mais prolífica do mundo, domina a indústria espacial com seu foguete Falcon 9, que coloca satélites e pessoas em órbita.

A empresa também lidera o setor de serviços de internet via satélite em órbita baixa com o Starlink, um sistema com mais de 9.000 satélites, muito à frente de concorrentes como o Amazon Leo, da Amazon.com Inc.

Executivos da SpaceX já cogitaram transformar o Starlink em uma empresa separada de capital aberto — ideia sugerida inicialmente pela presidente Gwynne Shotwell em 2020.

No entanto, Musk repetidamente questionou publicamente a possibilidade ao longo dos anos, e o CFO Bret Johnsen afirmou em 2024 que um IPO do Starlink aconteceria provavelmente “nos próximos anos”.

O site The Information, citando pessoas próximas às discussões, relatou separadamente na sexta-feira que a SpaceX informou investidores e representantes de instituições financeiras que planeja um IPO para toda a empresa na segunda metade do próximo ano.

Uma chamada oferta secundária ou tender, na qual funcionários e alguns acionistas iniciais podem vender ações, oferece aos investidores de empresas privadas como a SpaceX uma forma de gerar liquidez.

A SpaceX também trabalha no desenvolvimento do novo foguete Starship, anunciado como o mais potente já desenvolvido, capaz de lançar grande quantidade de satélites Starlink, além de transportar carga e pessoas para a Lua e, eventualmente, para Marte.

Jeff Bezos assumirá o cargo de codiretor executivo de uma nova startup de inteligência artificial voltada para aplicações industriais e aeroespaciais, informou o New York Times na segunda-feira (17). A empresa, batizada de Project Prometheus, já arrecadou US$ 6,2 bilhões em financiamento — parte dele investida pelo próprio fundador da Amazon —, tornando-se uma das startups em estágio inicial mais bem capitalizadas do mundo, segundo três fontes ouvidas pelo jornal.

A iniciativa marca o retorno de Bezos a um cargo operacional formal desde que deixou a presidência-executiva da Amazon, em julho de 2021. Embora esteja profundamente envolvido com a Blue Origin, sua empresa espacial, Bezos detém ali apenas o título de fundador.

Com o Project Prometheus, o bilionário entra de vez em um mercado de inteligência artificial cada vez mais competitivo, dominado por gigantes como OpenAI, Meta, Google e Microsoft, enquanto dezenas de startups disputam espaço com produtos e modelos próprios.

Leia também:

• ‘Inventar e Vagar’: o livro que ensina como Jeff Bezos construiu seu império
• A nova corrida espacial se tornou uma competição entre bilionários: Jeff Bezos contra Elon Musk

O novo empreendimento será comandado por Bezos em parceria com Vik Bajaj, físico e químico que já trabalhou de perto com Sergey Brin no laboratório de inovação do Google, conhecido como X (antigo Google X).

Segundo o The New York Times, o jornal norte-americano, o Project Prometheus já contratou cerca de 100 pesquisadores vindos de empresas de IA, como OpenAI, DeepMind e Meta, em meio a uma acirrada disputa global por talentos.

A tecnologia em desenvolvimento da startup é direcionada a aplicações de IA para engenharia e produção em setores como informática, indústria automotiva e aeroespacial — áreas alinhadas ao interesse de longa data de Bezos em acelerar o acesso humano ao espaço. Ainda não está claro onde a startup está sediada nem quando exatamente foi fundada, já que a empresa manteve um perfil discreto até agora.

Vida de Bezos

Desde que deixou o comando da Amazon, Bezos tem dividido os holofotes entre seus negócios e sua vida pessoal — incluindo um casamento repleto de celebridades em Veneza neste ano. Ele também intensificou seu envolvimento com a Blue Origin e ampliou seus investimentos no setor de IA. Segundo a CNBC, todos os aportes feitos em 2024 pela família Bezos, por meio da gestora Bezos Expeditions, foram direcionados a empresas de inteligência artificial.

O lançamento do Project Prometheus ocorre em um momento de forte pressão por mão de obra especializada. Grandes empresas de tecnologia vêm oferecendo salários recordes e bônus milionários para atrair cientistas de IA — alguns chegando a valores de US$ 100 milhões. O mercado continua marcado por uma rotatividade intensa, com talentos migrando entre OpenAI, Meta, Google e startups emergentes.

Mesmo diante de uma competição bilionária, o Project Prometheus desponta como um novo peso pesado. Com US$ 6,2 bilhões já captados, a empresa tem margem para adquirir poder computacional escasso, atrair pesquisadores de elite e desenvolver conjuntos de dados proprietários difíceis e caros de produzir em ambientes industriais, segundo especialistas ouvidos pelo Times.

💾

No início do ano, uma empresa de infraestrutura distante dos holofotes chamada Aligned Data Centers conseguiu arrecadar US$ 12 bilhões — mais do que algumas das principais startups de inteligência artificial do mundo conseguiram em uma única rodada. O objetivo: expandir drasticamente sua presença para atender à crescente demanda pelas instalações que alimentam os sistemas de IA.

Nove meses depois, com grande parte dessa expansão ainda em andamento, a Aligned está em negociações avançadas para ser adquirida pela Global Infrastructure Partners (GIP), fundo da BlackRock, em um acordo de US$ 40 bilhões. A transação está a caminho de ser uma das maiores do ano e, potencialmente, a maior já realizada por uma empresa de data centers. 

As negociações de aquisição, noticiadas pela Bloomberg na noite de quinta-feira (2), citando pessoas familiarizadas com o tema, são indicativas de quanto dinheiro os investidores estão dispostos a investir em empresas consideradas essenciais para o boom da IA. De acordo com um relatório do Goldman Sachs, empresas relacionadas à IA responderam por US$ 141 bilhões em emissões de crédito corporativo até agora neste ano, superando os US$ 127 bilhões em dívida total do ano passado.

À medida que as empresas de tecnologia se dizem preparadas para investir centenas de bilhões, senão trilhões, em infraestrutura física para IA, há uma demanda crescente por empresas como a Aligned, que prometem atender a essas necessidades. Isso é verdade mesmo que a maior parte da capacidade do data center da Aligned – e a receita esperada com ele – ainda esteja em fase de planejamento. 

• ‘Efeito DeepSeek’ aumenta as oportunidades de negócio para data centers do Brasil, diz Ascenty
• A vez do Brasil nos data centers: empresas colocam bilhões em campo à espera da explosão da IA

“Eles têm muita atividade planejada”, disse Ari Klein, analista da BMO Capital Markets, sobre a Aligned em particular. “Você provavelmente está vendo empresas pagando por essa atividade planejada, ou pelo que pode vir a acontecer.”

Fundada em 2013, quase uma década antes do boom da IA ​​generativa, a Aligned há muito tempo se concentra em fornecer data centers personalizados para empresas, com ênfase em eficiência e sustentabilidade. A empresa conta com o apoio da Macquarie Asset Management e possui 78 data centers sob gestão ou em desenvolvimento nas Américas, de acordo com seu portal. 

Assim como outros provedores de data centers, a Aligned tem enfrentado novas pressões nos últimos anos para construir mais e maiores data centers para IA. Em janeiro, a Aligned anunciou ter levantado US$ 12 bilhões em capital e dívida para acelerar seus planos de construir 5 gigawatts de capacidade de data center, o suficiente para abastecer metade da cidade de Nova York em um dia quente.

O investimento tinha como objetivo ajudar a Aligned a “aproveitar oportunidades impulsionadas pela crescente demanda por infraestrutura preparada para IA”, disse Andrew Schaap, CEO da empresa, em uma publicação nas redes sociais na época. “A Align está pronta para enfrentar este momento.”

O único desafio é que os data centers — e toda a infraestrutura, incluindo as fontes de alimentação para apoiá-los — levam tempo. Atualmente, a Aligned tem pouco mais de 600 megawatts de capacidade de data center em operação, com outros 700 megawatts em construção, de acordo com a DC Byte, empresa de inteligência de mercado que monitora o setor. Ainda assim, a capacidade combinada torna a Aligned uma “operação de tamanho razoável”, segundo o fundador da DC Byte, Edward Galvin. 

A Coreweave, provedora de nuvem que fechou acordos com a OpenAI e a Nvidia, possui 470 megawatts de capacidade ativa, de acordo com documentos públicos , com muito mais planejado. A CoreWeave reportou pouco mais de US$ 1,91 bilhão em receita em 2024. E embora a Aligned esteja em negociações para ser vendida por potencialmente US$ 40 bilhões, a Coreweave tem um valor de mercado de mais de US$ 65 bilhões.

A Aligned não divulgou publicamente seus números de vendas. Se cobrasse cerca de US$ 210 por quilowatt por mês, o padrão do setor para preços de data center, segundo a empresa de serviços imobiliários comerciais CBRE, a receita anual da Aligned seria de quase US$ 1,6 bilhão. Esse valor subiria para US$ 3,4 bilhões, incluindo a capacidade em construção. 

Representantes da Aligned não responderam a um pedido de comentário. 

A Aligned também está desenvolvendo um novo data center no Texas para a Lambda, uma empresa de infraestrutura em nuvem apoiada pela Nvidia . O projeto ainda está em construção.