That’s up 49% from the $3.5 billion valuation it achieved when announcing its $300 million Series C — which included primary and secondary funding — in March of 2025. The latest capital infusion brings San Francisco-based Mercury’s total primary and secondary funding to approximately $700 million since its 2017 inception.

led the latest financing, which included participation from returning backers , , , , and .

Mercury counts more than 300,000 companies as customers, including startups and larger entities such as , , , , and .

Interestingly, Mercury recently received from the banking regulator, the OCC, to establish its own bank. This is in contrast to many fintechs, which typically work with a sponsor bank but are not banks themselves.

The company hit $650 million in annualized revenue as of the 2025 third quarter, and claims to have achieved four consecutive years of profitability on both a GAAP net income and EBITDA basis.

AI’s effects

“AI is collapsing the friction between an idea and a company faster than anything I have seen in my career,” , co-founder and CEO of Mercury, said in a press release. “We are going to see more founders in the next five years than in the last twenty. But legacy banking in 2026 still works the way it did when I started my first company in 2006. I started Mercury because banking should do more than be a vault, it should help customers run the best business possible.”

Fintech startups, particularly those that apply AI to traditionally manual or burdensome processes, have benefited from increased investment in recent quarters. Global funding to VC-backed financial technology startups totaled $53.8 billion in 2025, per �������� . That’s a more than 29% increase from 2024’s total of $41.6 billion raised.

Disclosure: The author of this article is a freelance writer who also writes for Mercury’s independent magazine, Meridian.

Related �������� query:

Related reading:

• Fintech Startups Globally Raise More Money In Far Fewer Deals In Q1 2026

Illustration:

When and his wife were navigating the complex world of fertility treatment, the process was marked by stress and financial strain. But after finally achieving a successful outcome, AlSalim recognized how different his experience was from many others in his position.

Despite the ordeal, he noted, having a child afterward “is much better than a load of people who don’t have anything to show for it.”

The experience sparked a business idea to help others in the same situation his wife and he were in. “My son was 1 week old,” he said, recalling the exact moment the concept took hold.

AlSalim officially registered the company name, , in 2019, but the business’ true inception came later, as the founder refined the idea and sought investors.

Gaia is working on building what AlSalim believes is a fundamentally new category in the . The company uses artificial intelligence and machine learning — trained on millions of anonymized historical data points and fertility outcomes — to better understand risk and probability for fertility treatment.

The platform analyzes variables such as age, hormone levels, ovarian response, treatment protocols, embryo development and clinical outcomes to direct patients to “optimal” clinics based on their data profiles, and to generate personalized forecasts around fertility success. It also uses AI and machine learning to underwrite personalized outcome-based “flexible” financing plans for IVF, egg freezing and embryo transfer procedures.

“We tell you where to go, we protect your path, we finance your treatment, we support you,” AlSalim said in an interview with �������� News. “No one else today bundles care, capital and financial protection into a single product.”

And today, the New York-based startup — led by AlSalim as its sole founder — tells �������� News exclusively that it has secured a $100 million debt facility from to scale its operations across the United States.

The credit facility follows a $14 million Series A round raised in January 2025, led by , that brought Gaia’s total equity funding to $37 million across three rounds. Other backers include and .

Fertility remains a relatively niche area for healthcare startup investment. Last year, venture investors put $194.8 million toward startups in ��������’s fertility categories. Since the peak year of 2021, when $229.6 million went to fertility-related startups globally, annual investment in the sector has ranged between about $100 million and roughly $200 million, .

Treatment with ‘outcome protections built in’

Today, the fertility industry operates almost entirely on a “fee-for-service” model. Patients pay thousands of dollars per individual procedure, regardless of whether that procedure actually results in a baby. If a cycle fails, the patient is left with heartbreak and a depleted bank account.

Gaia flips this dynamic on its head by pricing the probability of success rather than the number of procedures, its founder said.

“We are not just a financing company,” AlSalim told �������� News. “We use data in order to create unique plans that are individualized with outcome protections built in.”

For an IVF cycle, which has a nationwide median cost of $22,000, Gaia says it offers complete predictability. If a member’s first IVF cycle fails, Gaia covers the next cycle at no extra cost. For embryo transfers, the plan includes unlimited transfers until a live birth is achieved.

The model works across other endpoints, too. For example, if a 30-year-old woman wants to freeze her eggs, Gaia uses its predictive engine to guarantee a target number of retrieved eggs based on her specific biomarkers. If she does not hit that number in the first round, Gaia funds a second cycle at no extra cost. Patients can choose to pay the fixed cost upfront or use Gaia’s financing to spread the cost over five years with monthly payments.

Closed-loop model

By owning the data and the risk from initial consultation to live birth, Gaia aims to build a closed-loop data asset that it believes will serve as a massive competitive moat.

Its model is resonating. Over the past 15 to 16 months, Gaia has experienced a significant growth inflection, according to AlSalim. The company has surpassed 1,100 memberships, with over 1,000 active members in the U.S., and has partnered with 200 clinic locations across 40 states.

The founder declined to provide hard revenue figures when asked about growth, saying that the company is “now developing a baby every 18 hours” while maintaining a of 85, which is considered “exceptional” in the healthcare industry by , creator of the customer loyalty benchmark.

Building a village

To sustain this velocity, Gaia has expanded its distribution channels beyond direct-to-consumer marketing to include local partnerships with acupuncturists and pharmaceutical companies, as well as direct clinic integrations.

Last year, the company launched an enterprise benefit product, marketing and selling directly to employers who want to offer comprehensive, risk-insulated fertility coverage to their workforce.

The corporate product has scaled rapidly, said AlSalim. Gaia’s enterprise client roster spans diverse sectors — from tech professionals in Silicon Valley to blue-collar manufacturing workers in Denver.

, managing director and head of U.S. Investments at Viola Credit, said his firm was drawn to Gaia because it believes the startup is addressing “a deeply important and underserved problem” with a model that is “both commercially compelling and mission-driven.”

Chen believes that Gaia stands out also because it is not “simply a financing product.”

Its approach, he said, “aligns incentives across patients, clinics, and financing in a way that feels genuinely differentiated,” he wrote via e-mail, “and we believe it can meaningfully improve access to fertility care.”

Related �������� query:

Related reading:

• VC Investment Soars For Fertility Services Even As Birthrate Dips

Illustration:

Just through mid-May, China-based robotics companies this year have raised $5.6 billion across 176 deals, �������� data shows. That sum matches total investment to the nation’s robotics companies in all of 2021, the peak of the funding cycle. Investment in the sector has also already eclipsed the $4.3 billion raised by China-based robotics companies in all of 2025.

Startup funding in Asia overall surged to $27.4 billion in Q1, its highest level in over three years, with China capturing $16.5 billion — 60% — of that total, according to recent �������� data. Robotics contributed meaningfully to that $16.5 billion total, with startups in the sector raising $3.3 billion across 126 deals.

Embodied AI boom

A review of �������� data shows that investors now are no longer funding mostly pre-programmed hardware, but increasingly backing China-based startups working on embodied AI —��or artificial intelligence with a physical body that interacts with the real world in real time.

That shift toward artificial intelligence-driven robotics mirrors a global surge in investment into robotics and other physical AI startups. It’s also thanks to the rise of advanced, open-source reasoning models that have fundamentally changed how robots operate. Startups are moving away from coding robots line-by-line toward Vision-Language-Action models that allow physical machines to observe, reason and execute physical tasks end-to-end.

In China, robotics startups at the intersection of the software and hardware integration are drawing the largest checks in the space and often back-to-back funding rounds. They include:

• , a 1-year-old humanoid robotics company that integrates embodied intelligence that last month raised a massive $513 million seed round led by and . The Shanghai-based company was valued at $1.9 billion.
• , which develops robotic systems and automation solutions for industrial and service applications, closed a $140 million Series A extension round in January from investors including . Then just three months later, it raised $293 million in a massive Series B round co-led by and
• In February, Beijing-based , which says it’s building a “universal brain” for robots, raised a $290 million Series A led by and . The 2-year-old company was valued at $1.5 billion. Then in April, it announced a $145 million Series A extension financing, bringing the total round to $435 million.
• Humanoid robotics company in February raised a $145 million Series B led by . The 2-year-old China-based company was valued at $1.4 billion. In April, it announced a $290 million extension to that round, bringing its total to $435 million
• Shenzhen-based , a builder of humanoid and quadruped robots, raised a $200 million Series B last month led by and . The 2-year-old company’s robots will be deployed for traffic, security and retail. It was valued at $1.5 billion.

Top investors

�������� data shows the most active investors in the space are largely Asia-based. The busiest this year has been Hong Kong-based , taking part in six deals, including a $200 million round last month for humanoid robotics and embodied intelligence developer .

Among lead or co-lead investors, three China-based firms — , and — have each taken part this year in deals totaling $290 million or more.

Exits gain steam

Venture investors are likely feeling confident as the sector notches notable liquidity events, including IPOs and acquisitions.

The of , targeting a $3 billion to $7 billion valuation, is a milestone for the industry. The company in March filed for an to list on the , and its IPO would likely spur other startups in the space to pursue their own public-market debuts.

The sector has already seen some notable exits.

They include Hong Kong-based ,�� a Shanghai-based startup that makes lightweight industrial robots. The company on May 18 listed on the , raising about $86 million. And it did not disappoint. Robotphoenix closed its first full day of trading at HK$53.75 ($6.86 U.S.), up nearly 80%. (Interestingly, Chinese robotics firms as their primary liquidity hub.)

On the M&A front, in what is widely considered a historic first for China’s embodied artificial intelligence sector, AI robotics unicorn in July 2025 engineered a two-stage consortium takeover to in legacy manufacturer for about $290 million. AgiBot’s co-founder formally stepped in to chair Swancor, effectively turning the publicly traded shell into a direct extension of AgiBot.

Ultimately, it seems that 2026 is the year China’s robotics companies are pivoting from raising early venture rounds to mass production, as a domestic market that currently accounts for more than 43% of global robotics venture investment, per ��������.

Related �������� queries:

Related reading:

• China Leads Asia’s Startup Funding To Its Highest Level In More Than 3 Years
• Amid Record Robotics Funding, Apptronik Raises $520M Series A Extension To Boost Production Of Humanoid Robot Apollo

Illustration:

Existing backer (KIP) led the financing, which is described as a bridge round. (KBIC) also participated. The raise brings Santa Fe Springs, California-based Xpanner’s total funding to $38 million since its 2020 inception.

Xpanner turns construction equipment that customers already own into automated assets “without replacing a single machine,” according to , the company’s co-founder and CEO.��

“There’s no upfront investment, no rip-and-replace,” Lee added. “Like a smartphone gaining new capabilities through app updates, customers expand their automation through simple software updates.”

The benefits for its customers include significant cost savings and shorter project durations, according to Lee.

Originally founded in South Korea, Xpanner moved its headquarters to the U.S. in 2023. Today, , , and are among its customers.��

Notably, all of Xpanner’s co-founders have deep industry experience. Lee spent two decades in executive positions at and , driving unmanned construction projects and corporate venture initiatives in the heavy equipment world. CFO spent over 12 years working in heavy equipment at Bobcat, followed by eight years in venture capital at Korea’s largest commercial bank. CTO led robotics and automation at for 20 years, becoming the first in the industry to commercialize semi-automation features for construction machinery.

Growth and a path to profitability

Xpanner is refreshingly transparent about its financials. The company grew revenue from $3 million in 2023 to $7 million in 2024 to $21 million in 2025, according to Park. It saw $8 million in revenue and $1 million in EBIT (earnings before interest and taxes) in the first quarter of 2026.

The startup is targeting $60 million in ARR by year’s end.

Impressively, the company says it maintains a gross margin above 80%, thanks mostly to its subscription-based AaaS model. It achieved monthly break-even in 2025, and Park said Xpanner is on track for full-year profitability this year.

“Once hardware is deployed, incremental subscription and service revenue flows at near-zero marginal cost,” he said.��

The company plans to use its new capital in part to strengthen its development capabilities by advancing its next-generation physical AI hardware and software platform, deepening its core component engineering, and expanding its data and AI infrastructure.��

Some of its customers are still on a perpetual modular model, which includes the one-time purchase of its X1 Kit hardware paired with its software. Looking ahead, Xpanner expects to be fully on its subscription model by the end of the year.��

The company is also actively expanding into adjacent verticals, including battery energy storage systems (BESS) and AI data center construction.

‘Strong gross margins, near-zero churn’

, managing director at KIP, told �������� News via email that his firm was impressed by Xpanner’s commercial traction and unit economics.

“Strong gross margins, near-zero churn, and rapid account expansion are signals that the value proposition is real and not pilot-driven,” he said.

director at KBIC, believes that most construction automation companies hit a scalability wall because they automate entire machines end-to-end. However, he said that since Xpanner’s task-specific approach scales through software rather than hardware redesign, the company “can expand wallet share inside accounts without proportional cost.”

“That’s a software-economics business operating in a hardware-dominated market, and it’s rare,” he wrote via e-mail.

Physical AI funding smashes records

Xpanner sits at the intersection of two sectors that have received strong interest from investors in recent years.

Startups working on physical AI — real-world applications of artificial intelligence, including technologies such as automated hardware and robotics — have already hauled in more than $37 billion in venture funding globally this year, , shattering the full-year record of $21 billion set in both 2025 and 2021.

At the same time, venture investment in real estate and property-related startups rebounded last year, largely driven by funding to AI-centric companies.

Related �������� queries:

• ��
• ��

Illustration:

As Ayoub describes it, while the medical profession was sold to him as a lucrative and stable career, the reality of modern healthcare hit home during a gap between his fellowship and his full-time role at . Living in New York City and unable to afford rent, he attempted to pick up extra shifts at a local urgent care. Despite a clear workforce shortage and his own need for income, he was told he couldn’t start for 90 to 120 days.

The culprit? A fragmented, manual credentialing process that acts as the industry’s primary bottleneck.

“The bottleneck is not the number of doctors, but the fragmented infrastructure connecting them to where they are needed,” said Ayoub, who also serves as an assistant professor of neurosurgery at the Donald & Barbara Zucker School of Medicine. “Most people assume the issue in healthcare staffing is a lack of doctors, but what we’ve seen is something different.��

There’s a large, underutilized workforce that simply can’t move between systems efficiently,

So in early 2025, Ayoub and began pondering a solution. Their initial ideas eventually turned into , a startup with an AI-powered platform designed to serve as an “automated Dropbox” for physicians.��

Today, the New York City-based startup is announcing it has raised $2.2 million in a pre-seed round led by , �������� News reports exclusively. also participated in the round.

AI-driven healthcare takes off

AI-related healthcare has seen a significant rise in venture funding globally, �������� shows. Investors put an estimated $14.9 billion into seed- through growth-stage funding to companies in AI-powered health tech categories in 2025, per �������� data. That’s up significantly compared to the $8.6 billion raised in all of 2024.

Many of the recently funded healthcare startups are AI-centric, and, like Saile, are focused on streamlining dated processes.��

In the current system, every time a doctor wants to work at a new facility — whether it be a hospital, a surgery center, or a telemedicine platform – he or she must manually resubmit a CV, licenses, and board certifications via email. In applying to an urgent care facility, Ayoub realized that while staffing agencies act as gatekeepers, the underlying infrastructure was broken.��

There was no centralized way for a doctor to maintain a compliant status and share it instantly across different job verticals.

Saile aims to solve that problem by storing and tracking all a doctor’s credentials in one place and providing alerts before documents expire so that a physician can always be compliant. It goes one step further by providing access to a shift marketplace. In a nutshell, the startup serves as a portable credential passport for physicians to be identified and assigned patients at various hospitals.��

The company’s five modular AI agents automate what currently takes months of manual coordination across recruiting, onboarding, credentialing, staffing and compliance.��

By combining credentialing and staffing into a single infrastructure layer, Ayoub says Saile has shortened the onboarding timeline by roughly 45 days, from about 90 to 120 days, and reduced administrative tasks for healthcare facilities by an estimated 40%.

“Other solutions either focus on one piece of the problem or offer staffing tied to a single job type,” Ayoub said in an interview. “Saile owns the entire journey…And facilities get direct access to a pre-vetted pool of local and regional physicians without juggling multiple vendors or paying for the friction in between.”

Investing in the ‘infrastructure layer’

What began as a bootstrapped project fueled by word-of-mouth in a tight-knit clinician community has quickly gained momentum. The app has grown to nearly 5,000 active user physicians nationwide. Operating with a lean core team of four, the company plans to use its new capital to expand its AI agent infrastructure, grow its marketplace capabilities, and deepen integrations with facility credentialing systems.

Saile has four core revenue streams, with a primary focus on a per-seat SaaS model for facilities. The approach is to offer facilities access to the pool of physicians, and then charge on a per-seat usage basis for the workflow and credentialing infrastructure that supports it.

, founder and partner at Matchstick Ventures, said his firm was drawn to the founder market fit it saw in Saile.

“Marc had felt the pain of this problem and actually had built this more or less for himself out the gate,” Brosher said in an interview with �������� News. “We love those combos where founders aren’t just randomly seeking out a solution to make a buck. This was very much a personal thing for him in the problem that he was solving.”

The firm also saw a “big” market opportunity in offering an “all-in-one” solution for doctors looking to pick up side jobs.

“People have tried to go after this a few different ways. They’ve either gone after credentialing, or they are a staffing agency,” Brosher added. “And when we look at this market, we feel like there needs to be disruption here…Ultimately, Saile is building the infrastructure layer beneath staffing. We feel like having that all-in-one infrastructure layer is actually where the real value is to be had.”

Related �������� queries:

Related reading:

• �������� Sector Snapshot: Funding To AI-Related Healthcare Startups Is Robust This Year

Illustration:

]]>

But while the fundamental need for food security remains high, startup investors have become increasingly risk-averse, driven by structural and economic pressures, �������� data shows.

And that aversion is reflected in the number of venture deals being conducted in the sectors so far in 2026.

The broad trend: Venture funding in agriculture and farming is still undergoing a market correction following the hyper-funding era of 2021, when startups in the space raised $10.5 billion across 1,419 deals, a review of �������� data shows. Many agtech startups that raised massive rounds during that 2021 peak have struggled to show real-world traction.

At the same time, as with most other industries, AI is having an impact. AI in agriculture has moved from predictive to agentic. The focus is no longer strictly on collecting data, but on autonomous orchestration — closing the loop between digital insight and physical action on the farm.

The numbers: So far this year through May 7, agtech startups have raised $1.4 billion, on track to be down slightly or on par with the $4.4 billion raised last year and the $4.6 billion raised in 2024. It’s also well below the peak $10.5 billion raised by agriculture-related startups in 2021 and the $10.3 billion that went to such companies in 2022.��

Deal count looks to be pacing down more significantly, with the 187 deals done year to date tracking much slower than the 784 rounds in 2025 and the 1,038 in 2024.That signals not only less overall funding but also larger round sizes for fewer companies.

Standout deals

The largest round in 2026 so far was raised by , a New Zealand startup that develops a smart collar for cattle, which enables virtual fencing and real-time monitoring. In late March, Halter raised funding led by at a valuation of just over $2 billion.

Boston-based , a weather technology company offering real-time forecasting services to predict and respond to climate-related threats, closed on at a $1 billion valuation in February. Private equity firm and co-led that financing.

������Գ���’s , which has built an amphibious water bomber aircraft for aerial wildfire suppression, raised a $135.2 million Series A round led by in March.

And , a UK-based producer of gene-edited crops, raised $105 million in a Series C round of funding co-led by and in March.

Interestingly, Indian startups have raised three of the 11 largest deals in the sector so far in 2026: , and .

Exits

The agtech exit landscape over the past year has been characterized by strategic consolidation rather than high-profile IPOs. As venture capital interest remains disciplined, established players – both within the agricultural sector and from broader tech – are acquiring startups to bolster their AI and automation capabilities.

The third quarter of 2025 saw three notable acquisitions in the space:

• picked up (founded in 1982, the company is manufacturer of driverless orchard sprayer for agriculture applications).
• announced it would buy , which aims to offer high-volume crop production units made from upcycled shipping containers to support farming in any climate, ��and
• said it would acquire (formerly Agritask), a crop supply intelligence company.

In the first half of 2026, the trend continued.��

In January, announced it would acquire biological insect control group . Interestingly, the company said it plans to.

Also in January, cannabis industry enterprise resource planning (ERP) platform announced it had acquired competitor , an 8-year-old cultivation management software startup.

(FBN), , and have long been viewed as IPO candidates but it remains to be seen when (or if) they’ll take the plunge.

Related �������� query:

Related reading:

• Automation Is A Bright Spot Amid Ho-Hum Agtech Funding

Illustration:

led the financing, which included participation from (Google’s early-stage AI fund), , , , and several angel investors. The raise brings Fazeshift’s total raised to $22 million since its 2023 inception.��

The San Francisco company was founded by a team with an unconventional pedigree: (CEO), a former consultant and mechanical engineer and (CTO), an -trained nuclear submarine officer.��

The two met at , but their lightbulb moment came while running a previous startup, , where they found themselves color-coding spreadsheets to track payments for just 10 customers and realized that the tools they were using failed to solve the basic problem of ensuring money actually hits the bank.

They realized that while there are more than a million accounts receivable (AR) clerks in the U.S. alone, many of them spend their time bouncing between systems such as , CRMs like , bank portals, and email threads because these systems do not natively talk to each other.��

Unlike accounts payable, which a company can standardize internally, Leksana contends, accounts receivable is a “snowflake” problem that remains one of the least automated functions in finance. Every customer has a unique set of requirements; for instance, a large retailer might demand that an invoice be submitted through a specific proprietary portal with Part A and Part B attached as PDFs.��

Fazeshift claims that it can automate more than 90% of manual AR tasks — from invoicing and collections to payment matching and reconciliation — by operating on top of existing systems and executing workflows across them. It essentially sits on top of a company’s current stack as a “brain.”

Competitors, according to Leksana, are generally focused on automating tasks, while Fazeshift is working on building what she described as an “intelligent control layer” that helps companies “collect faster, more predictably and with less effort, and that is continuously improving through proprietary payer behavior data.”

“What sets us apart is our ability to handle complex workflows that other tools fail to solve – especially in industries like wholesale, construction, staffing, and HVAC, where AR processes are highly fragmented and manual,” Leksana told �������� News in an interview.

An OS for the finance organization

After launching at the start of the Summer 2024 Y Combinator cohort, Fazeshift has seen its revenue grow 12x in a single year, attracting dozens of enterprise customers, including eight unicorns and its first public company, according to Leksana.

Customers include , , , and , as well as one of the largest independent wholesale distributors in the Southeast, the world’s top e-commerce aggregator, and a leader in music publishing, per Leksana.��

Looking ahead, Leksana believes that Fazeshift has the potential to expand beyond accounts receivables. The goal is for Fazeshift to become the primary operating system for the entire finance organization.

“Our long-term vision is to expand into a broader CFO suite,” she said, “building toward a future of autonomous finance where core operational work is executed by AI and human teams can focus on agent management, strategic work, and governance.”

Broken workflows for ‘critical functions’

, partner at F-Prime Capital, said her firm was impressed by Fazeshift’s efforts to meet the needs of companies still running AR mostly on spreadsheets and email.

“You’d be surprised how many Fortune 500 companies only started adopting software a few years ago and still have dozens, if not hundreds, of AR clerks on staff,” she wrote via email. “That gap between how critical the function is and how broken the workflows remain is exactly the kind of opportunity we look for.”

Wu also believes the market is at an inflection point where AI is moving from co-pilot to co-worker, and human teams are shifting from doing the work to reviewing and managing AI agents.

“Fazeshift is bringing us closer to an autonomous future for finance,” she said. The founders had “lived the pain of broken AR workflows firsthand at their last company and set out to build the platform they wished they’d had. When you meet founders like that, you move fast.”

Fintech startups, particularly those that apply AI to traditionally manual or burdensome processes, have benefited from increased investment in recent quarters. Global funding to VC-backed financial technology startups totaled $53.8 billion in 2025, per �������� . That’s a more than 29% increase from 2024’s total of $41.6 billion raised.

Related reading:

• Fintech Startups Globally Raise More Money In Far Fewer Deals In Q1 2026

Related �������� query:

• ��

Illustration:

]]>

With that in mind, its venture arm, , is backing startups that build the financial and technical infrastructure for a more autonomous economy.��

�������� News recently conducted an email interview with , managing director of , about the firm’s investment thesis, the kinds of startups it aims to back, and how it works with founders to build and scale projects in the American Express ecosystem.��

Amex Ventures’ recent investments reflect its focus on an autonomous future. In April, it led a round into business identity infrastructure platform and also wrote a check into , an agentic marketing platform that raised a $43 million Series B. It also backed Candex, a startup that uses AI and aims to help large companies pay small, one-time, or irregular vendors without the administrative headache or risk that comes with onboarding them.��

Since joining Amex 15 years ago, Tsang has been responsible for identifying and executing strategic investments in early- to growth-stage startups that fit the American Express thesis. He also leads the firm’s consumer services investment vertical, focusing on the future of membership, and oversees the firm’s global portfolio management operations.

Before joining American Express, he was part of the diversified industrials investment banking group at , where he focused on M&A and corporate finance transactions.

The interview has been edited for brevity and clarity.

�������� News: In the “2025 GenAI era,” we focused on tools that inform decisions. Now that we’re in the ‘2026 Agentic era,’ where agents execute transactions, how has the bar shifted for founders pitching Amex Ventures? Are you prioritizing the AI model or the trust/identity layer that will allow an agent to use a Platinum card autonomously?

Tsang: The bar for Amex Ventures investment has shifted toward agentic commerce systems that can navigate complexity and help facilitate end-to-end workflows for customers, with appropriate user authorization and controls. The most compelling founders operate companies that can handle much more of the full commerce journey, not just surfacing options, but incorporating context, executing user decisions, and ultimately completing tasks, with personalization increasingly becoming a key differentiator in how those experiences are delivered and retained.

Agentic commerce systems are sometimes framed too narrowly as just the final execution step, but the real opportunity — and where we are seeing the most innovation — is in systems that can understand preferences, constraints, and intent, and then orchestrate a complete experience that properly reflects them.��

Over time, this could evolve to support more comprehensive experiences, such as planning entire trips rather than just recommending a flight, or managing a series of related actions instead of a single step. AI also has the potential to help scale the kind of high-touch, tailored experiences that were once reserved for a small set of customers to a much broader base.

From our perspective, we are less focused on any single layer, whether that is the underlying model or the trust and identity infrastructure, and more focused on how these components come together to deliver a seamless and high-quality customer experience.��

Trust and security are foundational, especially in financial services. Ultimately, we are looking for founders who are building with the ambition and technical depth to address real-world complexity, while delivering meaningful outcomes for customers with a focus on security and compliance.��

Many founders struggle with the “CVC paradox” — getting a pilot with Amex is a huge win, but scaling it globally can take years. What is the most effective way a founder can leverage your firm to move from a localized experiment to a core membership benefit?

We see the most successful commercial partnerships follow a crawl, walk, run model — starting with a focused pilot, learning quickly, and then building toward broader integration and scale.

Our investment model creates shared incentives to build commercial partnerships that benefit both American Express and our portfolio companies, and we help advance these relationships over time. We look for founders who continue to iterate and innovate so that the scope of these partnerships can expand in a meaningful and sustainable way.

For example, we recently invested in the startup , an agentic marketing platform that helps enterprises understand, influence, and measure how brands are represented across AI-powered channels. At American Express, we are utilizing Bluefish’s technology to support our agentic search efforts, including our Answer Engine Optimization (AEO) strategy.

For travel and dining startups, customer acquisition cost (CAC) is skyrocketing due to influencer-driven marketing. How much does a startup’s ability to plug into the Amex ‘closed-loop’ ecosystem factor into your valuation, versus their organic growth?

When we evaluate companies, we start with the fundamentals. We are looking for businesses that can stand on their own, with strong products, clear value propositions, and sustainable growth models. That is core to any investment decision we make.

At the same time, one of the advantages of an investment from Amex Ventures is the potential to partner with American Express and engage with our ecosystem, which we see as a meaningful opportunity for many companies. When there is a clear path to creating mutual value through a commercial partnership, it can strengthen our conviction to invest.

That said, we do not invest simply because a company is likely to have a commercial relationship with American Express. To invest, we must believe in the underlying business on its own merits.�� The potential for an American Express partnership represents additional upside and an opportunity to accelerate the company’s growth together.

With tech M&A volumes up significantly this year, are you managing the current portfolio with an eye toward strategic buyouts by Amex’s parent company, or are you pushing for independent IPO-ready unit economics given the partially reopened window?

We are not managing the portfolio with a specific exit outcome in mind. Our focus is on investing in companies that can build strong, durable businesses and on creating strategic relationships with our portfolio companies that drive value on both sides.

In many cases, that means working closely with our portfolio companies to explore potential commercial relationships with American Express, which is core to our role as a corporate venture investor. In fact, nearly two-thirds of our portfolio companies have had a commercial relationship with American Express.

From there, a company’s path, whether toward an IPO or M&A, is ultimately driven by the founder and the broader investor group. Our role is not to steer that outcome, but to support the company in building long-term value and maintaining the flexibility to pursue the right exit opportunity.

When you look at the current “service-as-software” startups, do you see them as long-term standalone companies, or are we looking at a massive consolidation cycle where travel/dining tech eventually gets absorbed back into the major financial rails?

There are a number of ways this could play out, and it will likely vary by industry. In some sectors, particularly those with high levels of specialization or unique requirements such as regulation, there is a clearer path for companies to build durable, standalone businesses that deliver meaningful value over the long term.

At the same time, we are starting to see early signals that consolidation could emerge in certain areas, particularly at more horizontal layers of the AI stack, where capabilities can extend naturally across multiple use cases. It is still early, though, and the boundaries of how far that consolidation will go are being defined.

In lifestyle categories like travel and dining, I believe the outcome will be a mix. For example, in travel, there is already a range of booking options available to consumers. While there has been some consolidation amongst these companies, many have continued to find ways to differentiate and add value.��

We expect a similar dynamic to continue, with some companies scaling independently while others partner more deeply or join broader platforms over time.

Related �������� query:

• ��

Illustration:

Cambridge, Massachusetts-based Blitzy has now raised more than $204.4 million. led its latest financing. New investors including , and also participated, as did existing backers such as and . The company said it also received strategic investments from , and .��

Blitzy claims that its platform autonomously completes months of software development, including automated testing and quality validation. It further claims to increase engineering velocity by 5x “for some of the world’s largest enterprises.”��

While Blitzy did not reveal hard revenue figures, it said that its technology has been adopted by “dozens” of Global 2000 enterprises across 10 industries, including State Street and QAD. The company says it was built on the premise that frontier models alone would not solve enterprise software development.

“We believed that delivering production-ready code for the enterprise would come from fusing hyperscaled agent orchestration and a system that deeply understands the legacy codebases it is working within,” , co-founder and CEO of Blitzy, said in a press release.

Elliott, who also previously founded and is a former Army Ranger, started Blitzy with alum in 2023. Pardeshi holds more than 27 patents related to neural networks, image generation and AI-driven interface translation.

Big money for AI software development

Several other companies in the AI software development space have raised large rounds over the past year. They include:

• , which sells the popular AI coding assistant Cursor. The company has raised $3.4 billion and was most recently valued at over $29 billion. The company has since with that gives the -led space exploration company the right to buy Anysphere for $60 billion later this year.
• , a cloud-based development platform, has raised more than $870 million, including a March that valued the company at $9 billion.��
• , a Swedish AI vibe coding startup, has raised more than $550 million, including a December $330 million Series B financing at a $6.6 billion valuation.

Related �������� query:

Illustration:

The raise brings the Swedish company’s recent Series D funding round total to $600 million. At the time of the first close in March, Legora was valued at $5.5 billion.

, and also participated in the extension, according to CNBC.

The valuation was a big jump from the $1.8 billion Legora achieved just last October, when it raised The company has now raised a total of $866 million since being founded in 2023 by , and .

Nvidia has been an active startup investor, backing over three dozen companies so far in 2026, according to �������� data. The chip giant has a stake in several of the most valuable AI companies, including , , and . It also has a number of power-generation-related investments sprinkled in, indicating ongoing concern and interest in how we are going to feed all those power-hungry AI bots.

Record legal tech funding

Meanwhile, venture funding for legal tech startups reached a record high in 2025, driven by investor enthusiasm for AI’s potential to automate the legal profession. Per �������� , companies in the legal and legal technology sectors raised $4.08 billion in seed- through growth-stage funding in 2025. That’s an impressive 77.4% increase from the $2.3 billion raised by legal tech startups in 2024.

So far this year, legal tech startups have already raised more than $1.3 billion, .

Other startups in the industry that have closed on sizable fundings over the past year include:

• : A provider of legal practice management software, Filevine announced in September that it had closed on two previously undisclosed rounds totaling $400 million. led the first round and, interestingly, joined and to co-lead the second.
• : San Francisco-based Harvey, a provider of AI tools for legal professionals, closed on four separate funding rounds in 2025, including two rounds of $300 million each. To date, the 3-year-old company has raised more than $1 billion.
• : Toronto-based Blue J, developer of a GenAI tax research platform that counts legal professionals among its core users, raised $122 million in an August Series D financing led by and .
• : Palo Alto, California-based Eudia, which develops an intelligence platform for Fortune 500 legal teams, landed up to $105 million in a Series A financing led by .

Related �������� query:

Related reading:

• Swedish Legal Tech Startup Legora Triples Valuation To $5.55B With $550M Series D Led By Accel
• Legal Tech Investment Hits All-Time High With Filevine Funding

Illustration: