Why Catalyst is doubling down on ‘Anti-AI’ startups and vertical agents

“What makes AI business models especially exciting is their built-in feedback loop. Each deployment generates data that improves the product, enabling rapid iteration and scaling. We seek out companies that embrace this cycle and have the vision to grow it globally,” said Yair Shamir, Managing Partner; Edouard Cukierman, Managing Partner; and Ariel Anati, Partner, from Catalyst Private Equity. The firm joined CTech for its VC AI Survey to share insights on the technology and the sector.

Catalyst executives.

“In a way, we also explore a new market segment of ‘anti-AI’ companies – startups whose main purpose is to defend customers from AI companies or AI solutions,” they continued. “Maybe the market should explore such companies more in depth, as their growth and trajectory might be correlated with the current AI startups.”

You can read more below:

Fund ID
Name and Title: Yair Shamir, Managing Partner; Edouard Cukierman, Managing Partner; Ariel Anati, Partner
Fund Name: Catalyst Private Equity
Founding Team: Edouard Cukierman, Yair Shamir, Boaz Harel, Luc Muller, Ariel Anati
Founding Year: 1999
Investment Stage: Growth
Investment Sectors: AI, Cyber, Mobility

On a scale of 1 to 10, how has AI impacted your fund’s operations over the past year - specifically in terms of the day-to-day work of the fund's partners and team members?

6 - Hopefully by next year it will be 8 or even 9... AI has significantly enhanced our operational efficiency at Catalyst, particularly in deal sourcing and supporting our portfolio companies. Our team now leverages AI-powered tools for market analysis, competitive intelligence, and due diligence processes, which has accelerated our decision-making timeline by approximately 25%-30%. Every team member even uses the newest AI browsers on a daily basis. As a fund that has been operating for over 26 years through multiple crises, we've always emphasized the importance of data-driven decisions, and AI has amplified this capability. But that said, we don’t feel (yet) that the AI revolution has started in its full capacity in the fund management world, we believe we'll get there this year, when AI agents will be an integral part of our day-to-day operations.

Most notably and in parallel to our fund activity, our launch of the Catalyst Investors' Club platform in December 2022 incorporates AI-driven video analysis and investor matching algorithms, which has revolutionized how we democratize access to Israeli tech investments. This platform exemplifies how we're using AI not just internally, but as a core component of our investment strategy evolution.

Have you already had any significant exits from AI companies? If so, what were the key characteristics of those companies?

Yes, but it depends on who you're asking. We've had several notable AI exits, even before the term was so hyped. For example, in Mobileye, which represents one of the most successful vision AI companies globally. Additionally, our portfolio company Otorio, which develops AI-powered industrial cybersecurity solutions, was acquired by Armis just a few months ago.

At the end, the key characteristics of a successful AI exit is how the product provide you the most accurate and most relevant solution. A relevant market, the ability to create a new growing market and utilizing AI skills in the most efficient way (versus the competition).

Is identifying promising AI startups different from evaluating companies in your more traditional investment domains? If so, how does that difference manifest?

Well... evaluating AI startups is different from assessing traditional companies in the fact that most of the time (in a way) the exponential growth of AI companies even redefines the law of physics... Technical due diligence is far more complex, especially when you try to define between core model companies and wrappers companies (those that utilize AI models using APIs). We eventually look for teams that can demonstrate not just technical prowess but also the ability to navigate the complex regulatory landscape surrounding AI deployment.

Go-to-market strategies also differ; AI startups must first prove their technology’s accuracy and reliability before customers are willing to integrate it, unlike the more predictable sales cycles of traditional B2B companies. What makes AI business models especially exciting is their built-in feedback loop. Each deployment generates data that improves the product, enabling rapid iteration and scaling. We seek out companies that embrace this cycle and have the vision to grow it globally.

In a way, we also explore a new market segment of "anti-AI" companies – startups whose main purpose is to defend customers from AI companies or AI solutions. Maybe the market should explore such companies more in depth as their growth and trajectory might be correlated with the current AI startups.

What specific financial performance indicators (KPIs) do you examine when assessing a potential AI company? Are there any AI-specific metrics you consider particularly important?

The specific financial KPIs we prioritize really depend on a company’s stage and its AI vertical, but the backbone remains familiar: ARR, revenue growth, margins, and so on. What’s changed in the world of AI is not so much which metrics we track, but how we approach them. In traditional software or deep tech, it was easy to stack companies' side by side with standard benchmarks. With AI, we’ve learned that numbers must be complemented by context. Is that rapid growth coming from a genuinely differentiated AI advantage, or is it just the market’s current appetite for automation? Does high retention reflect the pain of switching out a deeply integrated AI product, or is it simply inertia? In AI, the "why" behind the metrics can matter as much as the metrics themselves.

Another pivotal difference is defensibility. When we see strong financial indicators, we’re now compelled to ask: how sustainable is this edge? Does the company have proprietary data that makes its growth repeatable, or are those metrics built on top of widely available third-party datasets? We’re also highly attuned to infrastructure efficiency, sometimes even modest revenue can signal outsized potential if the company is stretching every GPU cycle to the maximum. So, while the dashboard hasn’t changed, the interpretation has.

How do you approach the valuation of early-stage AI startups, which often lack significant revenues but possess strong technological potential?

As a fund focused on late-stage investments, we typically base valuations on cash flow projections (and potential growth). However, with early-stage AI startups where revenue is limited or projections are unreliable, we take a more qualitative approach. We assess factors such as the amount of capital needed to be raised, the strength and relevance of strategic partners, the maturity of the technology, and the leadership team’s capability. We pay close attention to whether the product offers a breakthrough solution or a novel approach to a persistent problem, especially if it can scale effectively. Ultimately, we value not just current performance, but the ecosystem a company is building, the credibility it has earned, and its potential to become a market leader through smart, impactful use of AI.

What financial risks do you associate with investing in AI companies, beyond the usual technological risks?

AI companies face significant financial risks that go beyond technological challenges. One of the most pressing is infrastructure cost, as training advanced models can require hundreds of thousands of dollars. Many companies misjudge how much they will spend on GPUs or cloud services, and we have seen promising startups fail simply because they ran out of funds for continued development and deployment. Regulatory compliance is another growing burden, especially for Israeli startups expanding into global markets. Adhering to General Data Protection Regulation (GDPR), industry regulations, and new AI laws like the EU AI Act can consume up to 30 percent of a startup’s budget and delay market entry if not managed well.

Other risks include dependency on external data sources. Companies relying on third-party datasets or APIs are vulnerable to price changes, usage limits, or loss of access, which can significantly affect their business model. Talent is another costly factor. In the end, strong financial management is just as vital as technical innovation for long-term success in AI.

Do you focus on particular subdomains within AI?

At Catalyst, we focus on three key AI subdomains where Israeli companies have proven global leadership. First is cybersecurity AI, which made up about 42% of all Israeli tech investments in 2024. Startups here, often rooted in military intelligence, excel in threat detection, behavioral analysis, and autonomous response, leveraging military-grade technology to stay ahead of global cyber threats.

Second, we prioritize computer vision, an area where Israeli firms lead in autonomous vehicles, industrial automation, and medical imaging. Our work with companies like Mobileye or Arbe shows how expertise in this field can drive transformative, high-value products with strong market adoption. We deliberately avoid consumer-facing generative AI and instead focus on B2B enterprise solutions, where Israel’s technical depth and enterprise sales strength create more defensible and scalable business models. This strategy allows us to back companies with real competitive advantages and strong global growth potential, aligning with our goal of building durable, high-impact ventures.

How do you view AI’s impact on traditional industries? Are there specific AI technologies you believe will be especially transformative in certain sectors?

AI is reshaping traditional industries by the convergence of different technologies together. Enabling a shift from reactive processes to predictive, data-driven decision-making. Israeli companies are at the forefront of this shift, applying AI to sectors such as industrial automation, healthcare, agriculture, and finance. In manufacturing and logistics, companies like Nexar are using computer vision and predictive analytics to improve real-time operations and data training, especially in fleet management. In healthcare, Israeli startups are leading advancements in medical imaging, diagnostics, and drug discovery, thanks to deep collaborations with top medical institutions and military-derived medical expertise.

In agriculture, AI is enhancing Israel’s long-standing leadership in water management and crop optimization through precision farming tools that improve yields and predict diseases. In financial services, Israel’s strengths in cybersecurity are driving AI innovations in fraud detection, risk analysis, and anomaly detection. Across all sectors, the most transformative AI technologies are those that allow companies to predict and prevent issues before they arise, whether that means catching fraud early, detecting health conditions sooner, or anticipating equipment failures. This proactive, preventive approach is where Israeli AI innovation is uniquely positioned to deliver large-scale, lasting impact.

What specific AI trends in Israel do you see as having strong exit potential in the next five years? Are there niches where you believe Israeli startups particularly excel?

In the next five years, we see strong exit potential in several Israeli AI trends. Artificial Intelligence Agents, which are autonomous systems that manage complex business workflows, are gaining traction. We believe that vertical agents' solutions will be the core growth in Israeli AI trends. AI infrastructure and optimization are another key area, as rising compute costs drive demand for tools that enhance training efficiency, GPU utilization, and deployment scalability. Data processing will be a key, that's why we invested in Speedata, for example, a unique chip company that focuses on creating the next generation of chips for data analytics tasks.

AI security and governance is rapidly growing, with Israeli firms applying their cybersecurity expertise to ensure compliance, safety, and auditability in AI systems. Vertical AI applications in defense, healthcare, and industrial automation also show high exit potential, particularly where Israeli startups offer mission-critical, domain-specific solutions.

Are there gaps or missing segments in the Israeli AI landscape that you’ve identified? What types of AI founders are you especially looking to back right now in Israel?

While Israel leads in applied AI and B2B enterprise tools, there are clear gaps in foundational LLM development, consumer-facing applications, and AI hardware. Building core LLM infrastructure remains capital-intensive and out of reach for most Israeli startups, which instead focus on fine-tuning global models or niche enterprise use cases. Similarly, custom chips and edge-AI devices are still early-stage areas, offering long-term growth potential but lagging behind the country’s software strengths.

We’re especially interested in founders who combine deep technical AI expertise with hands-on industry knowledge, those who understand both the technology and the practical workflows of sectors like healthcare, manufacturing, and financial services. Military intelligence backgrounds, particularly from elite IDF units, remain a major asset, as do global perspectives from founders with international experience. We back those who blend innovation, discipline, and commercial relevance to build scalable, globally competitive companies. To summarize, we look for founders who are able to adapt to new technologies, especially in AI domains.