OurCrowd: “2025 has been a breakthrough year for us in AI”

“At OurCrowd, we believe we’re only at the beginning of understanding the full impact of AI, not just as a set of technologies, but as a foundational shift in how industries operate, value is created, and startups are built,” said Guy Dassa, AI expert and investment partner at OurCrowd. “Israel has a unique opportunity to lead globally, particularly where AI converges with traditional strengths like cybersecurity, defense, and deep tech. But to maintain that edge, we need continued investment in infrastructure, talent development, and responsible innovation.”

Dassa joined CTech for its VC AI Survey to share its strategies and vision for the future of Israel and the investment space as the world adopts AI.

“We’re committed to backing bold founders who are building for the long term, tackling real problems, creating defensible moats, and driving change in sectors that matter,” he added. “The next wave of category-defining companies will come from those who blend technical excellence with commercial clarity and human insight.”

You can learn more in the interview below.

Fund ID
Name and Title: Guy Dassa, Investment Partner
Fund Name: OurCrowd
Founding Team: OurCrowd/ Jon Medved
Founding Year: 2013
Investment Stage: Agnostic
Investment Sectors: Israeli deal flow (Generalist).

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?

9 - We are using AI across virtually every aspect of our operations at the fund. It’s not just a tool we’re experimenting with; it has become a core enabler of how we work. From sourcing and screening new deals, where AI helps us parse massive volumes of data to identify promising startups and emerging trends, to streamlining our due diligence processes with faster document analysis, risk flagging, and market intelligence. AI is accelerating our speed and sharpening our insights.

On the portfolio management side, we’re leveraging AI for everything from monitoring performance signals to automating elements of reporting and investor communications. It’s also helping us support our portfolio companies more effectively, by providing AI-driven benchmarks, market research, and even co-developing go-to-market strategies in some cases.

Even on the communications front, from internal knowledge sharing to content creation for our newsletters, thought leadership, and investor updates, AI has added a layer of efficiency and scale we didn’t have before. We see AI not as a siloed initiative but as a horizontal capability woven into the daily workflows of our partners, analysts, and operations teams.

It should be noted that we are using enterprise license and keep confidential information secured.

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

Yes. 2025 has been a breakthrough year for us in AI. We had allocations in major access deals (directly on the cap table or through a Special Purpose Vehicle – SPV) like Groq and Databricks. We’ve had 11 exits from AI-driven companies since inception, underscoring the strength of our early conviction in the space and our focus on backing both infrastructure and application-layer innovation.

One of our standout performers was Anthropic, where we participated before the recent considerable upticks and saw tremendous growth and strategic interest over the past year as demand for frontier AI models accelerated. We also made investments in foundational players like OpenAI and Databricks, both of which continue to shape the AI landscape globally. Beyond that, we backed innovators like SSI, and a range of vertical AI startups tackling real-world challenges in sectors such as enterprise software, cybersecurity, and healthcare.

The common thread among these companies was a clear, defensible value proposition: either through proprietary models, unique datasets, or specialized domain knowledge. These were not generic AI plays — they built robust, scalable technologies with clear commercial pathways and strong teams capable of navigating a rapidly evolving market.

Many of these exits were driven by strategic acquisitions or secondary transactions as major corporates and late-stage investors raced to integrate or access best-in-class AI capabilities. For us, it validates the thesis that AI isn’t just a sector, it’s a foundational layer that will underpin nearly every future business model.

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

Yes. Evaluating AI startups often requires a fundamentally different lens compared to more traditional sectors. One of the key distinctions is the need to separate the infrastructure layer, such as large language models (LLMs) or foundational AI platforms from the application layer, where companies are building specific use cases or vertical solutions on top of that infrastructure.

While traditional startup evaluation focuses on traction, TAM, and defensible business models, AI investing often demands deeper technical due diligence: understanding how the model is trained, whether the data is proprietary or open, what kind of fine-tuning or reinforcement learning is being used, and how scalable and differentiated the approach really is.

There’s also the challenge of assessing “moats.” Many AI tools can appear impressive in demos but lack long-term defensibility unless they’re built on unique datasets, solve highly specific problems, or embed AI in a way that makes the product indispensable.

Moreover, timing and hype cycles can distort early signals. So, our AI evaluation process requires not only technical validation, but a clear understanding of how the product performs in the real world, how it integrates into workflows, and whether it’s truly delivering ROI, not just impressive outputs.

In short, while the investment principles remain the same: strong teams, clear market need, scalable business model - AI startups demand an added layer of scrutiny that bridges deep tech expertise with commercial viability.

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?

When evaluating AI companies, we always look at the standard financial KPIs: revenue growth, gross margins, customer acquisition cost (CAC), lifetime value (LTV), and burn rate. However, when it comes to AI, much of our analysis is sector-specific, as the most relevant KPIs often depend on the industry vertical and the type of AI application being deployed.

For example, in enterprise AI, we pay close attention to metrics like average contract value (ACV), sales cycles, and churn rates, particularly whether the AI product becomes embedded in customer workflows or is at risk of being replaced by in-house solutions or competitors.

In healthtech AI, clinical validation, regulatory progress, and time to reimbursement are often more meaningful than early revenue.

In deep tech AI (e.g., foundational models, LLMs, custom infrastructure), commercial KPIs may still be immature, so we look at the quality and uniqueness of the underlying models, data strategy, compute costs, and potential for API monetization.

We also assess AI-specific indicators such as:

• Data advantage – Does the company have access to proprietary or hard-to-replicate data?
• Model performance – Are they showing benchmark-beating results (e.g., accuracy, latency, precision/recall) in a real-world context?
• Human-in-the-loop requirements – How scalable is the solution without constant manual oversight?
• AI cost-efficiency – How well are they managing compute resources relative to revenue or usage?

Ultimately, we tailor our KPI analysis to the company's stage, vertical, and business model. In AI, one-size-fits-all metrics rarely apply. Context is everything.

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

At the early stage, particularly pre-seed and seed, AI startup valuations tend to fall within a relatively common range, shaped by market norms, geography, and investor appetite. In these cases, we benchmark against comparable rounds and apply a standard range based on team strength, TAM, and initial traction.

That said, AI companies with a truly differentiated value proposition - whether through proprietary data, exceptional technical talent, early signs of product-market fit, or validation from enterprise pilots - may justify a valuation premium. We are looking for companies that rise above what we call a “thin layer over GPT: (or any other LLM for that matter). In such cases, we’re willing to stretch beyond typical multiples, especially if we believe the company is building infrastructure or capabilities that will become foundational over time.

We also factor in the cost of building the tech, particularly in compute-heavy models or companies working at the infrastructure layer of AI, where the capital requirements are higher, but the strategic value is also elevated.

Ultimately, even at the earliest stages, we’re looking for indicators that the company isn’t just chasing the hype but is solving a real problem with a scalable AI solution, backed by a team capable of executing. If those signals are strong, valuation becomes a function of long-term conviction, not short-term revenue.

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

All of the above and then some. AI investing carries several amplified financial risks that go beyond typical early-stage tech considerations.

First, infrastructure costs can be prohibitively high, especially for companies training or fine-tuning large models. Cloud compute, GPU access, and ongoing model optimization can create significant burn without corresponding near-term revenue.

Second, data dependencies are a major risk. If a startup relies heavily on access to third-party data, whether for training, inference, or product functionality, changes in licensing terms, API pricing, or regulation (like GDPR, HIPAA, or the upcoming EU AI Act) can severely impact their cost structure or product viability.

Third, regulatory uncertainty is growing. AI companies working in sensitive domains like healthcare, finance, or public safety may face compliance burdens that delay go-to-market timelines or limit addressable markets. Even companies in less regulated sectors face reputational and operational risks from the broader debate around bias, transparency, and model explainability.

In addition, there’s always the risk that a large incumbent, like a hyperscaler or enterprise software giant enters the same space with more distribution, resources, and pre-trained models, making the startup’s offering obsolete. Similarly, the pace of innovation in GenAI means that a well-funded competitor can leapfrog a startup unless it has a meaningful moat, whether in data, IP, customer relationships, or vertical expertise.

These risks exist across the tech landscape, but in AI and GenAI, they are amplified by the velocity of change and the capital intensity required to compete. That’s why, at OurCrowd, we put significant focus on identifying durable competitive advantages early, not just great demos or hype-driven narratives.

Do you focus on particular subdomains within AI?

We invest across the full spectrum of AI subdomains, including machine learning, natural language processing (NLP), computer vision, and edge AI, but over the past 18–24 months, we’ve placed particular emphasis on generative AI (GenAI), which we see as a transformative force across multiple industries.

Our interest in GenAI spans both infrastructure and applications: from foundational model companies and tooling platforms to verticalized solutions in areas like enterprise automation, financial services, healthcare, and content creation.

We look for startups that don’t just plug into existing models, but add real value through proprietary data, workflow integration, and domain-specific innovation.

That said, we continue to back companies working in other core AI subdomains, particularly where machine learning and computer vision intersect with real-world challenges, such as robotics, cybersecurity, and advanced manufacturing. We’re also tracking emerging areas like multimodal AI and AI safety, where we expect significant growth and demand.

Ultimately, we believe the next generation of category-defining companies will emerge at the intersection of several of these subfields and our portfolio is built to reflect that convergence.

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

Absolutely! AI is already reshaping traditional industries, and the pace of change is only accelerating. There’s extensive research and coverage on this topic, and we’re seeing clear signals in our own portfolio and deal flow.

In general, industries that involve repetitive, rule-based, or data-heavy tasks are the most susceptible to near-term transformation. Tasks like basic analysis, screening, summarization, document processing, and routine content generation are increasingly being automated using AI, particularly generative models and NLP technologies. As a result, sectors like legal services, insurance, financial back-office operations, media, and customer support are seeing rapid AI-driven disruption.

By contrast, industries that rely on human nuance, creativity, or complex decision-making under uncertainty like various areas of medicine, law enforcement, and high-stakes strategy roles are less immediately exposed. That said, even in those fields, AI is becoming a powerful co-pilot rather than a replacement.

As the saying goes, “Human for the best, AI for the rest,” though the gap between “best” and “rest” is narrowing every day.

Technologies we view as especially transformative include:

• Generative AI for knowledge work and content automation
• Computer vision for manufacturing, agriculture, and autonomous systems
• Predictive analytics and ML for logistics, supply chain optimization, and finance
• Multimodal models that combine text, image, and audio for complex workflow

Ultimately, AI is not just a vertical, it’s a cross-cutting capability that will touch every sector. The question isn’t whether traditional industries will be affected, but how quickly and at what depth.

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?

The most compelling opportunity lies at the intersection of AI and cybersecurity, where Israel continues to lead globally. As threats become more sophisticated and automation becomes essential to defend against them, we’re seeing enormous innovation in AI-powered threat detection, predictive analytics, autonomous response systems, and fraud prevention.

Recent mega-deals speak for themselves:

• Wiz, the Israeli cloud security startup using AI to detect and prioritize risks, agreed to be acquired by Google for $32 Billion, making it not only the largest tech exit in Israeli history, but Google’s biggest acquisition ever.
• CyberArk, known for its identity and access management solutions, reached a $25 Billion acquisition agreement with Palo Alto Networks, Israel’s second-largest tech exit to date.

These historic deals show that Israeli AI–cyber companies aren’t just competitive, they’re category-defining. In fact, every 50 cents spent globally on cyber is invested in Israel, highlighting the country’s unmatched depth and momentum in this space.

Beyond cybersecurity, Israeli startups also excel in other high-potential AI niches:

• Defense and dual-use AI: Israel is producing cutting-edge autonomous systems, drone intelligence, and battlefield AI applications, with significant commercial and government interest.
• Critical infrastructure AI: Solutions for energy, water, and manufacturing that combine computer vision, sensor integration, and real-time optimization are gaining traction globally.
• Enterprise AI tooling: From data labelling and model compression to fine-tuning LLMs and privacy-preserving AI, Israel’s deep-tech DNA shines in core enabler technologies.
• Healthcare AI: While exits here tend to have longer horizons due to regulation, companies focused on diagnostics, medical imaging, and clinical automation are showing strong promise.

Israel’s edge comes from the fusion of elite technical talent, a military-born innovation culture, and a proven ability to scale real-world solutions fast. The convergence of GenAI with traditional Israeli strengths like cybersecurity and defense is set to fuel the next wave of global AI exits over the next five years.

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?

One notable gap we see is in the infrastructure layer of AI is the foundational tooling, platforms, and compute frameworks that support large-scale model development and deployment. These companies tend to be capex-heavy, which can make them harder to fund and scale in the early stages, especially in ecosystems like Israel that have traditionally focused more on lean, application-layer startups.

As a result, we believe Israel is currently underrepresented in the infrastructure space, an area with immense long-term value and exit potential.

When it comes to the types of founders we’re looking to back, we prioritize those with deep conviction and passion, especially entrepreneurs aiming to make a significant impact in sectors dominated by outdated legacy players. We’re particularly excited about founders targeting industries that are overdue for disruption, whether that’s insurance, logistics, manufacturing, construction, healthcare, or public sector services with AI-native solutions that are scalable and defensible.

We're also drawn to builders who aren’t just chasing hype or tooling around with off-the-shelf LLMs, but those who understand their end users deeply, can differentiate through data, and are capable of navigating both the technical and commercial complexity of scaling an AI-first company.

In short, we’re looking for mission-driven founders with a clear thesis, a real problem to solve, and the grit to go deep in complex industries, especially where incumbents are slow to adapt and AI has the potential to unlock significant efficiency or value.