The Incline Village, Nevada-based company is also acquiring Austin-based agentic AI startup as it looks to automate more of the labor-intensive work involved in investigating financial crime.

led the investment, which includes both primary capital and a secondary tender offer for employees. , , and others also participated. Socure did not disclose the terms of its acquisition of Fravity.

With the latest funding, Socure has raised over $742 million in disclosed funding since its 2012 inception. It was previously valued at $4.5 billion at the time of its Series E round in 2021. The company did not break down how much of its raise was primary and secondary capital.

Rapid growth as fraud surges

The transactions come as Socure says it is seeing both rapid growth in its own business and a sharp rise in increasingly sophisticated fraud. The company is refreshingly open about its financials, telling ������Ƶ News that it ended the second quarter with $364 million in annual recurring revenue, up 63% from a year earlier, and added 95 customers during the quarter, including , , and . It also claims to be growing “profitably.”

Socure uses AI and machine learning to help banks, fintechs and government agencies verify identities so they can “approve real customers instantly while stopping fraud.”

It now has more than 3,000 enterprise customers. They include 19 of the 20 largest U.S. banks, more than 600 fintech companies, major sportsbook and prediction-market operators, and 160 public-sector organizations. Specifically, some of those customers include , , , , and . The company’s revenue model mixes usage- and transaction-based SaaS.

AI creates both an opportunity and a problem

Socure co-founder and CEO said AI is creating both an opportunity and a problem for the business. For example, Socure saw an 8,000% increase in AI-driven fraud across its network last year, according to the company, as generative AI and other tools make it easier to create convincing fake identities and automate attacks.

At the same time, AI could help address one of the more costly parts of fraud prevention: investigating the large number of cases and alerts that automated systems flag for human review.

That is where Fravity comes in.

Automating fraud investigations

Fravity has built an AI-native platform that uses agents to automate fraud, risk and compliance investigations. Its technology will be incorporated into Socure’s RiskOS platform as RiskOS_Agents, initially focusing on watchlist screening and monitoring and know-your-business checks.

Socure and Fravity already share several enterprise customers that use the two products together, according to Socure. Across its existing deployments, Fravity has reduced cost per case by 80%, sped up case resolution fivefold and cut false positives by as much as 70%, the companies say.

The acquisition puts Socure more directly into what identity intelligence company estimates is a $71.1 billion financial crime investigation market. The problem is particularly acute at banks, where 53% spend at least an hour reviewing each alert, and 37% manually review more than 40% of alerts, according to Liminal.

As AI increases the volume and sophistication of fraud, Ayers argues that the identity layer — determining whether people and increasingly AI agents are who or what they claim to be — is becoming more critical to doing business online.

“I believe there are two types of companies that matter in the AI-driven global economy: those that are AI-native, and those that fight the consequences of AI acceleration,” he said in a statement.

Expanding beyond financial services

The investment follows a period of expansion for Socure beyond its financial services roots. In May, the company won a five-year, $163 million federal contract to provide identity-proofing technology for Login.gov. It is also pushing further internationally.

Socure had more than 550 employees as of March 2026, more than 100 more than it had about a year ago, according to Ayers.

Related ������Ƶ query:

• ��

Illustration:

By conventional Silicon Valley standards, had some of the credentials one might expect of a startup founder: , experience as a VC on Sand Hill Road, and time working inside a fast-growing venture-backed startup.

One thing he decidedly lacked was a technical background.

“I’ve never written a line of code in my life,” Levtov told ������Ƶ News in an interview. Years later, after building , a 12-year-old San Francisco-based supply chain software company that he says now works with 35 federal agencies and generates double-digit millions of dollars in annual recurring revenue, that remains true: “And I still haven’t written a single line of code.”

Over the years, Levtov raised $42 million in funding for Craft. His experience has given him a close-up view of both the disadvantages non-technical founders face, and the reasons Silicon Valley’s preference for technical founders may be too simplistic.

An unlikely route into tech

Levtov’s own route into technology was anything but direct. His family emigrated from the Soviet Union to England when he was a toddler, and with two musician parents, he began playing cello at age four. He later attended a specialist music school in London, studied at the Royal College of Music, and participated in a Columbia- exchange while earning an English literature degree from .

By graduation, Levtov had decided to keep music as a hobby and pursue business instead. He joined , later attended Stanford Business School, and eventually landed at , an ad-tech startup that grew from about 10 employees to roughly 200 during his time there.

“I was just totally bitten by the bug,” he recalls. “And I said, ‘This is what I want to do with my life. I want to build a company one day. Somehow, entrepreneurship is for me.’”

Before becoming a founder, though, Levtov spent time on the other side of the table as a venture capitalist at . He later left venture for an operating role at video service provider , and after moving back to Europe, eventually worked at helping Silicon Valley startups including , , and establish distribution partnerships.

His eventual startup grew out of an unsuccessful attempt to build an enterprise social network.

As part of that project, Levtov’s team created company profiles by collecting information from corporate websites, job pages, management pages and other sources.

Those profiles began showing up prominently in searches, convincing him there might be a business there.

The disadvantage of not being technical

But unlike a technical founder, he could not simply build the product himself.

“My first coder was literally a $20 an hour Odesk or person,” he said.

That dependence slowed everything down.

“For the non-technical founder, it’s just fundamentally a much longer time at the very beginning to get to something because a technical founder basically codes their idea on nights and weekends,” he said.

Instead, Levtov had to hunt down developers, explain his vision, and try to determine whether the result would match it. Once he’d done those things, he then had to find the capital to pay for it.

Still, the business gained traction.

Its company profiles eventually appeared in 100 million search results per month and drew about 2.25 million visitors organically, according to Levtov.

‘I guess that means not me’

When Levtov began raising venture funding in London in 2015 and 2016, he ran into another challenge familiar to non-technical founders: Investors preferred founders who could build the product themselves.

“I decidedly remember this clarity with which I found venture funds whose websites I go to and research. And what did they say? ‘We support technical founders in doing this and that.’ And it really was just this moment [of realizing], ‘oh I guess that means not me, right?’ ” he said.

Even so, Levtov does not describe himself as having been shut out of venture capital. He had Stanford and Venrock on his résumé and eventually secured funding from in the U.K., and later after moving back to Silicon Valley.

And he believes the preference for technical founders has some logic behind it.

“They’ve got a direct line between the business concept and the code in which it’s executed,” Levtov said.

His own experience showed him how costly that gap could be. He said there were times he hired the wrong technical person and did not have enough expertise to recognize the problem quickly.

“That is a real disadvantage: This inevitable disconnect, this gap between the non-technical person’s knowledge and, you know, the bare metal, as it were, or the most intrinsic innards of the software code by which this business product is going to live and breathe,” Levtov said.

He believes those mistakes slowed the company’s growth.

Finding the business inside the product

But the company’s eventual breakthrough also illustrated the potential advantage of approaching technology from the business side.

Someone at contacted the company and pointed out that its data could help track changes across a sprawling supply chain. The system could pick up signals such as changes in hiring, executive departures and new product offerings.

Lockheed became its first enterprise customer. Then, in 2020, the reached out about using the product to monitor 300,000 companies in the defense industrial base. The company closed a five-year, $6.5 million deal 94 days later, according to Levtov.

“We figured out that our company is actually a supply chain company, and we haven’t looked back since then,” he said.

Notably, those customers were not software developers asking for better developer tools. They were, noted Levtov, business users with business problems.

And this is where he believes his own background helped. A non-technical founder may not be able to evaluate code or engineering talent in a way that a technical founder can, he pointed out. But they may be stronger in areas such as understanding customers, managing people, fundraising and building relationships.

AI is further complicating that debate, since software can increasingly be built without traditional coding expertise. But Levtov stops short of arguing that technical founders no longer matter.

“It really just takes both. It takes both sides,” he said. “I think if you can have a technical founder and a non-technical founder, you’re probably in the ideal spot.”

Technical founders may have an edge at the earliest stages, Levtov said. As companies scale, the balance can shift toward skills like hiring, selling, positioning and dealmaking.

At different points in a company’s life, he said, “it’s really about the tech right now,” while at others, “it’s all about the dealmaking, or all about the positioning, or the marketing.”

Related ������Ƶ query:

Related reading:

• ‘Nobody Wanted to Give A Former Principal Money’: How An Educator Built An Edtech AI Startup With $63M From VCs
• How A Teenage Carpenter Became The Founder Of AI Construction Startup Trunk Tools
• From Humanities To AI: How Ali Hussain Built Fintech Tabs Into A $400M Startup

Illustration:

has been operating in that market since 2013. The New York-based company operates a marketplace for shares of privately held companies, giving employees and other shareholders a way to sell stock before a company goes public or is acquired.

announced plans to acquire EquityZen in October 2025 and completed the deal in January 2026, bringing the company under the investment bank’s umbrella.

, who co-founded EquityZen and serves as its chief strategy officer, has had a front-row seat to the secondary market’s evolution. ������Ƶ News spoke with Haslett about what secondary-market pricing says about today’s most sought-after startups, why AI companies are commanding premiums while many older startups trade at discounts, what the IPO market looks like beyond its biggest names, and why investors are taking a closer look at hard tech.

The following conversation has been edited for length and clarity.

������Ƶ News: The second quarter was one of the strongest venture-backed IPO quarters since 2021, but drove much of that activity. If you remove SpaceX, how open is the IPO market for the typical late-stage startup?

Phil Haslett: Generally, I’d say it’s better than it was three or six months ago. If you were a private late-stage technology company, you probably were going to wait until after SpaceX anyway, so that hurdle is gone.

Tech markets are also doing well. The stock market is at an all-time high, and there’s been a strong recovery in tech stocks overall. I assume that we’re gearing up for a busier summer than usual.

Another thing to consider is IPO performance beyond SpaceX. Some have had initial enthusiasm followed by a slowdown. has come down a bit. So companies may see it as a good time to go public, while post-IPO performance has been, in a word, “meh.”

But within AI, I think we’ve seen that there’s opportunity up and down the production curve — from energy for data centers, to the technology inside them, to orchestration of compute, to efficient spending on training and inference. There are a lot of interesting companies along that spectrum, and I think that bodes well for companies in the space that want to go public.

A few companies entered your Top 20, including , , and . Does that reflect a durable shift away from traditional software, or are investors chasing a small group of scarce, high-profile hard-tech companies?

Haslett: I think it reflects a thematic shift. The companies entering that list generally fall into AI infrastructure, space tech and robotics.

If those are industries we think will have generational growth opportunities, the logical conclusion is that each sector will have winners. SpaceX gets people thinking about opportunities in space and space tech, and by extension defense tech.

The same applies to AI infrastructure. If the market is that big, and we’ve seen companies go public over the last year or so, it stands to reason investors will be interested in other companies in that space. I think that’s more important than simply chasing scarce supply.

These businesses tend to be more capital intensive and may take longer to reach predictable revenue than a traditional SaaS company. How are secondary investors underwriting them?

Haslett: If a company needs more capital, investors have to decide whether the overall opportunity is big enough to justify waiting longer and having the company raise more.

If you have to build a factory or get regulatory approval, that can delay the company’s ability to increase its valuation or reach an exit. Investors discount that into what they’re willing to pay.

Secondary investors are making the same calculus as primary venture and growth investors, so you’d imagine much of that is already baked into headline valuations from primary raises.

What’s changed is that capital-intensive companies now have more financing options. Five or six years ago, a battery company or new chip manufacturer might have had little choice but to raise equity. In 2026, more credit and asset-based financing options are available.

That matters because if one of these companies underperforms or has a distressed asset sale, creditors and lenders get paid first. Secondary investors have to factor that in, too.

EquityZen says the average transaction occurred at a 38% discount to the last funding round, while many AI transactions traded at premiums. What does that say about how bifurcated the private market has become?

Haslett: I don’t know if it’s a mispricing. There are essentially two vintages of private companies right now.

Some companies weren’t built AI-first and have had to adapt. Many raised during the go-go years of 2021, at very high valuations, and may not have raised since. They’ve had to rethink their strategies, which can slow growth and execution. That gets reflected in the discount.

Then there’s a new wave of companies, from 2023 and beyond, that were built with an AI-first mentality. They started from a clean slate, may operate more efficiently, and have a cleaner story for the market.

Some of those companies are raising rounds in quick succession at higher valuations. Secondary investors may pay a premium because they believe the company’s trajectory is clear and the next valuation increase could happen quickly.

is an example from the 2021 cohort. It raised at roughly a $10 billion-plus valuation and just sold for substantially less. It’s still a good business, but when investors compare 20% growth with newer companies going from zero to hundreds of millions in revenue in just a few years, you can understand why their appetite changes.

We may see more companies from that era sell for less than where they raised in 2021.

Over the past few years, many private companies have conducted secondaries because they weren’t ready to go public. When should founders consider establishing a company-approved secondary program?

Haslett: Historically, companies started thinking about liquidity programs after they’d been around five, six, or seven years, largely to reward employees for their patience and provide liquidity to early investors.

Now we’re seeing younger companies engage in controlled liquidity and tender offers.

One reason is talent retention. There are only so many engineers and data scientists, and companies need to compete for them. Secondary liquidity has become more normalized.

More solutions are available than before. Morgan Stanley, for example, has significantly grown its tender-offer activity as investor interest and available tools have expanded.

There’s also more investor appetite. Investors are increasingly willing to gain ownership through tender offers or secondary transactions. Five years ago, that was far less common.

Right now, it’s a very founder- and employee-friendly environment, and investors are willing to support secondary liquidity because they want access. If markets turn, that pendulum could shift back.

For investors considering private-company shares, what does a secondary-market price tell them compared with the valuation at the company’s last fundraise?

Haslett: I think it gives them the true price.

A primary valuation is a point-in-time measure of what investors were willing to pay, and those investors generally received preferred stock with additional rights and liquidation preferences.

The secondary market is more telling of what you could actually get in your pocket now. For companies that embrace secondary liquidity, those prices help employees, former employees and early investors understand what their shares are actually worth.

How does EquityZen calculate popularity and distinguish durable investor demand from curiosity or hype?

Haslett: Our platform allows investors, typically retail accredited investors, to tell us what they’re interested in. They can browse companies, review our analysis, and indicate which companies they would invest in, if shares became available, and at what size.

That gives us a real-time metric of what our user base wants to invest in and how much. It helps guide where we spend our time bringing opportunities to clients.

The last thing we want is to work with a shareholder when we can’t find a buyer, or with a buyer when we can’t find shares for sale.

What does the recent consolidation in the secondary market tell you about how the market is evolving?

Haslett: There was a lot of attention toward the end of 2025 around consolidation in the secondary-market space. went to , and EquityZen went to Morgan Stanley.

To me, that reflects market growth, increasing adoption of secondary liquidity, and the fact that the biggest financial institutions are paying attention. I don’t expect that to change.

Your data showed that some software companies began trading at premiums again in the second quarter. What separates those gaining investor confidence from those still trading at deep discounts?

Haslett: Execution. Leadership and execution.

It’s about a company’s ability to take a legacy SaaS business and turn it into something AI-enabled across the business. Are you using AI tools to improve internal tasks? Are you building AI into your product for clients?

Companies that can combine the stickiness and customer loyalty they’ve already built with their domain expertise and AI are going to do just fine. The ones that are slower to adopt are going to get pummeled.

Six months ago, there was concern that when a company like announced a cybersecurity or legal tool, companies in those sectors would immediately lose value. I think some of that was a knee-jerk reaction.

Customers already using your software have some patience, but they also expect you to keep improving the product and give them a reason not to switch. The companies that are slow to react, or too proud to react, are the ones I think will get hit hardest.

, and 1��are examples of software that is deeply ingrained in large enterprises. If companies can keep their products working well and keep adapting them, they still have a shot at being successful standalone businesses. It comes down to management execution.

Related ������Ƶ query:

Illustration:

spent much of his childhood in his family’s St. Paul, Minnesota, convenience store, where he packed cigarette cartons and watched his father run a business seven days a week, 365 days a year.

Hussain was the son of a first-generation immigrant who arrived from Karachi, Pakistan, and worked his way from employment at a to owning his own corner store. That experience instilled in Hussain a work ethic that stuck with him. But his father made a clear trade-off with him in high school.

“My dad didn’t want me to necessarily come back to the store,” Hussain recalls. “He’s like, ‘Look, like this is what I did and built. Go use school as a mechanism to leave.’ ”

Ultimately, Hussain went on to form , a New York-based AI startup that automates parts of finance and accounting. Founded in 2023, the company has raised around $90 million, employs about 180 people, and was last valued at $400 million, according to Hussain.

But unlike many tech startup founders, Hussain didn’t study computer science in college. Instead, he earned a humanities degree at , won a Marshall Scholarship to , left academia abruptly to work at , and spent six years learning the operational ropes at early-stage startup before launching Tabs in 2023.

From St. Paul to Oxford

Hussain leveraged a scholarship from the to attend Cornell, where he fell in love with comparative politics and history. Fixated on academia, he graduated and immediately headed to Oxford to pursue a Ph.D. Two months in, reality hit.

“I realized this is a terrible idea,” Hussain admits. “I grew up … way too scrappy packing the cooler to survive through a postdoc and potentially a very structured 10-year career, which seemed very hard and long and not in my control.”

Deciding to reset his trajectory at 23, Hussaini took a chance on management consulting at BCG in the Midwest. Though it provided an intensive crash course in business operations, spreadsheet modeling and corporate processes, the structured corporate hierarchy lacked the agency he had seen in his father’s store.

By 2015, he decided to embed himself directly into tech, taking a massive pay cut to join Latch — then a 10-person seed-stage startup — as its first operations hire.

“Had I tried to do this directly out of Oxford or out of BCG, I think [it] would have been impossible,” Hussain told ������Ƶ News in an interview. “One of the things that often keeps many non-traditional founders out is … the ability to access capital, but also understand the playbook of how to build, how to design around a real problem, and build a team.”

Over six years at Latch, as the company grew to tens of millions in revenue, Hussain picked up a few lessons about building venture-backed companies. He learned to pursue large markets, to surround himself with people whose strengths complement his own, and to build for major shifts in technology.

Humanities vision meets deep tech

In 2023, Hussain applied those principles to start Tabs, an AI platform that automates revenue recognition, billing and collections. From the beginning, the founder knew he had to leverage his strengths. He also knew his weaknesses. Hussain recognized that he brought commercial vision and operational execution, not the ability to write code, to the table. So he partnered with a deeply technical co-founder, , to balance his own background.

“I came from the humanities,” Hussain noted. “Tabs is a deeply technical and complex problem to solve, and so having someone who could augment my vision … was a very important part.”

Investors took notice. Early relationships and the operational credibility Hussain built during his “apprentice” years paid off. Tabs quickly raised a $4 million pre-seed round co-led by and . Since then, the startup has grown to roughly 180 employees, raised about $92 million in total capital, reached a $400 million valuation in its last round, and maintained triple- to quadruple-year-over-year revenue growth.

To Hussain, non-traditional backgrounds in tech are a strategic advantage that fosters the resilience required to survive early-stage uncertainty.

“I think a lot of non-traditional folks … have to embrace a ton of volatility, even ahead of being a founder, to make the sacrifices to learn,” Hussain said. “Sometimes it’s just the non-traditional background that allows you to embrace non-traditional ways of learning that ultimately get you into entrepreneurship.”

Related ������Ƶ query:

Related reading:

• ‘Nobody Wanted to Give A Former Principal Money’: How An Educator Built An Edtech AI Startup With $63M From VCs
• How A Teenage Carpenter Became The Founder Of AI Construction Startup Trunk Tools

Illustration:

Venture investors appear to increasingly see physical AI as the next leg of the broader AI boom. Notably, according to a recent article in , many firms known for early bets on software, internet services and social media companies are writing more checks to companies building “physical technologies and materials tied to the artificial-intelligence boom.”

������Ƶ data backs this up.

In the first half of 2026, global venture funding in the space totaled $47.4 billion across 521 deals, per our data. That’s up dramatically — almost 4x — compared to the second half of 2025 when physical AI startups raised $12 billion across 470 deals. It’s also up significantly — by nearly 80% — from the $26.4 billion raised across 436 deals in the first half of 2025.

To give you an idea of just how much more money is going into physical AI companies, here’s a comparison. In the three years spanning 2022 to 2024 combined, venture investors put a total of $41.9 billion into physical AI companies — still several billion less than we’ve seen raised in just the first half of this year alone.

And before we go any further, I should clarify that by our criteria, physical AI includes industries such as robotics, autonomous vehicles, aerospace, drones, industrial automation and sensors.

Noteworthy deals

Several multibillion-dollar megadeals drove the spike in H1 investment. One very large deal in particular accounted for nearly one-third of all venture dollars: Mountain View, California-based ’s raised in February. , , and co-led the financing, which was raised at a staggering $126 billion valuation.

Other companies that have brought in large rounds this year include:

• In May, defense tech startup raised another $5 billion in funding at a $61 billion valuation — double the $30.5 billion valuation it received less than a year earlier.
• San Diego-based in March landed a $2 billion Series G round co-led by and . Its valuation jumped to $12.7 billion.
• In March, Austin-based , a defense tech startup focused on autonomous sea vessels, raised $1.75 billion in Series D funding, bringing its total funding to around $2.6 billion. led the round, which set Saronic’s valuation at $9.25 billion — more than double its Series C level in 2025.

Exits

The physical AI space has also produced several notable exits so far in 2026, although activity has been more concentrated in aerospace, defense and drones than in areas like robotics.

has been the clear outlier, raising $75 billion in its June IPO at a $1.77 trillion valuation. Other notable public debuts include Herndon, Virginia-based space intelligence company , which raised $416 million, and Arlington, Virginia-based autonomous drone maker , which raised $320 million. On the M&A side, one of the most notable deals was roughly $900 million acquisition of Tel Aviv’s humanoid robotics startup , a transaction the company explicitly tied to its push into physical AI.

Investor POV

, general partner at , told ������Ƶ News via email that while funding in physical AI has historically been concentrated in robotics and humanoids, defense, and foundational models, he sees the opportunity as much broader. Physical AI, in his view, represents the convergence of software, hardware, sensors and IoT, and services across a wide variety of real-world applications. What is changing, according to Ziegler, is AI’s ability to process data from those systems at such a scale and speed to generate useful operational insights, while the underlying hardware becomes cheaper and more accessible.

“Even our mobile phones now have LIDAR scanners on them,” he noted, “democratizing the ability to map objects and spaces.”

For Edison Partners, the appeal is particularly strong in high-value, traditionally analog industries where physical AI can become mission-critical infrastructure. Ziegler pointed to manufacturing, supply chain, utilities, agriculture, transportation, government, and physical and spatial intelligence as areas of interest. Many of these companies resemble vertical software businesses, he said, with “attractive unit economics, large deal values and multi-year deployments,” while their combination of software, sensors and hardware can generate proprietary datasets that become increasingly valuable over time. Edison is especially interested in applications where the return on investment is measurable through predictive maintenance, risk management, asset integrity, security and autonomous operations.

The economics of building these companies have also improved considerably over the past two years. Ziegler compared the shift to what cloud infrastructure did for SaaS.

“The costs to build these companies have come down, and AI infrastructure and multi-modal tech to do so is now available,” he said.

Meanwhile, compute and foundation-model capabilities have become more accessible, reusable models and physics-based simulation have improved, training data is more plentiful, and sensor and hardware costs have declined. At the same time, companies are increasingly bundling hardware into recurring or mixed-revenue models and moving toward outcome- or usage-based pricing. That combination, Ziegler said, makes the hardware itself a distribution mechanism for software and data, with “hardware [as] the distribution model for creating a data intelligence flywheel.”

, partner and head of growth at , told ������Ƶ News via email that while physical industries remain capital intensive, AI and other enabling technologies are changing how efficiently companies can build and scale. Historically, the capital required to reach meaningful scale made investors wary, but he argues that “tech barriers are plummeting, experienced talent is pouring in, and market demand is rising.”

That convergence is driving more investment into areas including energy, robotics and autonomy, inference, chips and compute, and data center infrastructure. As a result, he said, funding is increasingly shifting away from experimentation and toward companies that can hit production milestones, land customers and scale efficiently.

For Eclipse, physical AI is not a new theme but a core investment thesis dating back to the firm’s founding in 2015. Fath said the opportunity has become more compelling because “the technical and economic conditions are now catching up to that longstanding conviction,” allowing companies to iterate, deploy products and reach customers faster.

Eclipse defines physical AI broadly as “intelligence embedded in systems that perceive, reason, and act in the real world,” while generally avoiding investments in standalone large-language-model providers. Fath described the firm’s focus as investing on the “shoulders,” rather than the “head.” This means that Eclipse backs both the infrastructure that enables generative AI, such as chips, compute, energy and data centers, and the companies applying AI to build new businesses in the physical world.

He views the current landscape as the result of technology, talent, capital, demand and policy finally aligning. More powerful compute, foundation models, simulation and developer tools are allowing smaller teams to build faster with less capital and labor, Fath points out. Looking ahead, he expects value to accrue throughout the physical AI stack, but believes the strongest moats will belong to companies that vertically integrate and own multiple layers.

Ultimately, he said, “customers value operational efficiency, reliability, and revenue, not technical sophistication alone.” The companies that can turn technical capability into dependable systems at commercial scale — and then use their data and infrastructure to expand into additional products — are likely to capture the most value.

Related ������Ƶ query:

Related reading:

• Anduril Raises Another $5B As Defense Tech Startups Shatter Funding Records
• SpaceX Shares Close Up 19% After Largest IPO Of All Time

Illustration:

Before founding , an AI startup for the construction industry, spent more than a decade working in construction. She grew up in a low-income household in Austria, where her father was a carpenter. When the family needed extra money he would take his children to job sites, and Buchner began working as a carpenter as a teenager.

“I’m a blue-collar worker by background,” she explained.

Buchner never really left the industry. She worked her way up and eventually became a general contractor, managing large construction projects in Europe. But her career took a turn after a worker died on one of the sites she was running.

Experiencing the fatality changed the way she thought about her work.

“Once you realize that people die under your management, you just look at life differently,” Buchner said.

She decided to build a health and safety app for construction workers. That company was separate from (and before) Trunk Tools, but the project pulled Buchner into software and eventually led her to pursue a Ph.D. focused on data science in construction and early artificial intelligence.

While doing that research, she realized something that would later become central to Trunk Tools: Construction companies produce huge amounts of information, but much of it is fragmented across different systems and buried in documents, making it difficult to analyze.

“The data was there,” Buchner said, but it was “so messy and so unstructured” that traditional data analytics could only go so far.

Turning construction data into an AI business

Buchner saw an opportunity to make that information more useful. She also decided she wanted to build the company in the U.S. rather than Europe.

In 2019, she moved to California to attend . She founded Trunk Tools in 2021 while she was still in business school, graduating in 2022. The early years were messy, as they often are for startups, and Buchner said the company did not really find its first meaningful product and market until 2023.

By then, advances in large language models were beginning to make it more practical to work with large quantities of unstructured information.

That was particularly relevant in construction. Buchner estimates that an average construction site can involve 3 million to 4 million pages of documentation spread across numerous systems. Trunk Tools integrates with those systems and creates a layer on top that can analyze the information and use it to carry out various tasks.

The company sells primarily to general contractors and subcontractors, and has also begun working with some project owners.

Trunk Tools’ AI agents can review contracts and construction drawings, flag inconsistencies and help with specifications, submittals and bidding, and pass that information among different workflows. For example, one agent can uncover an issue in a drawing and give that information to another agent to figure out next steps and how any changes might affect other parts of a project.

Such an ability to connect information across a project is important because construction projects are unusually complex, Buchner said.

“For a human brain, it’s impossible to calculate the second- and third-order effects of changes because of how complex and connected it is,” she said.

In one case, Buchner said, a project owner requested a change that a construction company might otherwise have simply begun implementing. Trunk Tools calculated that the change would add nearly $4 million to a roughly $100 million project. Once the contractor presented the cost to the owner, the owner decided not to proceed.

In another instance, a worker got chemicals in their eye, and someone nearby used a Trunk Tools agent to ask what to do. The system returned step-by-step instructions for responding to the exposure.

“With stuff like that, you have to act fast,” Buchner said, calling it a particularly meaningful example for the company.

From early product to rapid growth

New York-based Trunk Tools now has more than 100 employees. Buchner is a solo founder and said her Ph.D. gave her enough of a technical background that she did not feel she needed to seek out a technical co-founder. She did, however, hire experienced AI engineers early and has since built out a larger technical leadership team.

The startup has also expanded the number of products it offers. Buchner said Trunk Tools had two live AI agents about a year ago and now has more than 10. The pace of development has become fast enough, she said, that the company sometimes struggles to train its own sales and customer-success teams on new products as quickly as they are being released.

The company has raised about $70 million in funding from investors such as , and . Revenue grew fourfold last year and is on track to grow roughly 3.5x this year, according to Buchner. Its most recent funding round was a $40 million Series B in 2025 that valued Trunk Tools at $325 million.

An unconventional founder advantage

Fundraising was more difficult at the very beginning, Buchner shared, in part because she did not fit the mold investors often associate with startup founders.

“I think it was hard very early on because I might not look like the traditional tech founder,” she said.

But after Trunk Tools gained traction, Buchner believes her industry background became an advantage. She had spent years doing the work her customers do and could answer detailed questions about construction in investor meetings.

“I walk into a room full of investors, and I know that … I’m definitely the smartest person in the room when it comes to my industry,” she said.

Bringing AI to a tech-wary industry

Historically, the construction sector has been slower than others such as finance and legal services to adopt new software or technologies. One reason for this, Buchner believes, is that there is often a long distance between the person buying a product and the person expected to use it.

Executives may make purchasing decisions while sitting in an office. Meanwhile, the actual users are often out in the field working on job sites and under constant pressure to keep projects moving. Contractors also typically operate on relatively thin margins. So if a bad decision is made around purchasing technology, the potential cost is greater.

Still, Buchner believes that AI is being adopted faster than other technologies she has seen enter the construction industry.

Over time, Trunk Tools has also learned that simply selling the software is not always enough. The company now provides training and change-management services for customers trying to roll out AI across large organizations.

It’s another area where having an industry background makes a difference, in Buchner’s view. Rather than sticking to a model of what a conventional software business is supposed to look like, she said, she is more focused on what construction companies actually need.

“I’m just thinking about solving a problem,” she said.

Looking back, Buchner said she always expected to start a business someday. For years, she assumed it would be her own construction company.

Instead, the carpenter-turned-contractor ended up building a software company for the industry she had already spent much of her career working in.

Related ������Ƶ query:

Related reading:

• ‘Nobody Wanted to Give A Former Principal Money’: How An Educator Built An Edtech AI Startup With $63M From VCs

Illustration:

led the raise, which brings the Austin-based startup’s total funding to $40 million since its 2022 inception. Returning backers , , and also participated in the round. ClearJet’s earlier investors include , formerly JetBlue Ventures, and .

ClearJet’s model is straightforward. Rather than build its own fleet of planes or trucks, it connects shippers with unused cargo capacity on commercial flights already traveling between U.S. cities to move e-commerce packages around the U.S. Customers include major multibillion-dollar retailers, e-commerce platforms, 3PLs and marketplaces.

In just three years, the startup has built what founder and CEO calls a “super carrier,” a network that now spans 95 U.S. airports and connects retailers with major U.S. airlines and multiple final-mile delivery providers. This network gives retailers a way to ship packages so that they travel directly between cities on passenger planes already in the air rather than through the traditional networks used by major parcel carriers.

Unlike a traditional parcel carrier, ClearJet doesn’t own the planes transporting those packages. Instead, its asset-light “Uber for cargo” model taps available capacity on flights that are already traveling between cities. The startup says its approach can cut shipping costs by as much as 35% while speeding deliveries by one to three days.

“We’re basically connecting with the already moving aircraft,” Guggenheim told ������Ƶ News in an interview. “These flights are going from A to B city. We’re taking those same routes, and that’s just why we’re so fast. That’s also why we’re so cost efficient.”

The approach appears to be working. ClearJet is profitable, its revenue has more than tripled year over year, and it is approaching nine figures in top-line revenue, according to Guggenheim.

The market opportunity is still large. The startup says it moves more than 30 million packages annually, which is still a fraction of the roughly 1.8 billion U.S. parcels it considers eligible to move by air.

Global funding to supply chain management and logistics startups has reached $8.4 billion in 2026 so far, per ������Ƶ . This puts this year on pace to top 2025’s total of $9 billion considering we have over four months left in the year.

How it works

Retailers connect to ClearJet through an API and can generate a two-day shipping label. ClearJet picks up the packages, takes them to an airport, handles sorting and screening, places them on commercial flights, and then injects them into final-mile networks at their destination. Those providers can include , the , , , and , Guggenheim said.

“We call it the super carrier because it truly is that, and it gives all the power back to the retailer,” Guggenheim said in an interview with ������Ƶ News.

One of ClearJet’s first large retail customers had previously relied on FedEx for goods arriving from Asia, with deliveries taking seven days from factory to customer, according to Guggenheim. Under ClearJet’s model, products arrive at Los Angeles International Airport, where the company takes possession of the cargo, sorts it and flies it into 14 different airports before handing the packages to final-mile carriers.

The result, Guggenheim said, was a reduction in delivery time from seven days to five — and $35 million in cost savings for the customers.

That combination of time and cost savings was what caught Edison Partners’ attention.

, who leads the firm’s vertical SaaS and AI practice, told ������Ƶ News that Edison had spent years looking at ways to use excess capacity in supply chains without requiring companies to make massive investments in physical infrastructure.

“We looked at a few supply chain businesses over the years,” Ziegler said. “Candidly, most of them went bankrupt because they took an asset-heavy approach to the middle mile.”

ClearJet took the asset-light approach. And the company’s airline relationships, regional sortation infrastructure, regulatory license and technology architecture make it difficult to copy its model, according to Ziegler.

“When you think about what they built, it is a very unique aviation infrastructure platform, and it’s difficult to replicate,” he said. “He’s [Guggenheim] proven the business model, and the unit economics work.”

Backstory

The idea for ClearJet grew out of Guggenheim’s own frustrations as a longtime e-commerce entrepreneur.

He started his first company in 1997 after meeting and his family and building direct-to-fan e-commerce businesses for them. Guggenheim later worked with a range of music and sports clients, including helping launch the first beyonce.com. His company went on to support more than 2,000 Plus stores with over $1 billion in GMV, he said.

Along the way, logistics became one of his biggest headaches.

Shipping had become the second-largest cost of goods outside of the product itself for his business, he said. And in 2019, after spending $55 million with , Guggenheim said he received an email giving him five days’ notice that his account was being canceled because it wasn’t profitable enough.

“And so I said, ‘there has to be a better way,’ ” he said.

Guggenheim began thinking about the thousands of domestic passenger flights traveling around the U.S. every day and wondering whether their unused cargo space could become part of an alternative parcel network.

He had no connections with the airlines, he said, so he began cold-emailing executives until he reached the president of . At a cargo industry event, Guggenheim got about five minutes to pitch his idea in what he now calls his “Shark Tank moment.”

“I said, ‘I want to start flying packages from LA to New York. Do you have any flights?’” Guggenheim recalled. “And he put his arm around me and said, ‘You and I are going to be best friends.’”

Guggenheim went on to build relationships with , , , and , the latter of which participated in an earlier ClearJet financing. He also began recruiting people from the airline and parcel industries and building the technology underpinning the network. ClearJet formally launched in May 2023.

One logistical obstacle was the aircraft themselves. Guggenheim said most U.S. passenger aircraft are narrow-body planes whose cargo doors are too small to accommodate the pallets typically moved by freight forwarders.

ClearJet addressed this by designing its own overpack bags specifically for e-commerce parcels. Those bags can travel through airports much like passenger luggage before being unloaded and handed to the appropriate delivery company.

AI component

ClearJet’s AI models choose each parcel’s path based on cost, speed and geography across its network. AI is also built into how ClearJet routes packages. Its models choose a parcel’s path based on factors including cost, speed and geography.

The company is also developing AI agents to automate more of the operational work around those shipments.

For example, ClearJet is creating AI agents to handle tasks such as rating, booking, tracking and managing delivery problems, according to Guggenheim.

This includes features such as responding to weather disruptions by moving packages onto a different flight or through a different city, something ClearJet can do because it isn’t tied to just one airline.

“We’re building an entire agent team, an army of agents that do everything that the humans were doing,” Guggenheim said.

ClearJet raised its seed round in early 2023 and $13.4 million in a Series A in 2024, according to Guggenheim. He declined to disclose the company’s valuation but described the Series B as a significant step-up from its previous financing.

Next, ClearJet plans to expand into returns and international shipping, and expand its network of airports. Guggenheim also wants to give consumers much more detailed visibility into where packages are during their journeys, similar to the real-time experience they have become accustomed to with services such as DoorDash.

“We have a really big appetite for giving consumers a better visual experience with their packages,” he said.

For Ziegler, the bigger bet is that ClearJet can become infrastructure for a delivery market increasingly built around getting products directly to individual consumers.

“We saw this as an opportunity to actually create a category-defining business,” he said.

ClearJet has just under 50 full-time employees and hundreds of contractors operating seven days a week.

Related ������Ƶ query:

Illustration:

To dig into this trend, ������Ƶ News analyzed a dataset of repeat founders who have gone through YC’s cohorts. That analysis revealed some very interesting insights. The dataset, shared with us directly from YC, consisted of 454 repeat founders through the program as well as 935 founder-company records spanning 2005 through 2026.

What we found is that repeat participation to date has mostly been a two-chapter journey: 428 founders (94%) went through YC exactly twice, while only 25 appeared three times. and co-founder was the sole four-time founder.

Other highlights: Founders typically returned to YC five years after their previous appearance, with an average gap of 5.1 years. However, the data reveals two distinct themes. Nearly 30% of return participations occurred within two years — including 38 in the same calendar year — while 61 returns happened after a decade or more. Some founders jump straight into their next venture, while others take years off to build experience before coming back around.

Repeat founder numbers peak in the most recent data, hitting 65 in 2025. But that doesn’t automatically mean people are returning at higher rates. In recent years, YC cohorts have grown significantly, and the 2025-26 numbers include newer batch formats alongside potentially incomplete data.

It’s also clear that returning to YC isn’t always a solo journey.

Several complete founding teams returned together for subsequent companies, including those behind , and , as well as and .

A trend YC partners are watching closely

, a general partner at the San Francisco-based accelerator who worked the spring 2026 batch, has enjoyed a front-row seat to the shift. In that cohort, he said he had “a bunch of repeat, second-time founders” he’d worked with before — several during their previous YC company.

“It definitely feels like more of a trend now,” Epstein said. Still, he’s careful not to overstate the novelty.

“It’s not a new thing. But the alumni base of past YC founders continues to grow,” he said in an interview with ������Ƶ News, and that naturally translates into more people eligible to come back.

Epstein has worked with more than 1,000 startups at YC. Before that, he was a startup entrepreneur himself, co-founding (YC W10), a marketplace for graphic design assets that he sold to in 2014 before spinning it back out as an independent company in 2017.

Ask him what separates second-time founders from first-timers, and he points to experience using the program itself.

“They know exactly how to get the most out of the advice, network and resources available to them,” he said. “Having been through the startup grind, they get really good at focusing on the signal that matters and cutting out the noise.”

That experience also helps them avoid a specific, costly mistake.

“The biggest mistake I see second-time founders avoid is overhiring or overspending pre-product-market fit,” Epstein said. “The biggest regret of all the successful first-time founders I know is that they hired too many people, moved way slower and didn’t like working at their own companies anymore.”

Leaner teams, powered by AI

That instinct toward leanness shows up in another pattern: Many repeat founders are choosing to start solo the second time around.

“Some of them (repeat participants) are solo founders, but they’re not building alone,” Epstein said. “They already have networks of people they can bring in as founding employees. This helps them move faster, and feels more fun and less lonely.”

He compares this shift to how cloud computing eliminated the need for startups to raise large sums just to pay for servers.

“It wouldn’t surprise me if 10-15 years from now you look back at all the money startups had to raise to hire people and realize that’s not a requirement,” he said.

AI is accelerating that shift, and Epstein sees it pulling former company builders, including himself and YC CEO , back into hands-on product work.

“It’s so easy to get back into it and start building again. And it’s incredibly exciting,” he said. That mix of hard-won product sense and new tooling, he believes, is changing what one person can build alone.

“They actually become the people that can produce at 10x or 100x what a traditional engineer would be able to build,” he said.

As an example, Epstein pointed to , a founder he first worked with on in 2020 who’s now building an AI tool that helps founders manage their projects and automate tasks.

Even so, Epstein believes founders keep coming back for the same core reasons: personalized advice from partners, a community of ambitious peers, access to top investors and alumni, and the urgency of the batch environment.

“The pressure cooker environment of the batch, which pushes them to move even faster, and distribution to thousands of companies within the network,” he said. “It’s extremely hard to replicate those things on your own.”

From Opkit to Sazabi

One of the repeat founders Epstein has worked with is , whom YC has now backed twice.

Callaway’s path to Silicon Valley began almost by accident. As a college sophomore, he skipped a lined-up investment banking internship after reading about a software bootcamp in San Francisco — a decision he calls “probably the single most important” of his life.

From then on, his goal was clear: “I wanted to do my own venture-backed tech startup, and I wanted to do a YC venture-backed tech startup.”

After gaining experience at and fintech , he founded his first company, , in YC’s fully-remote summer 2021 batch. Opkit was a healthcare-fintech startup building insurance verification and revenue-cycle-management software.

“It was, in retrospect, not the right thing for me to be working on, but a really fun and interesting and rewarding first venture,” he said in an interview. Opkit was later acquired by .

That experience shaped his second company, , a name chosen deliberately in contrast to Opkit.

“Opkit wasn’t very personal to me. It was more of an MBA case study approach to starting a business,” he said. “With Sazabi, it needs to really be in alignment with who I am and my passions and interests.”

Sazabi, an AI-native observability platform competing with incumbents like , draws directly on work Callaway has done throughout his career — a return, in his words, to “what I know best.” He sees it as part of a common pattern: First-time founders often avoid building in the field they know best, then return to it with their second company.

Callaway hadn’t originally planned to go through YC again, and the reconnection happened almost by chance through an email that looped in his former partner on Opkit, Epstein. Once Callaway decided to return, he was more strategic about timing, even deferring his batch to build out more of the product first.

“I wanted to use YC as a go-to-market acceleration event,” he said, something he likely wouldn’t have known to do without having gone through the program before.

The founder was back at YC in person for the first time this spring. He described the second-time experience as something entirely new: “It was really something special.”

This time around, Callaway also noticed a more experienced cohort than his own first batch, along with new concerns specific to the AI era. “There’s a lot of anxiety around what the durable moat is in an AI world when lines of code are effectively free,” he said.

On fundraising, he drew a pointed comparison to 2021. “Spring 2026 felt similar to fall 2021,” he said, “but unlike 2021, where interest rates and ZIRP drove a lot of that energy, in 2026 it’s driven by AI and by real material gains.”

The company’s thesis is resonating with investors. In late June, Sazabi announced an $8 million seed round led by , and Y Combinator, with participation from and more than 60 angels from companies including , and .

“AI has changed how software gets written. Now it is changing how software gets operated,” Callaway said. “Sazabi is rebuilding observability from first principles for a world where agents are part of every engineering team.”

Overall, as AI continues to lower technical barriers and YC’s alumni pool keeps growing, second-time founders like Callaway are becoming an increasingly visible part of the accelerator’s lineup.

Illustration:

In November 2022, was doing something rare for a longtime educator: taking time off. Having launched his career as a teacher in Atlanta Public Schools, Khan became an assistant principal before founding his own public high school in Denver. After a year spent coaching principals at the district office, he decided to take a “personal sabbatical.”

Then, ChatGPT came out.

Khan began tinkering with the new technology, fascinated by its potential.

“I actually went out to my old school building, the one that I founded, and started using it with teachers,” Khan recalls. He ran workshops and asked the educators to use the tool in as many scenarios as possible.

The responses were varied. Most teachers barely touched it. A few tried, but felt doing the work manually was faster. However, some had a lightbulb moment.

“There were one or two teachers who told me, ‘This has completely revolutionized the way I teach,’ ” Khan said in an interview with ������Ƶ News.

Seeing that divide sparked something in him.

“I thought this technology could impact every teacher, not just teachers who are really enthusiastic about using new technologies,” Khan said. “So the task then was, ‘How can we take all the power of this new technology and make it really accessible to teachers?’ ”

Building the ‘vertical AI’ for K-12

That experiment set the groundwork for , a platform designed as an all-in-one AI operating system for K-12 educators and students. For teachers, the tool acts as a daily assistant. It performs tasks like building rubrics, differentiating assignments for varied learning levels, and generating practice worksheets and reading materials.

The platform also helps educators offer monitored AI experiences directly to students. Those experiences range from algebra tutors to writing assistants customized with state exam rubrics that deliver tailored feedback to help students revise their essays.

“You can think of MagicSchool as the vertical AI solution for K-12 schools,” Khan said. “Enterprises are adopting generative AI in other fields … Like in law, there’s and that are vertical AI for legal firms. We’re kind of that, but for K-12 schools.”

Today, the company’s primary customers are school districts that want to provide a safe, governed environment for generative AI that aligns with data privacy rules and local curriculum priorities. MagicSchool now partners with large school systems, including Denver Public Schools, and Florida’s Broward County Schools and Hillsborough County Schools, as well as private institutions.

“One in five children in America go to a school that is in partnership with MagicSchool,” Khan noted. Additionally, roughly 8 million educators worldwide have signed up for the platform, he said.

The uphill battle to raise capital

Despite the platform’s rapid adoption, Khan’s path to raising capital for MagicSchool was a challenge. In the beginning, he worked with hourly contractors and lacked a formal business model.

“I had no real business plan,” Khan said. “The most successful tech companies from my perspective as a consumer were the ones that just got a lot of users, and that was my goal — I was like ‘let’s just get a lot of people using this, and we’ll figure it out from there.’ ”

Once MagicSchool’s user base neared 1 million, Khan began pitching venture capitalists. However, when compared to standard Silicon Valley profiles, his background as an educator initially proved to be a hurdle rather than a selling point.

“Nobody wanted to give a former principal money,” Khan said, recalling “quite literally hundreds of meetings” before securing an institutional investor.

“I think that investors are taught to pattern match,” he noted. “They’re saying, ‘Hey, well, did you go to ? Are you a tech person? Did you work at ? ‘… I have none of those things on my resume.”

Even edtech-focused investors were hesitant, leaving Khan frustrated as he watched other founders secure millions based purely on tech-heavy resumes.

“I remember seeing other edtech companies right around our size raise seed rounds … with no product, no sales, no nothing,”�� he recalled. “ I would think, ‘People know what our product is. Millions of teachers know what our product is, and nobody’s heard of that one.’ ”

To overcome the skepticism, Khan relied strictly on impressive growth metrics, convincing investors during every fundraising stage that the business’ “traction was undeniable.”

The strategy paid off. Following early angel investor checks, MagicSchool went on to raise a $2.4 million seed round led by Colorado-based . To date, the company has raised nearly $63 million in total funding, driven by strong financial growth, including 3x year-over-year revenue growth at the end of last year, according to Khan. The startup’s other backers include , , and.

Expertise as the next wave of innovation

Although Khan no longer manages the high school he founded, he stays connected to the classroom through district visits. The school remains a top-performing public school in Denver under a former founding team member, he noted.

“Of course, I miss that,” Khan admits. “There’s nothing that can replace the relationship you build over a long period of time with students.”

Yet, his deep-rooted experience in education ultimately became MagicSchool’s greatest asset — a trend Khan sees taking hold across the broader AI landscape as domain experts step up to build industry-specific tools.

“I think that what we’ve learned over time is that the model is no longer the differentiator,” Khan says. “How you contextualize the model with the real problems that people have in the work that they’re doing, and specific expertise, is the thing that’s going to unlock the next wave of innovation and impact for generative AI.”

Related ������Ƶ query:

Illustration:

Menlo Ventures XVII will invest primarily in seed and Series A companies, while Menlo Inflection IV will provide growth capital to startups at Series B and beyond. The new funds will target companies throughout the AI market, from foundational models and infrastructure to enterprise, healthcare and consumer applications.

The new capital gives the Silicon Valley firm more flexibility to back companies from their earliest days through later funding rounds that can require hundreds of millions of dollars. It also shows how important AI has become to a firm previously known for investments in companies including , and .

In recent years, has become the most prominent company in Menlo’s AI portfolio. The firm first invested in the AI model developer in 2023 and has added to its investment in later rounds. Menlo’s other AI investments include app-building platform , music-generation startup , AI model marketplace , voice productivity company , AI infrastructure companies and , robotics startup , and AI research company .

, a partner at Menlo since 2015, has played a central role in developing that strategy. He invests across AI infrastructure, developer tools and AI-native software and has led Menlo’s investments in companies including Anthropic, Lovable, OpenRouter, AI-powered software delivery platform , code security startup and legaltech startup .��

Before joining Menlo, Murphy spent 15 years as a general partner at Kleiner Perkins, where he was an observer at Google from the firm’s initial investment through its IPO, helped launch the $200 million iFund with Apple and worked on investments including DocuSign, AppDynamics, Upstart and Shazam. Earlier in his career, he held operating roles at Netboost and Sun Microsystems.

������Ƶ News spoke with Murphy about why AI is pushing Menlo toward larger and more concentrated investments, what the firm has learned from its relationship with Anthropic, and where he sees the next opportunities — as well as potential bottlenecks — across the AI market.

The interview has been edited for brevity and clarity.

������Ƶ News: Inflection IV puts Menlo in competition with some of the biggest late-stage investors in the world. How do you keep the firm’s close, founder-focused approach when you’re writing much larger checks?

Murphy: AI companies need more capital than previous generations of software companies. They’re staying private for longer, and the winners are quicker to break from the pack.
For us, a larger fund gives us the ability to partner with founders from company formation through hypergrowth. Through our venture fund, we invest in seed and Series A companies, but the inflection fund gives us the scale and flexibility to back the clear winners as they emerge.��

This was our strategy with Anthropic, Suno, Wispr, OpenRouter and Lovable.��

You’ve recently invested $100 million in companies including Lovable and Suno. Is that level of concentration becoming a bigger part of Menlo’s strategy, or is it reserved for a small number of standout AI companies?

The Anthropic investment is an example of us doubling down when we had incredible conviction. Remember, we first invested in the [Series] C round, which gave us a chance to get close to the team, see how well they were executing, and understand where they were going.

When we led the [Series] D round, it was still the largest investment the firm had ever made. We learned from that experience and success, and it’s become a standard part of our approach now. Also and importantly, the market has changed.��

There’s a gold rush around later-stage AI, and the companies that break out are growing at rates we’ve never seen before, at scale. These companies need capital to sustain that growth and, frankly, have earned higher private valuations given the growth rate.��

We’re changing how we invest, but overall we’re pursuing more of a barbell right now. On the later end, we’re much more aggressive, stage- and capital-wise, for the right companies.��

That said, the bar is still very high. Many AI categories are overfunded, and there is a huge amount of speculation. The winners of this era separate quickly, and we believe they will compound at unprecedented rates.��

Your relationship with and Anthropic gave Menlo an early view into where the AI market was heading. What are you seeing now that you think other investors may still be missing?

I don’t know that it’s counterintuitive, but I’d say we are moving from Phase 1 to Phase 2 of the market and are seeing an entirely different set of opportunities and challenges.��

In Phase 1, developers just picked a model to start building AI. In Phase 2, we are seeing companies get to scale using AI and looking to optimize their spend and infra choices. A whole host of companies are seeing tailwinds alongside Claude and Claude Code, such as OpenRouter, Fireworks, Modal and .��

It will be a multi-model world. One size won’t fit all, and we’ve been active in that area as well, including more vertical models such as for life sciences and for robotics.��

The Anthology Fund has helped you spot promising AI companies early. As the application layer matures, what specific bottlenecks are you seeing founders run into when building enterprise-grade defensibility on top of frontier models?

The Anthology Fund has been an incredible source of deal flow and has given us a broad aperture around what areas of AI are disproportionately taking off. It’s been a great program for getting closer to a broad set of application and infrastructure companies and building relationships before deciding where to lean in.��

I wouldn’t say it’s been the key factor in identifying bottlenecks across the AI ecosystem. For sure it is part of it, but from the broad set of portfolio companies and new companies we meet, the No. 1 bottleneck has been how to take all the new code that has been written and get it into production faster, safely, and securely.��

This has created a big tailwind for companies helping with software delivery, like Harness with application and code security, like Semgrep; and code review and testing like .��

Additionally, the rise of custom models based on open-source/open-weight models has created a number of bottlenecks as companies look for compute, training, sandboxes, and more. Both development and runtime resources have become essential to accommodate this next wave, and companies like Modal and Fireworks are addressing that with their offerings and the compute capacity they’ve been able to aggregate across various compute providers, including Nebius and CoreWeave.

Valuations across the AI market have risen dramatically. Which parts of the market do you think are most likely to produce strong, sustainable businesses: infrastructure,��model tools or industry-specific applications?

We’ve been active across models, infrastructure, and applications. All are showing tremendous potential and tailwinds right now. At the moment, infrastructure is seeing a disproportionate spike in opportunities as enterprises and AI-native companies embrace a multi-model approach and scramble to keep up with the compute and infrastructure management needs that it requires. Coding tools are now mainstream and putting tremendous pressure on organizational processes to release software faster and more efficiently, which is leading to tailwinds for companies like Harness and Gimlet.

It’s fair to say the majority of companies are optimizing for market share right now rather than gross margin, but there are many opportunities for margin improvement over time, and this is a rare land-grab moment.

You’ve backed new AI research labs before they even have a product, including . At that stage, what convinces you that a team has something truly different, and that it can compete with much larger technology companies?

As I mentioned, we believe in a multi-model world, where one size won’t fit all needs and use cases. We have an explicit strategy to gain early exposure to some of the most compelling AI research teams with distinctive techniques or capabilities, even at a very early stage. Many of these companies are raising $100 million-plus rounds, and while we occasionally lead in this category, we prefer to write smaller checks initially. This helps us build broader exposure across the category and talent pool, and then double down once we see one really taking off.

Frankly, there are too many right now, and all claim some differentiated technique or team. Of the roughly 60 model companies, we believe we’ve invested in more than five of the best and expect to lean into one or two of them as they ramp.

Related ������Ƶ queries:

Related reading:

• Anthropic Backer Menlo Ventures Raises $3B In New Funds To Back AI Startups Across Stages��

Illustration: