Dario Amodei, the CEO of Anthropic, expressed on Thursday that he cannot ethically comply with the Pentagon's demand for unrestricted access to the company’s artificial intelligence systems. In his statement, Amodei emphasized Anthropic's recognition that military decisions should be left to the Department of Defense, rather than private enterprises. However, he pointed out that there are specific instances where AI could threaten democratic principles and that some applications of the technology exceed the limitations of current capabilities.

The situations he highlighted include mass surveillance of U.S. citizens and the deployment of fully autonomous weapons that lack human oversight. The Pentagon believes it should have the freedom to utilize Anthropic’s models for any lawful military purpose, arguing that a private company should not dictate how its technology is used.

Amodei's remarks come just before the deadline set by Defense Secretary Pete Hegseth, which was scheduled for Friday at 5:01 p.m. The Department of Defense has attempted to pressure Anthropic by labeling it as a supply chain risk—typically reserved for foreign threats—or by considering invoking the Defense Production Act, which could compel companies to prioritize production for national defense initiatives.

Amodei pointed out the inconsistency in the threats, noting, “One approach identifies us as a security risk, while the other claims Claude is vital for national security.” He affirmed that while the Department has the right to select contractors aligned with its objectives, he hopes they will reconsider given the significant value Anthropic's technology offers to the military.

Currently, Anthropic stands as the sole AI research lab with systems that are ready for classified military use, although reports suggest the DOD is preparing xAI for similar capabilities.

Amodei stated, “We strongly prefer to support the Department and our military personnel, provided that our two requested safeguards remain intact.” He also mentioned that if the Department decides to discontinue its partnership with Anthropic, the company is committed to facilitating a seamless transition to a different provider, ensuring that ongoing military operations and critical missions remain uninterrupted.

In essence, he conveyed, “We can amicably part ways if necessary. There's no reason for hostility on either side.”

Anthropic's CEO remains resolute as Pentagon deadline approaches.

A tense, dimly lit executive suite: the Anthropic logo dominates one wall as the CEO in a sharp suit stands with arms crossed, backlit by moody natural light streaming through floor-to-ceiling windows. Beyond, a looming Pentagon silhouette under stormy skies and a single U.S. flag at attention underscore the ticking deadline.

Anthropic CEO stands firm as Pentagon deadline looms - TechCrunch

- Anthropic CEO Dario Amodei refuses Pentagon's request for unrestricted AI access.  
- Amodei cites concerns over mass surveillance and fully autonomous weapons as risks.  
- Pentagon threatens Anthropic with supply chain risk designation if demands are not met.

Anthropic CEO Dario Amodei said Thursday that he “cannot in good conscience accede to [the Pentagon’s] request” to give the military unrestricted access to its AI systems.

“Anthropic understands that the Department of War, not private companies, makes military decisions,” Amodei wrote in a statement. “However, in a narrow set of cases, we believe AI can undermine, rather than defend, democratic values. Some uses are also simply outside the bounds of what today’s technology can safely and reliably do.”

The two cases are: mass surveillance of Americans and fully autonomous weapons with no human in the loop. The Pentagon believes it should be able to use Anthropic’s model for all lawful purposes, and that its uses shouldn’t be dictated by a private company.

Amodei’s statement comes less than 24 hours ahead of the Friday 5:01 p.m. deadline Defense Secretary Pete Hegseth has given Anthropic to either acquiesce to his demands, or face the consequences. The Department of Defense has attempted to force Amodei’s hand by either labeling Anthropic a supply chain risk — a designation reserved for foreign adversaries — or invoke the Defense Production Act and effectively force the firm to do its bidding. The DPA gives the president the authority to force companies to prioritize or expand production for national defense.

Amodei pointed out the contradiction in those two threats. “One labels us a security risk; the other labels Claude as essential to national security.”

He added that it’s the Department’s right to choose contractors most aligned with its vision, “but given the substantial value that Anthropic’s technology provides to our armed forces, we hope they reconsider.”

Anthropic is currently the only frontier AI lab that has classified-ready systems for the military, though the DOD is reportedly getting xAI ready for the job.

Techcrunch event Save up to $300 or 30% to TechCrunch Founder Summit 1,000+ founders and investors come together at TechCrunch Founder Summit 2026 for a full day focused on growth, execution, and real-world scaling. Learn from founders and investors who have shaped the industry. Connect with peers navigating similar growth stages. Walk away with tactics you can apply immediately.



Offer ends March 13. Save up to $300 or 30% to TechCrunch Founder Summit 1,000+ founders and investors come together at TechCrunch Founder Summit 2026 for a full day focused on growth, execution, and real-world scaling. Learn from founders and investors who have shaped the industry. Connect with peers navigating similar growth stages. Walk away with tactics you can apply immediately



Offer ends March 13. Boston, MA | REGISTER NOW

“Our strong preference is to continue to serve the Department and our warfighters—with our two requested safeguards in place,” Amodei said. “Should the Department choose to offboard Anthropic, we will work to enable a smooth transition to another provider, avoiding any disruption to ongoing military planning, operations, or other critical missions.”

TLDR, he’s saying: “We can just part ways. There’s no need to be nasty about it.”

Anthropic CEO stands firm as Pentagon deadline looms

The healthcare AI sector is currently experiencing a reality check, as investors are now looking for startups to demonstrate clear financial and operational benefits. During last week’s ViVE conference in Los Angeles, several investors shared their thoughts on the return on investment (ROI) for healthcare AI solutions.

Larry Cohen, CEO of Health2047, the venture studio of the American Medical Association, highlighted that the strained finances of health systems mean AI acquisitions increasingly require the approval of Chief Financial Officers (CFOs). This shift necessitates a robust justification for financial returns. "There's a notable change in how products are being pitched; the conversation has moved beyond just ease of use and satisfaction towards illustrating financial benefits," Cohen stated. He emphasized that AI solutions must now demonstrate their capacity to be self-sustaining financially.

Uma Veerappan, a vice president at Flare Capital Partners, pointed out that many healthcare AI startups are still in the early stages of validating their financial returns, especially for tools designed for clinical applications. In her view, the initial focus may need to be on user adoption rather than immediate ROI. “Given the nascent stage of many of these companies, it's not critical for ROI to materialize within the first six months. What's pivotal is showcasing product stickiness. Over a longer timeline, financial and clinical ROI will become essential benchmarks,” Veerappan noted.

Rachel Feinman, managing director at Tampa General Hospital Ventures, explained that ROI can vary greatly depending on the type of AI tool in question. For instance, tools like ambient clinical documentation scribe systems tend to provide indirect ROI; healthcare organizations may invest in them to enhance their appeal to prospective physicians, even if they don’t contribute to immediate revenue increases. "Healthcare systems are beginning to promote their use of ambient listening technology for providers, making it a top consideration for recruitment," Feinman said. On the other hand, AI systems designed to automate administrative tasks—such as revenue cycle management and data processing—often deliver clearer and more rapid financial returns.

Vig Chandramouli, a partner at Oak HC/FT, observed a trend where more AI startups are moving away from conventional SaaS pricing models toward outcome-based pricing, such as charges per transaction or completed task. This approach allows clients to feel they are remunerating solely for measurable outcomes. "Businesses are being compelled to adopt pricing that reflects three to five times ROI. Value capture is central to nearly every business pitch," he remarked. However, Chandramouli cautioned that outcome-based pricing carries hidden risks, particularly if a startup miscalculates the costs associated with achieving successful outcomes, which could undermine unit economics. He stressed the importance for startups to meticulously define contract terms to clarify what constitutes a successful action, distinguish between fixed and variable fees, and assess the realism of projected revenues.

AI Startups Face ROI Evaluation: Insights from 4 VCs on Current Priorities

A moody, dimly lit boardroom scene: a lone entrepreneur stands under a warm natural spotlight at the head of a dark mahogany table, facing four shadowy VCs seated in deep shadow. Behind them, a large digital screen displays a bold ROI graph overlaid with a stylized AI chip icon. Scattered pitch decks, a small pile of coins, and tense body language tell the story of an AI startup’s high-stakes reckoning. Dark theme, cinematic natural light, dramatic storytelling.

AI Startups’ ROI Reckoning: What 4 VCs Say Matters Now - MedCity News

- Healthcare AI startups must demonstrate clear financial returns to gain CFO approval for purchases.  
- Long-term adoption is increasingly prioritized over immediate ROI for healthcare AI tools.  
- Shift towards outcome-based pricing models is emerging among AI startups for clarity on value.  

Healthcare AI is entering a reality check phase, with investors now expecting startups to prove clear financial and operational value.

Here are five soundbites from interviews with investors last week at the ViVE conference in Los Angeles about how they are thinking about ROI for healthcare AI tools.

The CFO test

Health systems’ tight finances mean that AI purchases increasingly require CFO approval, which means there needs to be a solid case for financial returns, said Larry Cohen, CEO of Health2047, the American Medical Association’s venture studio.

“You can see it in the way people talk about their products and try to pitch what they’re doing. Now, everybody just doesn’t talk about how much easier it is or how much happier everybody is. Now, they’re coming around to financial ROI,” he declared.

In today’s market, Cohen said, AI tools have to show they can pay for themselves.

The long game

Many healthcare AI companies are still in the early innings when it comes to proving financial returns, particularly for tools that assist with clinical use cases, pointed out Uma Veerappan, vice president at Flare Capital Partners.

In the near term, she believes demonstrating adoption could matter more than showing immediate ROI for some startups.

“With the early stage nature of many of these companies, I don’t necessarily believe that ROI needs to show within the first six or so months of deployment. Rather, I think the ability to show you have a sticky product is what’s really important. After a longer period of time is probably when financial and clinical ROI become more table stakes,” Veerappan remarked.

ROI depends on the tool

ROI looks different depending on the kind of AI tool, noted Rachel Feinman, managing director of Tampa General Hospital Ventures.

Some tools, such as ambient clinical documentation scribes, deliver ROI indirectly rather than through immediate financial gains, she pointed out. For example, health systems may adopt these tools to stay competitive in recruiting physicians, even if they don’t increase visit volume or short-term revenue.

“Already, health systems and clinics are starting to advertise that they have ambient listening for providers, and providers are putting that at the top of their list as it relates to where they want to go — that’s where they want to be,” Feinman explained.

At the same time, she said AI that automates manual administrative work — especially for tasks like revenue cycle automation and data abstraction — typically offers clearer and more immediate financial returns.

ROI-driven business models

Vig Chandramouli, partner at Oak HC/FT, said that he is seeing more AI startups shifting away from traditional SaaS pricing toward models tied directly to outcomes, such as charging per completed task or transaction. This approach helps customers feel they are paying only for measurable results, he stated.

“Companies are getting forced into pricing at three to five times ROI. Value capture is front and center on almost every pitch,” Chandramouli declared.

He warned, however, that transaction-based pricing can create hidden risks. If a company underestimates the true cost of delivering each successful transaction, the unit economics can quickly erode.

In order for the shift to ROI-based pricing to work, Chandramouli thinks startups will have to carefully unpack each contract — clarifying what counts as a successful action, what fees are fixed versus variable and how realistic projected revenue is.

Photo: Malte Mueller, Getty Images

AI Startups’ ROI Reckoning: What 4 VCs Say Matters Now

During a public Q&A on Saturday evening, OpenAI CEO Sam Altman faced a challenging landscape regarding the company's recent acquisition of a Pentagon contract from competitor Anthropic. At approximately 7 p.m., Altman took to X to clarify OpenAI's position after Anthropic opted out, primarily due to concerns about mass surveillance and automated weaponry—issues that OpenAI appears willing to embrace.

In response to critical inquiries, Altman refrained from taking a firm stance on policy matters, emphasizing his faith in the democratic process and the authority of elected officials. "I very deeply believe in the democratic process," he stated in one of his replies, reiterating the need to uphold constitutional principles.

An hour into discussions, Altman expressed surprise at the volume of skepticism, noting an unexpected level of debate surrounding whether power should rest with democratically elected governments or unelected corporations. "I guess this is something people disagree on,” he commented, reflecting on the gravity of the situation for both OpenAI and the broader tech landscape.

This moment serves as a pivotal turning point for OpenAI, which is transitioning from a thriving consumer-oriented startup to a significant player in national security. During his Q&A, Altman's approach echoed standard practices in the defense sector, where military leaders typically defer to civilian oversight. However, as OpenAI navigates these new waters, it seems to lack the necessary tools and strategies to handle this transformation effectively.

Altman’s public engagement coincided with a crucial period for the company, following the Pentagon’s decision to blacklist Anthropic for its demands regarding surveillance and automated weaponry, leading to OpenAI securing the same contract. While he framed the acquisition as a means to ease tensions, he appeared unprepared for the backlash from both employees and users alike.

In recent years, OpenAI has sought engagement with the U.S. government, but this situation feels radically different. When Altman addressed Congress in 2023, he navigated discussions with a blend of enthusiasm and caution about the company’s potential, aiming to galvanize investor support while steering clear of regulatory constraints. Now, however, the urgency of AI innovation and funding requirements necessitate a more serious engagement with government entities, revealing a lack of preparedness on both sides.

The most pressing conflict at hand is the situation with Anthropic and the U.S. Defense Secretary Pete Hegseth’s comments suggesting the lab could be marked as a supply chain risk. This looming threat has the potential to cripple Anthropic, as pointed out by former Trump administration official Dean Ball, who noted that this could sever the company from essential hardware and hosting services, creating an unprecedented crisis. This scenario serves as a chilling precedent for other tech firms considering relationships with the government.

Even if Hegseth modifies his stance, significant damage has already been done, as Ball articulated. The overarching sentiment is that the dynamics have effectively shifted, mandating corporations and political figures to operate under the increasingly dominant tribal logic.

While this development poses a direct crisis for Anthropic, it also creates complications for OpenAI, which faces mounting internal pressure to maintain ethical boundaries. At the same time, right-wing media scrutiny intensifies, searching for any indications that OpenAI might not align strictly with political allies. Amid these tensions, the Trump administration appears intent on complicating matters further.

Though OpenAI may not have set out to become a defense contractor, its ambitious aspirations have forced it into competition with established firms like Palantir and Anduril. Engaging with the Trump administration necessitates making political choices, and there are no neutral players in this field, meaning that forging relationships with one group inevitably alienates another. The repercussions of these decisions could lead OpenAI to face challenges, whether through lost contracts or employee dissatisfaction.

The current crackdown occurs at a time when numerous prominent tech investors hold substantial influence in Washington. Yet, many seem favorably inclined towards tribal dynamics, perceiving Anthropic as having aligned more with the Biden administration, thus raising concerns among industry insiders. With the tables turned, few appear willing to champion the concept of free enterprise.

This situation presents a formidable challenge for any company, as those aligned with specific political factions may gain short-term benefits but would find themselves vulnerable to shifts in the political landscape. Historically, the defense sector has been characterized by well-established, heavily regulated players like Raytheon and Lockheed Martin, who, due to their long-standing connection with the Pentagon, could navigate changes in administration while maintaining focus on technological advancement.

Today’s emerging tech firms may be quicker on their feet than their predecessors, yet they seem less equipped for the sustained challenges ahead.

There is a lack of effective strategies for the collaboration between AI companies and the government.

A moody neoclassical government hall at dawn, long shadows stretching across a grand circular table shaped like interlocking puzzle pieces—adult executives in dark suits on one side, suited officials with clipped folders on the other—staring at a single missing piece glowing faintly in the center. Warm shafts of natural light spill through towering arched windows, highlighting tense body language and the unresolved gap at the heart of the scene.

No one has a good plan for how AI companies should work with the government - TechCrunch

- OpenAI's CEO Sam Altman faced backlash for accepting a Pentagon contract previously rejected by Anthropic.  
- Altman emphasized the need for democracy while expressing surprise at public dissent on government power.  
- OpenAI's transition to national security raises concerns about its preparedness for government engagement.  

As Sam Altman discovered Saturday night, it’s a fraught time to do work for the U.S. government. Around 7 p.m., the OpenAI CEO announced he would be fielding questions publicly on X, as a way of demystifying his company’s decision to pick up the Pentagon contract that Anthropic had just walked away from.

Most of the questions boiled down to OpenAI’s willingness to participate in mass surveillance and automated killing – the exact activities Anthropic had ruled out in its negotiations with the Pentagon. Altman typically punted to the public sector, saying it wasn’t his role to set national policy.

“I very deeply believe in the democratic process,” he wrote in one response, “and that our elected leaders have the power, and that we all have to uphold the constitution.”

An hour later, he confessed surprise that so many people seemed to disagree. “There is more open debate than I thought there would be,” Altman said, “about whether we should prefer a democratically elected government or unelected private companies to have more power. I guess this is something people disagree on.”

It’s a telling moment for both OpenAI and the tech industry at large. In his Q&A, Altman employed a stance that’s standard in the defense industry, where military leaders and industry partners are expected to defer to civilian leadership.

But what’s more telling is that, as OpenAI transitions from a wildly successful consumer startup into a piece of national security infrastructure, the company appears unequipped to manage its new responsibilities.

Altman’s public town hall came at a heightened time for his company. The Pentagon had just blacklisted OpenAI rival Anthropic for insisting on contractual limitations for surveillance and automated weaponry. Days later, OpenAI announced it had won the same contract Anthropic had given up. Altman portrayed the deal as a quick way to deescalate the conflict – and it was surely a lucrative one. But he seemed unprepared for how much blowback it generated from both the company’s users and its employees.

Techcrunch event Disrupt 2026: The tech ecosystem, all in one room Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $400. Save up to $300 or 30% to TechCrunch Founder Summit 1,000+ founders and investors come together at TechCrunch Founder Summit 2026 for a full day focused on growth, execution, and real-world scaling. Learn from founders and investors who have shaped the industry. Connect with peers navigating similar growth stages. Walk away with tactics you can apply immediately



Offer ends March 13. San Francisco, CA | REGISTER NOW

OpenAI has been engaging with the U.S. government for years — but not like this. When Altman was making his case to the Congressional committees in 2023, for instance, he was still mostly following the social media playbook. He was bombastic about the company’s world-changing potential while acknowledging the risks and enthusiastically engaging with lawmakers — a perfect combination for stirring up investors while heading off regulation.

Less than three years later, that approach is no longer tenable. AI is so obviously powerful and the capital needs are so intense that it’s impossible to avoid a more serious engagement with the government. The surprise is how unprepared both sides seem to be for it.

The biggest immediate conflict is Anthropic itself, and U.S. Defense Secretary Pete Hegseth’s stated plan Friday to designate the lab as a supply chain risk. That threat looms over the whole conversation like an unfired gun. As former Trump official Dean Ball wrote over the weekend, the designation would cut Anthropic off from hardware and hosting partners, effectively destroying the company. It would be an unprecedented move against an American company, and while it might ultimately be reversed in court, it will cause damage in the interim and send shockwaves through the industry.

As Ball describes the process, Anthropic was carrying out an existing contract under terms that had been established years earlier – only to have the administration insist on changing the terms. It’s far beyond anything that would fly between private companies, and sends a chilling message to other vendors.

“Even if Secretary Hegseth backs down and narrows his extremely broad threat against Anthropic, great damage has been done,” Ball wrote. “Most corporations, political actors, and others will have to operate under the assumption that the logic of the tribe will now reign.”

It’s a direct threat to Anthropic, but also a serious problem for OpenAI. The company is already under intense pressure from employees to maintain some semblance of a red line. At the same time, right-wing media will be on alert for any sign of OpenAI being a less-then-staunch political ally. In the middle of everything is the Trump administration, doing its best to make the situation as difficult as possible.

It can be argued that OpenAI didn’t set out to become a defense contractor, but by virtue of its massive ambitions, it’s been forced to play the same game as Palantir and Anduril. Making inroads during the Trump administration means picking sides. There are no apolitical actors here, and winning some friends will mean alienating others. It remains to be seen how high a price OpenAI will pay, either in lost business or lost employees, but it’s unlikely to emerge unscathed.

It might seem strange that this crackdown is coming at a time when there are more prominent tech investors holding influential positions in Washington than ever, but most of them seem entirely happy with tribal logic. Among Trump-aligned venture capitalists, Anthropic has long been perceived as currying favor with the Biden administration in ways that would damage the larger industry – a perception underscored by Trump advisor David Sacks’ reaction to the ongoing conflict. Now that the reverse has happened, few seem willing to stand up for the broader principle of free enterprise.

This is a difficult position for any company to be in – and while politically aligned players may benefit in the short term, they’ll be just as exposed when political winds inevitably shift. There’s a reason why, for decades, the defense sector was dominated by slow-moving, heavily regulated conglomerates like Raytheon and Lockheed Martin. Operating as an industrial wing of the Pentagon gave them the political cover they needed to avoid the politics, staying focused on the technology without having to press reset every time the White House changed hands.

Today’s startup competitors might move faster than their predecessors – but they’re much less prepared for the long term.

No one has a good plan for how AI companies should work with the government

Artificial intelligence (AI) is evolving beyond basic automation, effectively addressing the bottlenecks often encountered in finance and enterprise operations. Leading this shift is Sphinx Labs, which recently secured $7.1 million in seed funding, along with Basis, which has reached an impressive valuation of $1.15 billion. These startups are innovating by introducing autonomous systems that function similarly to digital employees.

Sphinx Labs focuses on developing AI agents to eliminate the delays caused by human involvement in financial compliance tasks. Co-founder Alexandre Berkovic pointed out that despite advancements in compliance software, the process's speed remained hampered by the human element: analysts still needed to log in, assess data, and make decisions manually. He remarked, “The efficiency gains of upgrading platforms and data were marginal.” 

The unique strategy of Sphinx involves deploying agents that integrate directly with current compliance systems. According to Berkovic, “Sphinx operates just like a human would, utilizing a username and password to access various systems.” These AI agents are trained on company-specific guidelines, policies, and risk assessments, enabling them to conduct reviews that adhere to the organization's standards.

Initially, the AI offers recommendations for human analysts to approve or reject. However, as trust in the system grows, Sphinx plans to automate the entire decision-making process. The company collaborates with prominent banks and cryptocurrency payment processors, particularly in high-volume sectors where compliance can impose significant operational burdens. Berkovic noted that in some instances, Sphinx has managed to automate over 92% of compliance tasks, leaving only 8% for manual review. It has also identified up to 17% more genuine positive cases of suspicious activities that might otherwise be overlooked.

In addition to Sphinx’s achievements, several other funding highlights emerged last week. MatX has raised $500 million to develop cutting-edge AI chips that aim to compete with Nvidia in high-performance AI processing, according to Bloomberg reports. This funding reflects ongoing investor enthusiasm for foundational AI infrastructure as computational needs escalate.

Encord, on the data management front, obtained $60 million in Series C funding to enhance its AI-driven data infrastructure solutions. The company specializes in managing and annotating diverse datasets, paving the way for enterprises to evolve beyond text-based AI into more practical applications.

In terms of enterprise AI and workflow automation, Basis has secured $100 million in a Series B funding round led by Accel and Google Ventures, achieving a valuation of $1.15 billion. This firm focuses on creating AI agents for accounting firms, streamlining complex workflows related to tax, audits, and bookkeeping.

Moreover, Koah has completed a $20.5 million Series A financing round with plans to develop AI-native monetization systems that seamlessly integrate with generative AI chat platforms, fostering revenue models tailored to these interactive environments.

Finally, while not solely focused on AI, Honest Health has raised $140 million to enhance its platform for value-based care, aiding health systems in improving care coordination and cost management as they move away from traditional fee-for-service practices. This showcases the ongoing interest among investors in technology-driven innovations aimed at transforming healthcare through data.

Investors Increase Investments in the Agent Economy

A sunlit, modern trading room where a diverse group of sharply dressed investors leans over a sleek glass table. On the table sits a miniature cityscape with dozens of tiny service robots bustling along illuminated streets. Each investor places golden poker chips into a slot labeled “Agent Economy,” while a wall-mounted ticker displays a steeply rising graph. Warm, natural light floods the scene, highlighting eager faces and the tokens in mid-motion to convey the rush of rising bets on autonomous agents.

Investors Ramp up Bets on the Agent Economy - PYMNTS.com

- Sphinx Labs raised $7.1 million to develop AI agents for financial compliance operations.  
- AI agents aim to automate 92% of compliance tasks, significantly reducing human bottlenecks.  
- Basis secured $100 million to create AI agents for automating accounting workflows.

Artificial intelligence (AI) agents are moving beyond simple automation to ease the human bottleneck across finance and enterprise workflows. Led by Sphinx Labs’ $7.1 million seed round and Basis’ new $1.15 billion valuation, startups are deploying autonomous systems that operate like digital employees.

By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions .

Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

Startup Spotlight

Sphinx Labs is building AI agents designed to replace the human bottleneck inside financial compliance operations. The idea emerged from a simple observation. As Co-founder Alexandre Berkovic explained to PYMNTS, no matter how much compliance software improved, “it didn’t really change the time to review cases, and the reason was that the bottleneck was humans.” Analysts still had to log in, review data and make decisions manually. “The efficiency gains of making a better platform and better data were marginal.”

Sphinx’s approach is to deploy agents that operate directly inside existing compliance systems.

“Sphinx works just like as a human. So you’re going to give Sphinx a username and a password, access to different systems,” Berkovic said. The system ingests company-specific “procedures, policies, SOPs, risk matrices, anything really specific to that company” and executes reviews in line with those internal standards.

Initially, the agents suggest decisions for human approval. “Analysts are going to approve or reject our decision,” he said. Over time, as confidence builds, “we’re going to start automating the decision end to end.”

The company works with “big banks and payments processors in crypto and FinTechs,” particularly in high-volume environments where compliance creates heavy operational overhead. Berkovic said that in some deployments, Sphinx has “been able to automate over 92% of compliance operations,” with the remaining 8% routed to manual review. It has also “been able to find up to 17% more true positives,” detecting suspicious activity human teams might miss.

Advertisement: Scroll to Continue

Sphinx recently raised $7.1 million in a seed round led by Cherry Ventures, with participation from Y Combinator, Rebel Fund, Deel Ventures and Singularity Capital.

We’d love to be your preferred source for news. Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks! Add as Preferred Source

Other funding rounds from last week include:

AI Infrastructure and Compute

MatX raised $500 million to build next-generation AI chips designed to compete with Nvidia in high-performance AI workloads, according to Bloomberg.

The company is developing silicon optimized for large-scale model training and inference, positioning itself as part of a growing wave of startups attempting to challenge the dominance of incumbent GPU providers. The round underscores sustained investor appetite for foundational AI infrastructure as demand for compute continues to rise.

On the data layer, Encord secured $60 million in a Series C round to scale its AI-native data infrastructure platform. The company provides tools to manage, curate and annotate multimodal datasets, including video, imagery and sensor data, as enterprises expand beyond text-based AI applications into physical and real-world deployments. As model performance increasingly hinges on high-quality training data, startups like Encord are targeting a core bottleneck in the AI stack.

Enterprise AI and Workflow Automation

Basis raised $100 million in a Series B round led by Accel and Google Ventures at a reported $1.15 billion valuation. The company builds AI agents for accounting firms, automating complex workflows across tax, audit and bookkeeping.

Koah, meanwhile, closed a $20.5 million Series A to scale AI-native monetization infrastructure. The startup is building contextual advertising systems designed to integrate directly within generative AI chat experiences, aiming to create revenue models tailored to conversational interfaces rather than traditional display formats.

Outside AI infrastructure, Honest Health raised $140 million to expand its value-based care platform for health systems. The company partners with providers to improve care coordination, analytics and cost management as hospitals transition away from fee-for-service models. While not exclusively an AI company, its technology-enabled approach reflects continued investor interest in data-driven healthcare transformation.

Investors Ramp up Bets on the Agent Economy

During the week of the NFL Draft, Pittsburgh will see over $1.2 million dedicated to funding in the artificial intelligence startup sector, with Mark Cuban, a native of Mt. Lebanon, serving as one of the judges.

On April 22, an event titled "Powering the Future of Sport: A Draft Week Showcase" will take place at Carnegie Mellon University, organized in collaboration with the Pittsburgh Steelers and the AI Strike Team. This exclusive event will unfold at the Robotics Innovation Center in Hazelwood Green, where Pennsylvania's Governor Josh Shapiro is scheduled to make a brief speech, and guests will witness live demonstrations of cutting-edge AI technology, according to a recent press release.

Carnegie Mellon President Farnam Jahanian expressed that hosting this event during the NFL Draft is an opportune moment to showcase how sports are serving as a testing ground for the future of AI through aspects like play, performance, engagement, and entertainment.

Cuban, noted for his role as an investor on ABC’s “Shark Tank” until May 2025, will be joined on the judging panel by Ed Stack, the executive chairman of Dick's Sporting Goods and Foot Locker, along with former Pittsburgh Steeler Will Allen.

Participants in this competition will vie for a prize that includes $1 million in Amazon Web Services compute credits—essentially a form of promotional currency—and an additional $275,000 in investment capital from Magarac Venture Partners located in Bakery Square.

The judges will be specifically looking for startups that incorporate AI and robotics to enhance athlete performance, training and safety, game and athlete analytics, fan engagement, as well as advancements in venue and broadcast technologies and health and wellness initiatives, as outlined in the application criteria.

Startups from both the U.S. and abroad are welcome to apply, although preference will be given to those committed to establishing a business presence in Pennsylvania, according to the event's website.

The criteria for applicants mention that a "substantial operating presence" could encompass a range of things, such as setting up a headquarters or primary office, launching a significant satellite office, or positioning engineering, R&D, manufacturing, or deployment operations in the state.

Dan Rooney, the Steelers' vice president of strategy, commented on the evolution of the NFL, noting, "Artificial intelligence is transforming the NFL, with millions of data points collected each game reshaping how teams compete and how fans engage."

Mark Cuban participates in panel for Pittsburgh AI startup amid NFL Draft week

Natural light floods a modern Pittsburgh conference hall. Mark Cuban stands at a sleek podium, leaning forward mid-conversation with three startup founders seated at a long table strewn with laptops and small AI hardware prototypes. Floor-to-ceiling windows behind them frame the iconic steel bridges and downtown skyline under a bright spring sky. On one side of the stage, two glossy football helmets rest on a stand, subtly nodding to NFL Draft week.

Mark Cuban joins Pittsburgh AI startup panel during NFL Draft week - TribLIVE.com

- Over $1.2 million in funding for AI startups will be awarded in Pittsburgh.  
- The event, hosted by Carnegie Mellon and the Steelers, occurs during NFL Draft week.  
- Judges will evaluate startups focusing on AI in sports performance, analytics, and engagement.

More than $1.2 million in artificial intelligence startup funding will be awarded in Pittsburgh during NFL Draft week, with Mt. Lebanon native Mark Cuban serving as a judge.

On April 22, Carnegie Mellon University in partnership with the Pittsburgh Steelers and the AI Strike Team will host Powering the Future of Sport: A Draft Week Showcase, an artificial intelligence event.

The invitation-only event will be held at Carnegie Mellon University’s Robotics Innovation Center in Hazelwood Green, where Pennsylvania Gov. Josh Shapiro will deliver brief remarks and attendees will view live demonstrations of AI technology, the news release said.

“With the NFL Draft coming to our city, it is a fitting moment to highlight how sport – through play, performance, engagement, and entertainment – has become a proving ground for the next era of AI,” saidCarnegie Mellon University President Farnam Jahanian.

Cuban, who was a regular investor on ABC’s “Shark Tank” until May 2025, will be joined by Ed Stack, executive chairman of Dick’s Sporting Goods and Foot Locker and former Pittsburgh Steeler Will Allen on the panel of judges.

The prize includes $1 million in Amazon Web Services compute credits, which act as promotional currency, and $275,000 in investment capital provided by Magarac Venture Partners, in Bakery Square.

The judges’ panel will be looking for startups implementing AI and robotics in athlete performance, training and safety; game and athlete analytics, fan engagement, venue and broadcast technology and health and wellness, according to the application requirements.

Both U.S. and international companies are eligible to apply, but preference will be given to companies that commit to establishing an operating presence in the Commonwealth of Pennsylvania, the website said.

Applicant requirements state that “substantial operating presence” may include, but is not limited to: establishing a headquarters or primary office, opening a significant satellite office and locating engineering, R&D, manufacturing, or deployment operations within the region.

“Artificial intelligence is transforming the NFL, with millions of data points each game reshaping how teams compete and how fans engage,” Dan Rooney, Steelers vice president of strategy, said in a statement.

Mark Cuban joins Pittsburgh AI startup panel during NFL Draft week

The customer service sector is currently undergoing significant changes, largely driven by advancements in AI technology. While apprehensions grow among investors and industry executives regarding the Business Process Outsourcing (BPO) field, AI-enhanced customer support firms like Decagon, Parloa, and Sierra are successfully attracting substantial funding from venture capital sources.

One standout in this space is 14.ai, a startup supported by Y Combinator, which has pioneered an AI-centric approach by transforming traditional customer support teams at numerous startups they collaborate with.

Recently, 14.ai secured $3 million in seed funding, led by Y Combinator with additional backing from General Catalyst, Base Case Capital, SV Angel, and notable founders from companies such as Dropbox, Slack, Replit, and Vercel.

The founders, Marie Schneegans and Michael Fester, are a married couple who met in Paris over a decade ago. They initially built separate companies; Schneegans co-founded Workwell, focused on corporate intranet solutions, while Fester developed Snips, a business specializing in local-first assistants for smart devices—later acquired by Sonos in 2019.

Motivated to work together, they relocated to the U.S. and decided to address the challenges in customer service, opting to create a hybrid model rather than a traditional SaaS platform. Thus, 14.ai emerged as an AI-driven customer support agency.

“We don’t just create software solutions for our clients. 14.ai operates as a complete AI-driven customer service agency that combines software with services. Handling customer support software can be complex for businesses, so we take on their entire workload, utilizing our tailored technology stack,” Fester explained.

The firm claims it can seamlessly integrate with existing support systems within just a day, quickly tackling support ticket backlogs. It has the capability to manage inquiries across multiple platforms, including email, phone calls, chat, TikTok, Facebook, Telegram, and WhatsApp.

As an example, they recently collaborated with a men's health supplements startup, Sperm Worms, founded by a fellow Y Combinator alumnus, which was struggling with an overwhelming number of support tickets. After taking over on a Thursday morning, 14.ai had resolved outstanding tickets from various channels by Thursday afternoon.

Currently, the startup consists of a small team of six, who rotate availability to provide round-the-clock support for their clients. With the newly secured funding, 14.ai aims to expand its workforce in the coming months.

The startup focuses exclusively on hiring AI engineers and plans to further increase this area of expertise. By analyzing customer support workflows and other operational aspects, such as sales and revenue generation, 14.ai seeks to automate processes, enabling humans to devote less time to routine tasks.

“We function not merely as a support agency, but as a catalyst for revenue growth, capturing valuable insights from customer interactions early on,” Fester noted.

14.ai aims to eliminate three key expenses from a startup's financials: ticketing systems, AI software add-ons, and labor costs. Their client roster spans various industries, including luxury skincare brand Yon-KA, smart glasses manufacturer Brilliant Labs, and Creative Lighting.

Additionally, the startup intends to enhance its solutions by allowing AI to manage most operational tasks autonomously. They are currently running GloGlo, a brand of glucose gummies designed for individuals with Type 1 diabetes, with the goal of achieving autonomous operation through AI.

Tom Blomfield, a Y Combinator partner, believes 14.ai successfully merges human expertise with AI in customer service. He stated that with effective implementation, AI could handle approximately 60% of tasks, leaving the remaining 40% for human involvement.

“As AI takes on an increasing share of the workload, the dynamics between AI and human roles will evolve. Current systems often require customers to repeatedly manage painful workforce reductions,” he shared with TechCrunch via email.

“In contrast, 14.ai acts as the entire customer service department, integrating both AI and human elements. They can effectively reallocate customer support agents across clients at varying stages of AI integration, optimizing resource distribution,” he added.

Importantly, AI-focused agencies are precisely what Y Combinator has highlighted in its 2026 startup calls for proposals.

Married founders' company, 14.ai, is taking over customer support roles at startups.

Dimly lit open-plan startup office at midnight, a sleek AI terminal with a large glowing “14.ai” logo stands center stage, rows of abandoned customer support desks and headsets fading into deep shadows; in the foreground a married founder duo seen from behind, arms linked, surveying the silent room—moody, dark theme.

A married founder duo’s company, 14.ai, is replacing customer support teams at startups - TechCrunch

- AI is transforming the customer service industry, impacting traditional BPO operations significantly.  
- 14.ai, co-founded by Marie Schneegans and Michael Fester, raised $3 million in seed funding.  
- The startup aims to combine AI and human support, optimizing customer service operations efficiently.  

The customer service industry is in a bit of flux, thanks to AI. Investors and corporate leaders have rung alarm bells for the BPO (Business Process Outsourcing) industry. On the other hand, AI-powered customer support startups such as Decagon, Parloa, and Sierra have picked up millions of dollars in funding from venture capitalists.

14.ai, a Y Combinator-backed startup, is taking an approach of building an AI-native agency that has replaced legacy customer support teams at many startups it has worked with.

The company has raised $3 million in seed funding led by Y Combinator, with participation from General Catalyst, Base Case Capital, SV Angel, and the founders of Dropbox, Slack, Replit, and Vercel.

The startup was founded by a married duo, Marie Schneegans and Michael Fester. The two met in Paris more than a decade ago and went on to build separate companies. Schneegans was a co-founder at corporate intranet company Workwell. Fester previously founded Snips, a company that worked on local first assistants for smart devices, which was acquired by Sonos in 2019.

After this, they wanted to build a company together, so they moved to the U.S. The duo picked up customer service as the problem to tackle, but didn’t want to build a pure-play SaaS company. They founded 14.ai to operate as an AI-native customer support agency of sorts.

“We’re not building software for customers. 14.ai is an AI-native customer service agency. We combine software and services in one package. For customers, operating software is hard, especially for customer service. We take over their entire operation, and we use our own purpose-built stack for customer service,” Fester said.

The company said it can integrate with a support system within a day and start clearing the support ticket backlog very quickly. It can monitor tickets across various channels, including email, calls, chat, TikTok, Facebook, Telegram, and WhatsApp.

“We started working with a men’s health supplement company called Sperm Worms by a former YC founder, who had a lot of backlog of tickets. His team of customer service agents was in the Philippines, and they were not being able clear tickets efficiently. We took over on Thursday morning, and by Thursday afternoon, we had cleared tickets from all channels like social media, SMS, email, chat, and voice,” Schneegans said.

The company currently has six people working, and they all take turns to be available around the clock for the clients they work with. The startup said that with the new funding, it aims to increase the headcount in the next six months.

14.ai is only working with AI engineers and plans to hire more AI engineers. The startup learns workflows of customer support and other functions, such as sales and revenue growth, and tries to automate tasks through its software so humans have to spend less time on particular issues.

“We are not just a support agency, but also a revenue growth engine because we capture all kinds of conversations early on for a client and get insights from them,” Fester said.

The company wants to take off three key items from a startup’s balance sheet, including ticketing systems, AI software add-ons, and human labor costs. The startup caters to many clients in different sectors such as luxury skin care brand Yon-KA, smart glasses maker Brilliant Labs, and lighting company Creative Lighting.

The startup also wants to improve its own product by experimenting and letting AI handle most tasks. For that, it runs GloGlo, a glucose gummies brand for Type 1 diabetics, and tries to operate autonomously with AI.

Tom Blomfield, a partner at Y Combinator, thinks that 14.ai strikes the right balance between using AI and humans for customer service. He said that with the right integration, AI can solve 60% of the task automatically, and the remaining 40% could be handled by humans.

“As the AI takes over more and more of the work, the balance between AI and humans will change over time. With the existing platforms, the customer is left to handle round after round of painful headcount reductions,” he told TechCrunch over email.

“In contrast, 14.ai becomes the customer service department, both AI and human. They can reassign customer support agents between customers who are at different stages of the AI adoption journey, and carry out that load balancing much more effectively,” he added.

Notably, AI-powered agencies are one of the things Y Combinator mentioned in its requests for startups in 2026.

A married founder duo’s company, 14.ai, is replacing customer support teams at startups

Businesses are gearing up for a future dominated by agentic commerce, where AI agents take on the responsibility of purchasing products for consumers.

Currently, this phenomenon isn't widespread. At present, AI-driven answer engines utilizing large language models (LLMs) like OpenAI’s ChatGPT and Google’s AI Mode and Gemini are assisting consumers in their shopping journeys, but they’re not yet significantly impacting sales figures. Data from Similarweb indicates that between August 2025 and January, these AI answer engines directed 49.5 million visitors to the online marketplaces of five major retailers—Amazon, Walmart, Target, Temu, and eBay. Amazon accounted for 28% of this traffic, closely followed by Walmart, which attracted 27%.

While this represents a noticeable increase, it still constitutes a small portion of overall ecommerce traffic. The exact nature of agentic commerce remains uncertain as it evolves, but various potential scenarios are already emerging.

6 Innovative AI Search Startups Transforming the Future of Retail

A bright, sunlit futuristic shopping arcade with six sleek, freestanding digital kiosks in a row—each displaying a distinct glowing abstract “search” icon—shoppers casually tapping and browsing products on the touchscreens as colorful merchandise lines the stalls around them. Warm natural light streams through a glass ceiling onto smooth wood and glass architecture, evoking innovation and the next chapter of retail powered by AI.

How 6 Hot AI Search Startups Are Shaping the Future of Shopping - ADWEEK

- Businesses are preparing for agentic commerce with AI agents facilitating consumer purchases.  
- Current AI engines are guiding purchases but driving limited sales at e-commerce sites.  
- Between August 2025 and January, AI traffic was fractional, benefiting major retailers.  

Businesses are preparing for an era of agentic commerce, when AI agents buy products on behalf of consumers.

That’s not happening at scale yet. Right now, AI answer engines powered by LLMs like OpenAI’s ChatGPT and Google’s AI Mode and Gemini guide people toward purchases, but are not driving significant sales. Between August 2025 and January, AI answer engines drove 49.5 million visitors to the ecommerce sites of five big retailers—Amazon, Walmart, Target, Temu, and eBay—according to Similarweb, with Amazon grabbing 28% of that traffic followed by Walmart with 27%.

That’s a growing but still fractional amount of total ecommerce traffic, and while it’s unclear what agentic commerce will actually look like once it matures, there are already multiple scenarios.

How 6 Hot AI Search Startups Are Shaping the Future of Shopping

Canva has reassured its workforce that the integration of its recent artificial intelligence acquisition, Leonardo.AI, into the design platform will not result in any job losses. The company dismissed media speculations regarding layoffs as inaccurate.

According to a report from The Australian, Cliff Obrecht, Canva's chief operating officer and co-founder, held a video conference last week with Leonardo's 150 employees in North Sydney, indicating that the organization would be undergoing restructuring. Obrecht reportedly informed the team that while many employees would be aligned with ongoing AI initiatives, some would be encouraged to explore other job opportunities within Canva.

Lachlan, Canva’s Head of Corporate and Internal Communications, clarified to Startup Daily that the company is not considering layoffs, stating that discussions of job losses were never communicated to the staff. He criticized The Australian for not reaching out for a comment before publishing their article, emphasizing that had they done so, the facts could have been clearly communicated.

The article appeared against the backdrop of growing concerns surrounding job security in the tech sector, particularly following the announcement by US fintech company Block of a 40% workforce reduction, amounting to 4,000 jobs, and Sydney’s WiseTech Global stating it would eliminate 2,000 coding positions, both citing AI advancements as the rationale.

Canva expressed its commitment to intensifying its AI integration efforts with Leonardo, explaining that a small team will continue to develop Leonardo as a separate product while others will transition into one of over 100 AI-focused product and research teams. Since its acquisition by Canva in 2024 for over $300 million, Leonardo has functioned as an independent entity, with its technology already woven into significant aspects of Canva’s offerings.

A spokesperson highlighted Leonardo’s reputation as a leader in image and video generation, an essential component for Canva's expansive user base of 265 million. The spokesperson stated that the integration phase would see team members from Leonardo shift into various teams across Canva, aiming to leverage their expertise across Canva’s broader AI product suite while still investing in Leonardo as a standalone offering, ensuring support for its existing community.

Additionally, Obrecht recently announced Canva's acquisition of two more startups, Mango.AI and Cavalry, from the US and UK. Mango.AI, a 10-month-old marketing algorithm startup based in San Francisco, was still operating in stealth mode, while Cavalry is a UK-based platform specializing in 2D animation for motion designers.

With these new acquisitions, Canva has expanded its portfolio to five companies in the last two years, following previous acquisitions of Affinity and Leonardo in 2024, as well as MagicBrief the year prior.

Canva refutes claims of job cuts as Leonardo.AI joins the company.

A sunlit, airy Canva design studio with the bold Canva logo painted on an exposed brick wall. A small team of adult designers gather around a large monitor where the Leonardo.AI logo slides seamlessly into the Canva branding like two interlocking puzzle pieces. Desks strewn with colorful mockups, notebooks and potted plants basking in warm natural light, conveying a mood of optimistic collaboration and no layoffs.

Canva denies job losses as Leonardo.AI comes into the fold - Startup Daily

- Canva denies reports of job losses from integrating Leonardo.AI into its platform.  
- Leonardo's team members will transition to roles across Canva's AI product teams.  
- Canva continues to acquire startups, enhancing its AI capabilities and product offerings.

If you like this article, share it with your friends.

Canva assures no job losses as it merges Leonardo.AI into its platform

Canva says no people will lose their jobs as integrating its prize artificial intelligence acquisition, Leonardo.AI, into the design platform, refuting media reports of potential job losses as inaccurate.

The Australia reported that Canva chief operating officer and cofounder Cliff Obrecht told Leonardo’s 150 North Sydney-based staff in a video call last week that the business was being restructured.

Obrecht reportedly told the Leonardo employees that while many in the Leonardo team would “directly map” to existing AI efforts, others were encouraged to look at other roles in the business.

Canva’s Head of Corporate and Internal Communications Lachlan told Startup Daily that the business is not exploring any redundancies and no job losses were discussed or communicated to the team, saying suggestions otherwise are “categorically false”, taking a swipe at The Australian for failing to check with the company prior to publication.

“It is disappointing that The Australian did not seek our response before publishing, had they done so, we would have unequivocally clarified the facts,” he said.

Get the best of Startup Daily straight to your inbox Want to know the latest in startup news? Subscribe to our daily news and analysis coverage on what’s happening to ANZ startups, investors and the broader ecosystem. And best of all, it's FREE! By continuing, you agree to our Terms & Conditions and Privacy Policy.

“We provided a clear on-the-record statement, which was not included in their reporting.”

The report came amid heightened sensibilities around AI and tech jobs after US fintech Block announced it was cutting 40% of its workforce, 4000 jobs, and Sydney logistics software platform WiseTech Global announced 2000 coding jobs would go, with both companies citing AI as the reason.

Doubling down on AI

Canva said the business was accelerating its planned integration of Leonardo “to double down on our AI ambitions” with a small team continuing to work on Leonardo as a standalone product while others join the mother ship as part of 100-plus AI product and research teams.

Leonardo remained a standalone business for the past 18 months since Canva acquired the Sydney generative AI startup in 2024 in a deal worth more than $300 million. Its tech is already integrated into key parts of the Canva product.

“Leonardo has proven itself a global leader in image and video generation, which is foundational to what we’re building at scale for Canva’s 265 million users,” a spokesperson said.

“As part of this phase of the integration, team members across Leonardo will transition into new teams across Canva.

“This is simply about better integrating their world-class talent to work across Canva’s broader suite of AI products at even greater scale. At the same time, we’ll continue investing in Leonardo as a standalone product, with a dedicated team focused on serving and growing its existing community.”

Last week Obrecht also announced Canva was acquiring two more startups, Mango.AI and Cavalry, from the US and UK respectively, as well as adding a “chief algorithms officer” to its C-suite.

Mango.ai is 10-month-old San Francisco marketing algorithm startup that was still in stealth mode. Cavalry is UK-based 2D animation platform used by motion designers.

The acquisitions take the number of companies Canva’s acquired in the last two years to five, following on from Affinity and Leonardo in 2024, and MagicBrief last year.







Canva denies job losses as Leonardo.AI comes into the fold

In recent years, venture capitalists have been investing heavily in artificial intelligence companies, reflecting the technology's growing influence in Silicon Valley and across the globe. However, not every AI startup is capturing the interest of investors.

Although many companies are rebranding with "AI" in their titles, certain startup concepts are losing favor among investors. A discussion with venture capitalists revealed trends regarding which AI software-as-a-service (SaaS) ventures are currently undesirable.

Aaron Holiday, a managing partner at 645 Ventures, noted that investors are keen on startups focused on creating AI-native infrastructure, specialized vertical SaaS with unique data, systems that facilitate user tasks, and platforms integrated into essential workflows. Conversely, investors are becoming less interested in companies that merely add thin layers to existing workflows, offer generic horizontal tools, provide lightweight product management solutions, or deliver surface-level analytics—essentially, services that could be automated by AI agents.

Abdul Abdirahman from F Prime pointed out that vertical software lacking proprietary data advantages is no longer appealing. Igor Ryabenky, managing partner at AltaIR Capital, emphasized the need for significant product differentiation. He argued that if a company’s unique offerings are primarily in user interface improvements and automation, they may struggle to succeed in the current market. 

"The barrier to entry has lowered, making it increasingly difficult to establish a strong market presence," he explained. New entrants should prioritize genuine ownership of workflows and a clear comprehension of problems from the outset. Instead of relying on extensive codebases, speed, focus, and flexibility in pricing strategies—away from rigid per-seat models—are essential for success today.

Jake Saper, general partner at Emergence Capital, pointed out trends in workflow ownership. He identified distinctions between Cursor and Claude Code, stating that ownership of the developer's workflow is becoming increasingly important, as many developers prefer execution capabilities over a detailed process. 

Saper warned that products focusing on "workflow stickiness" face challenges as agents take over tasks historically handled by humans. He remarked that previously, getting users to engage with a product was a significant advantage; however, if AI agents can perform tasks, human-driven workflows lose their significance.

Furthermore, integration solutions are becoming less desirable as technologies like Anthropic's Model Context Protocol (MCP) streamline connections between AI models and external systems, reducing the necessity for downloading multiple integrations or creating custom solutions. 

Transitioning into the future, Abdirahman noted that tools for workflow automation and task management will likely diminish as agents increasingly handle those tasks, referencing the performance of public SaaS companies that are losing footholds as AI-native competitors emerge with superior technologies.

Ryabenky highlighted that SaaS companies struggling to attract funding are often those whose offerings are easily replicated. Generic productivity software, basic customer relationship management systems, and shallow AI integrations face risks because strong AI-focused teams can quickly recreate them.

Ultimately, investors are inclined toward SaaS solutions that possess depth, expertise, and integration into crucial workflows. Focusing on deep AI integration into products and adjusting marketing strategies to highlight these advancements will be critical. Ryabenky concluded that capital is shifting towards businesses that have control over workflows, data, and specialized knowledge while moving away from easily copied offerings.

Investors reveal their disinterest in certain aspects of AI SaaS companies.

A dark, moody boardroom scene: on a polished black table a glass jar labeled “AI SaaS Hype” lies overturned, spilling folders stamped “Buzzwords,” “Unproven ROI,” “Generic Features.” Around it sit sharply dressed investors with arms folded and skeptical expressions. Soft natural light streams through a narrow window, casting dramatic shadows over the spilled papers.

Investors spill what they aren’t looking for anymore in AI SaaS companies - TechCrunch

- Investors are focusing on AI startups with proprietary data and deep workflow integration.  
- Generic tools and superficial AI layers are no longer enticing for venture capitalists.  
- Companies must demonstrate real problem-solving capabilities to attract investor interest today.  

Investors have been pouring billions into AI companies over the past few years, as the technology continues to hold sway in the Valley and thus the world. But not all AI companies are grabbing investor attention.

Indeed, even as it seems every company these days is rebranding to include “AI” in its name, some startup ideas are just no longer in favor with investors. TechCrunch spoke with VCs to learn what investors aren’t looking for in AI software-as-a-service startups anymore.

Popular SaaS categories for investors now include startups building AI-native infrastructure, vertical SaaS with proprietary data, systems of action (those helping users complete tasks), and platforms deeply embedded in mission-critical workflows, according to Aaron Holiday, a managing partner at 645 Ventures.

But he also gave a list of companies that are considered quite boring to investors these days: Startups building thin workflow layers, generic horizontal tools, light product management, and surface-level analytics — basically, anything an AI agent can now do.

Abdul Abdirahman, an investor at the firm F Prime, added that generic vertical software “without proprietary data moats” is no longer popular, and Igor Ryabenky, a founder and managing partner at AltaIR Capital, went deeper on that point. He said investors aren’t interested in anything, really, that doesn’t have much product depth.

“If your differentiation lives mostly in UI [user interface] and automation, that’s no longer enough,” he said. “The barrier to entry has dropped, which makes building a real moat much harder.”

New companies entering the market now need to build around “real workflow ownership and a clear understanding of the problem from day one,” he said. “Massive codebases are no longer an advantage. What matters more is speed, focus, and the ability to adapt quickly. Pricing also needs to be flexible: rigid per-seat models will be harder to defend, while consumption-based models make more sense in this environment.”

Techcrunch event Disrupt 2026: The tech ecosystem, all in one room Your next round. Your next hire. Your next breakout opportunity. Find it at TechCrunch Disrupt 2026, where 10,000+ founders, investors, and tech leaders gather for three days of 250+ tactical sessions, powerful introductions, and market-defining innovation. Register now to save up to $400. Save up to $300 or 30% to TechCrunch Founder Summit 1,000+ founders and investors come together at TechCrunch Founder Summit 2026 for a full day focused on growth, execution, and real-world scaling. Learn from founders and investors who have shaped the industry. Connect with peers navigating similar growth stages. Walk away with tactics you can apply immediately



Offer ends March 13. San Francisco, CA | REGISTER NOW

Jake Saper, a general partner at Emergence Capital, also had thoughts on ownership. To him, the differences between Cursor and Claude Code are the “canary in the coal mine.”

“One owns the developer’s workflow, the other just executes the task,” Saper continued. “Developers are increasingly choosing the execution over process.”

He said any product dealing with “workflow stickiness” — meaning trying to attract as many human customers as possible to continuously use the product — might find themselves in an uphill battle as agents takeover the workflow.

“Pre-Claude, getting humans to do their jobs inside your software was a powerful moat, but if agents are doing the work, who cares about human workflow?” he told TechCrunch.

He also thinks integrations are becoming less popular, especially as Anthropic’s model context protocol (MCP) makes it easier than ever to connect AI models to external data and systems. This means someone doesn’t need to download multiple integrations or build their own customer integrations; they can just use the MCP.

“Being the connector used to be a moat,” Saper said. “Soon, it’ll be a utility.”

Also, no longer en vogue include the “workflow automation and task management tools that enable the coordination of human work become less necessary if, over time, agents just execute the tasks,” Abdirahman said, citing examples, mainly public SaaS companies whose stocks are down as new AI-native startups arise with better, more efficient technology.

Ryabenky said the SaaS companies struggling to raise right now are the ones that can easily be replicated, he said.

“Generic productivity tools, project management software, basic CRM clones, and thin AI wrappers built on top of existing APIs fall into this category,” he said. “If the product is mostly an interface layer without deep integration, proprietary data, or embedded process knowledge, strong AI-native teams can rebuild it quickly. That is what makes investors cautious.”

Overa, what remains attractive about SaaS is depth and expertise, with tools embedded in critical workflows. He said companies should right now look into integrating AI deeply into their products and update their marketing to reflect that, Ryabenky continued.

“Investors are reallocating capital toward businesses that own workflows, data, and domain expertise,” Ryabenky said. “And away from products that can be copied without much effort.”

Investors spill what they aren’t looking for anymore in AI SaaS companies

On Friday, President Donald Trump announced his decision to halt government collaboration with Anthropic, a prominent artificial intelligence startup. The Pentagon plans to categorize the company as a supply-chain risk, significantly impacting the lab following a recent confrontation regarding technology regulations.

Trump specified that a six-month wind-down period would be implemented for the Department of Defense and other government entities that utilize Anthropic’s products. He warned that if the company does not assist with the transition, he would employ “the Full Power of the Presidency” to ensure compliance, which could lead to serious civil and criminal repercussions.

These developments represent a remarkable condemnation from the U.S. government towards Anthropic, a key player in the nation’s efforts to maintain a competitive edge in AI-related national security. This move potentially places Anthropic in a contentious position, likening it to suppliers deemed adversarial by Washington. Notably, Alphabet's Google and Amazon.com are among the financial supporters of Anthropic.

Trump instructs U.S. agencies to reject Anthropic's AI after Pentagon identifies startup as a supply threat.

Dark, tense scene in a shadowy Pentagon hallway: a silhouetted Donald Trump stands at the door, sharply pointing as two stern military officers drag away a heavy wooden crate stamped with the Anthropic logo; a U.S. flag hangs quietly in the gloom, pale moonlight filtering through a high window.

Trump directs U.S. agencies to toss Anthropic's AI as Pentagon calls startup a supply risk - The Japan Times

- Trump directs the government to cease collaboration with AI startup Anthropic.  
- The Pentagon will classify Anthropic as a supply-chain risk amid technology concerns.  
- A six-month phase-out period is set, with possible severe consequences for non-compliance.

U.S. President Donald Trump said Friday he is directing the government to stop work with Anthropic, and the Pentagon said it would declare the startup a supply-chain risk, dealing a major blow to the artificial intelligence lab after a showdown about technology guardrails.

Trump added there would be a six-month phase-out for the Defense Department and other agencies that use the company's products. If Anthropic does not help with the transition, Trump said, he would use "the Full Power of the Presidency to make them comply, with major civil and criminal consequences to follow."

The actions mark an extraordinary rebuke by the ​United States against one of the premier companies that has kept it in the lead on national security-critical AI, threatening to give Anthropic a pariah status that Washington until now had reserved for ‌enemy suppliers. ‌Alphabet's Google and Amazon.com are among Anthropic's financial backers.

Trump directs U.S. agencies to toss Anthropic's AI as Pentagon calls startup a supply risk

In Cambridge, Massachusetts, Mikey Shulman, the CEO of Suno, engages in a lively music creation session at their innovative studio. Accompanied by a research scientist, they are experimenting with AI to develop a new song. The duo inputs phrases like "Afrobeat, flute, drums, 90 beats per minute," and voilà – a vibrant rhythm resonates through their 19th-century office space. This process resembles the enchanting experiences many have had with platforms like ChatGPT and text-to-image generators, allowing users to craft music without prior instrumentation skills or extensive practice.

However, the emergence of AI-generated music has sparked controversy within the music industry. Traditional record labels have expressed their frustration, filing lawsuits against Suno and competitor Udio, contending that these startups are infringing on copyright by leveraging the legacies of established musicians. In light of thousands of AI-generated tracks popping up on platforms like Spotify, Suno and Udio are actively seeking to collaborate with music industry giants and establish a sustainable presence.

Shulman, who co-founded Suno in 2022, emphasized the importance of partnership with the music industry rather than opposition. "Music is so culturally significant that we must integrate AI into the existing musical landscape," he said. In 2024, major labels Sony Music, Universal Music, and Warner Records accused both companies of copyright violations for allegedly misusing their artists' work. While Suno has reached a $2.45 billion valuation and settled with Warner, only Udio has managed to secure licensing deals with Universal, Merlin, and Warner, while struggling with ongoing legal battles with Sony and other European entities representing musicians.

Udio, which began operations in 2023 with a smaller user base, experienced a backlash when users were prohibited from downloading their AI-generated creations after striking licensing agreements. Despite this setback, Udio CEO Andrew Sanchez remains hopeful, focusing on creating avenues for fans to engage with willing artists' music using AI.

Skepticism towards AI-generated music is prevalent among established musicians. Tift Merritt, a prominent songwriter and Artists Rights Alliance co-chair, recently spearheaded a campaign highlighting the need for AI companies to respect copyright laws and secure licensing agreements rather than sidestepping them. Merritt criticized the current exploitation of musicians’ intellectual property, stating that AI developments prioritize technological advancement over artistic integrity.

Shulman has acknowledged prior missteps in addressing these concerns, clarifying that his intention was never to dismiss the joys of music creation. He admitted that his earlier comments about music not being enjoyable for most often tied to the challenges of achieving perfection in production.

Meanwhile, Sanchez aims to present Udio as an alternative that prioritizes collaboration with artists. He understands that not all musicians will welcome AI, but he hopes to demonstrate that their innovations can facilitate creativity rather than undermine it. 

Outside of these startups, independent creators like Christopher “Topher” Townsend leverage AI for their music production. Townsend, a self-sufficient artist, uses various AI tools to compose and market his work. He acknowledges the fears some musicians have regarding AI but remains optimistic about its potential to democratize music creation.

Jonathan Wyner, a music production professor at the Berklee College of Music, views generative AI as a valuable tool that can enhance the music-making experience. He believes AI can streamline processes and enable unique musical possibilities for aspiring creators. However, many artists remain cautious, concerned about how record labels will negotiate their relationships with AI companies, especially in ways that may favor major artists over independent ones.

This conversation continues to unfold, with many artists urging the music community to differentiate between responsible AI use and companies like Suno, which they view as exploiting artists’ work without proper consideration for copyright.

AI song generator startups Suno and Udio, previously at odds with the music industry, now aim to become part of it.

A shadowy backstage in a grand concert hall: on opposite walls, two oversized neon logos – Suno in warm amber light, Udio in deep crimson – glow over a stage strewn with shattered vinyl records and torn sheet music swirling into smoky dark clouds. At the edge of the stage, silhouetted music executives in sharp suits stand with arms crossed, glaring up at an empty microphone bathed in a single stark spotlight. Moody red and purple ambient light casts long, dramatic shadows across the scene.

AI song generator startups Suno and Udio angered the music industry. Now they're hoping to join it - AP News

- Suno and Udio create AI-generated music, facing legal challenges from major record labels.  
- Both startups aim to negotiate partnerships with the music industry for future cooperation.  
- Artists express concerns over copyright infringement and exploitation of their intellectual property rights.  

Add AP News as your preferred source to see more of our stories on Google.

Add AP News on Google Add AP News as your preferred source to see more of our stories on Google. Share

CAMBRIDGE, Mass. (AP) — Suno CEO Mikey Shulman pulls up a chair to the recording studio desk where a research scientist at his artificial intelligence company is creating a new song.

The flute line sounds promising.

The percussion needs work.

Neither of them is playing an instrument. They type some descriptive words – Afrobeat, flute, drums, 90 beats per minute – and out comes an infectious rhythm that livens up the 19th century office building where Suno is headquartered in Cambridge, Massachusetts. They toggle some editing tools to refine the new track.

Much like early experiences with ChatGPT or AI text-to-image generators, trying to make an AI-generated song on platforms like Suno or its rival, Udio, can seem a little like magic. It takes no musical skills, practice or emotional wellspring to conjure up a new tune inspired by almost any of the world’s musical traditions.

But the process of training AI on beloved musicians of the past and present to produce synthetic approximations of their work has angered the music industry and brought much of its legal power against the two startups.

Now, after their users have flooded the internet with millions of AI-generated songs, some of which have found themselves on streaming services like Spotify, the leaders of Suno and New York-based Udio are trying to negotiate with record labels to secure a foothold in an industry that shunned them.

“We have always thought that working together with the music industry instead of against the music industry is the only way that this works,” said Shulman, who co-founded Suno in 2022. “Music is so culturally important that it doesn’t make sense to have an AI world and a non-AI world of music.”

Sony Music, Universal Music and Warner Records sued the two startups for copyright infringement in 2024, alleging that they were exploiting the recorded works of their artists.

Since then, the pair have strived to make peace with the industry. Suno, now valued at $2.45 billion, last year struck a settlement with Warner, and Udio has signed licensing agreements with Warner, Universal and independent label Merlin. Only one major label, Sony, has not settled with either startup as the lawsuits move forward in Boston and New York federal courts. Suno also faces legal challenges in Europe brought by groups representing music creators.

The first of the settlement deals, between Udio and Universal, led to an exodus of frustrated Udio users who were blocked from downloading their own AI-generated tracks. But Udio CEO Andrew Sanchez said he’s optimistic about what the future will bring as his company adapts its business model to let fans of willing artists use AI to play with and potentially alter their works.

Udio CEO Andrew Sanchez is photographed in the Tribeca neighborhood of New York, Wednesday, Jan. 21, 2026. (AP Photo/Richard Drew) Udio CEO Andrew Sanchez is photographed in the Tribeca neighborhood of New York, Wednesday, Jan. 21, 2026. (AP Photo/Richard Drew) Add AP News on Google Add AP News as your preferred source to see more of our stories on Google. Share Share Facebook

Copy Link copied

Print

Email

X

LinkedIn

Bluesky

Flipboard

Pinterest

Reddit Read More

“Having a close relationship with the music industry is elemental to us,” Sanchez said in an interview. “Users really want to have an anchor to their favorite artists. They want to have an anchor to their favorite songs.”

Many professional musicians are skeptical. Singer-songwriter Tift Merritt, co-chair of the Artists Rights Alliance, recently helped organize a “Stealing Isn’t Innovation” campaign by artists — including Cyndi Lauper and Bonnie Raitt — to urge AI companies to pursue licensing deals and partnerships rather than build platforms without regard for copyright law.

“The economy of AI music is built totally on the intellectual property, globally, of musicians everywhere without transparency, consent, or payment. So, I know they value their intellectual property, but ours has been consumed in order to replace us,” Merritt said in an interview in Raleigh, North Carolina.

Shulman contends technology “evolves very often faster than the law,” and his company tries to be thoughtful about “not breaking the law” but also “deliver products that the world really wants.”

Suno CEO doesn’t really think ‘people don’t enjoy’ making music

When the music industry first confronted Suno over alleged copyright infringement, the company’s antagonistic response alienated professionals like Merritt.

Symbolizing the divide was a clip last year in which Shulman was quoted as saying, “it’s not really enjoyable” to make music most of the time. Shulman started learning piano at age 4 but later dropped it. He took up bass guitar at 12, playing in rock bands in high school and college. He said that experience gave him some of the best moments of his life.

Mikey Shulman, one of the founders of AI music generator startup Suno, speaks to a reporter at company headquarters, Tuesday, Jan. 13, 2026, in Cambridge, Mass. (AP Photo/Robert F. Bukaty) Mikey Shulman, one of the founders of AI music generator startup Suno, speaks to a reporter at company headquarters, Tuesday, Jan. 13, 2026, in Cambridge, Mass. (AP Photo/Robert F. Bukaty) Add AP News on Google Add AP News as your preferred source to see more of our stories on Google. Share Share Facebook

Copy Link copied

Print

Email

X

LinkedIn

Bluesky

Flipboard

Pinterest

Reddit Read More

“You need to get really good at an instrument or really good at a piece of production software,” Shulman said on the “The Twenty Minute VC” podcast. “I think the majority of people don’t enjoy the majority of the time they spend making music.”

“Clearly, I wish I had said different words,” Shulman told the AP. The context, he added, was that “to produce perfect music takes a lot of repetitions and not all of those minutes are the most enjoyable bits of making music. On the whole, obviously, music is amazing. I play music every day for fun.”

Udio CEO pitches his company as the friendly alternative

Sanchez, the Udio CEO, also loves making music. He’s an opera-loving tenor who’s sung in choirs and grew up crooning Luciano Pavarotti in his family’s home in Buffalo, New York.

Founded in 2023 by a group that included several AI researchers from Google, the startup now employs about 25 people. It has fewer users and raised less venture capital than Suno, which likely gave Udio a stronger incentive to be first to settle with record labels, said copyright lawyer Brandon Butler.

“A service (like Suno) that gets more venture backing is in some sense hungrier to find revenue streams and more on the hook to all those backers to make sure that they achieve profitability, which would make settling and compromising less attractive,” said Butler, director of the copyright advocacy group Re:Create. “Whereas a company with fewer backers, with less capital, with less access would be weaker and less able to resist the risk that they’re incurring by being involved in litigation.”

Still, Udio embraces its underdog status.

“So many tech companies actively cultivate this I-am-a-tech-company-crusader and that’s part of their identity,” Sanchez said. “That alienates people who are creative and I am uniformly opposed to that.”

Sanchez said he knows not every artist is going to embrace AI, but he hopes those who leave the room after talking with him realize he’s not imposing a kind of “AI bravado.”

“If you took what we’re doing and pretended that the word AI wasn’t a part of it, people would be like, ‘Oh my gosh. This is so cool.’”

Some see potential in AI-assisted music creation

In the basement office of his Philadelphia, Mississippi home, Christopher “Topher” Townsend is a one-man band, making and marketing Billboard-chart-topping gospel music — none of which he sings himself — and doing it in record time.

The rapper, whose lyrics reflect his political conservatism, downloaded Suno in October and, within days, created Solomon Ray, a fictional singer that Townsend calls an extension of himself.

Townsend uses ChatGPT to write lyrics, Suno to generate songs and other AI tools to create cover art and promotional videos under the Solomon Ray name.

“I can see why artists would be afraid,” Townsend said. ”(Solomon Ray) has an immaculate voice. He doesn’t get sick. You know, he doesn’t have to take leave, he doesn’t get injured and he can work faster than I can work.”

Trying to dispel that fear for aspiring artists is Jonathan Wyner, a professor of music production and engineering at the Berklee College of Music in Boston, who sees generative AI as just another tool.

“To the creative musician, AI represents both enormous potential benefits in terms of streamlining things and frankly making kinds of music-making possible that weren’t possible before, and making it more accessible to people who want to make music,” he said.

Such a vision remains a tough sell for artists who feel their work has already been exploited. Merritt says she’s particularly concerned about labels making deals with AI companies that leave out independent artists. An open letter she co-signed this week says “many in our community are embracing responsible AI as a tool for creation” but targets Suno as a “smash and grab” business that artists should avoid.

“Artists need to know the difference – all AI platforms are not the same, and Suno, which is being sued for copyright infringement, is not a platform artists should trust,” says the letter from Merritt and six others.

Singer and songwriter Tift Merritt, co-chair of the Artist Rights Alliance, poses for a photo with her guitars at her home in Raleigh, N.C., Thursday, Jan. 29, 2026. (AP Photo/Allen G. Breed) Singer and songwriter Tift Merritt, co-chair of the Artist Rights Alliance, poses for a photo with her guitars at her home in Raleigh, N.C., Thursday, Jan. 29, 2026. (AP Photo/Allen G. Breed) Add AP News on Google Add AP News as your preferred source to see more of our stories on Google. Share Share Facebook

Copy Link copied

Print

Email

X

LinkedIn

Bluesky

Flipboard

Pinterest

Reddit Read More

—————-

O’Brien reported from Cambridge, Massachusetts and New York. Ngowi reported from Cambridge and Somerville, Massachusetts. AP journalists Sophie Bates in Philadelphia, Mississippi and Allen G. Breed in Raleigh, North Carolina, contributed to this report.

Anthropic's CEO remains resolute as Pentagon deadline approaches.

More from AI Startups & Investments

AI Startups Face ROI Evaluation: Insights from 4 VCs on Current Priorities

There is a lack of effective strategies for the collaboration between AI companies and the government.

Investors Increase Investments in the Agent Economy

Mark Cuban participates in panel for Pittsburgh AI startup amid NFL Draft week