The AI startup nation: Meet five companies showing how Israel still leads in new tech

Israel has long been known as “the startup nation,” and when that comes to AI, it’s proving true once again.

There are more than 2,300 AI startups active in Israel today, according to a new report co-produced by Google Israel and RISE Israel, an independent technology think tank formerly known as SNPI. That number accounts for roughly one quarter of the technology companies in the country. Long known for having the highest number of startups per capita in the world, Israel's tech sector is rich with innovation and healthy competition, disrupts existing markets and creates new ones, and fosters an amazing community of technical talent.

Even so, Israel must continue to capitalize on the AI moment if it wants to hold onto the technology crown. In fact, the report notes that AI startups in Israel are not launching at the pace of previous startup eras and there may be an emerging shortage of talent. This underscores the need for the public and private sectors to take AI seriously, both as a market and within their organizations. Technology has given Israel an edge in the past, and it can be a similar advantage in the future.

The Israeli startup scene has always been strong across multiple sectors, never putting all its pomegranates in one basket. The same is true for AI, where across many different vertical sectors and technology categories, we are seeing AI-focused Israeli startups leading. To name just a few: Pinecone leads in vector databases; AI21 has mastered creative output and agents for enterprises data interaction; Lightricks knows all the tricks of creative output and text-to-video generation; AUI knows natural language processing for customer service chat; and Tabnine is a leader in code assistance.

An important step for ensuring even more growth in AI is for Israeli companies to work together to promote the importance of AI and identify opportunities where it can have the greatest impact. This means bringing knowledge, technology infrastructure, solutions, regulation, and human capital, to the table, so that everyone can benefit. The RISE Israel report also showed that Israel should continue to focus on education, to ensure employees and graduates have the skills to develop and work with AI.

Google Cloud has been doing its part to build on these AI endeavors, including the opening of a cloud region in Tel Aviv a year-and-a-half ago, and expanding data residency for data stored at-rest for Gemini, Imagen, and Embeddings APIs on Vertex AI. And, earlier this year, we announced a new $4-million Google AI Startup Fund, designed to provide equity-free awards to local startups that have received funding from the Israeli Innovation Authority's fast grant channel.

As leaders from within and beyond the AI space gather this week for the Google Cloud Summit Tel Aviv 2024, we wanted to highlight a number of our top AI customers from Israel — who are among the 60% of funded gen AI startups, and 90% of AI unicorns, that are global customers of Google Cloud.

CytoReason

On average, it takes more than ten years for a new drug to go from initial discovery to the market. CytoReason hopes to change that by developing computational disease models using vast amounts of data from countless indications and treatments.

The company maps human diseases, going tissue-by-tissue, cell-by-cell. Through its AI-led platform, CytoReason aims to help pharma companies shorten clinical trials and reduce the high costs of drug development. It already has relationships with Pfizer, Roche, and Merck, among other leading pharmaceutical organizations.

With the huge amounts of data that CytoReason handles, its SQL servers could no longer handle the ever-growing requirements, so it switched to BigQuery — Google’s data warehouse platform — to help store, manage, and access its data more quickly and efficiently. To help utilize that data and to scale its models faster, the company then worked with a Google Cloud partner to optimize how it uses and deploys Google Kubernetes Engine (GKE).

For example, CytoReason has recently been working with one of its clients, Sanofi, on inflammatory bowel disease (IBD), using machine learning to help identify patient subtypes and pair them with personalized IBD targets.

It’s just one example of CytoReason’s mission, which is to move bioscience “from trial and error, to predictable medicine.”

Fairtility

The fertility journey, especially the IVF journey, is as much an emotional experience as a physical one. Improving visibility into the clinical and laboratory data to both IVF professionals and patients can help improve outcomes and make the process smoother.

Health-tech startup Fairtility is harnessing AI and data to revolutionize reproductive care and reimagine IVF for the modern fertility clinic. Among the company’s offerings is CHLOE, a digital lab assistant that helps assess the viability of embryos using visual AI analysis.

Ultimately the goal is to empower clinics around the world with the latest AI technology to improve outcomes. This should help reduce the number of IVF cycles to achieve a live birth, minimize the emotional, physical, and financial burden on prospective parents, and enable more collaborative discussions between clinicians and patients.

Google Cloud's flexibility, efficiency, and robust support for computationally intensive AI applications, have allowed Fairtility to accelerate innovation and make a meaningful difference in the lives of patients worldwide.

Hour One

One of the most popular applications of generative AI has been the creation of images and videos from nothing but words. Hour One seeks to offer enterprises more meaningful applications in the form of its video avatars, who can serve in roles as varied as brand ambassadors or customer service.

Hour One uses AI to enable businesses with no technical expertise to produce professional-grade videos in minutes for highly affordable prices. By using generative AI, Hour One can significantly lower the cost of video production, helping companies deliver content in an engaging way.

Creators simply log into the platform, describing the video they want, and then the company’s generative AI solutions create a cinematic video, complete with a virtual human presenter narrating either a provided script or one generated by AI. Among the popular applications so far have been workforce training, such as for HR, IT, or sales enablement.

To execute these intensive workloads, Hour One needed more processing power to support its compute-heavy gen AI content creation. By partnering with Google Cloud, Hour One gained powerful GPUs to enable better image quality, faster turnover, and more sophisticated algorithms, allowing it to scale up its business.

Lightricks

For those who aren’t making video from text prompts alone, but instead want to enhance their original material, AI offers a plethora of features and enhancements to creativity and more efficient workflows. This is especially true for Lightricks’ latest product, LTX Studio — a pioneering digital storytelling platform powered entirely by generative AI.

For more than a decade, Lightricks has offered a cadre of video and imaging editing apps, many of them award-winning, including the popular selfie-editing tool, Facetune. Lightricks also is the creator behind Videoleap and Photoleap, helping creators build sparkling images and videos with powerful editing capabilities. The company’s suite of apps has been downloaded more than 750 million times.

AI features have long helped enhance these apps, and to make them function even better. Lightricks began to leverage the performance and ample memory capacity of Google Cloud TPUs, including the most recent generation v5p, which helped to successfully train its generative text-to-video model without splitting it into separate processes.

By having the optimal AI-specialized hardware, Lightricks could significantly accelerate each of its training cycles, allowing it to swiftly conduct a series of experiments. The ability to train its model quickly in each experiment facilitates rapid iteration, an invaluable advantage for its research team in this competitive field of AI-generated media.

Wiz

Data scientists and engineers are bringing models to production faster than ever to deploy more AI applications. Oftentimes, though, they are unaware of how to secure their AI infrastructure against modern cloud attacks. Wiz, an Israeli cloud security startup, works to help security teams empower their data teams to deploy more useful AI applications faster and more responsibly.

One step in helping to secure these models is platform integration. That’s why Wiz recently launched support for Google Cloud Vertex AI. Now, Wiz customers can build, train, and deploy ML models at scale while ensuring they are effectively monitoring and managing the security risks associated with managing AI/ML models. Wiz allows Vertex AI customers to identify misconfigurations, external exposure, sensitive data, and identity risks behind Vertex AI services.

Wiz is also using Gemini to add AI-powered features to help customers improve their experiences with the Wiz Cloud Security Platform, interacting in a more straightforward and efficient way.

Opening image created with MidJourney, running on Google Cloud, using the prompt: An electric grove of cypress trees that almost look like they're made of pure energy, sprouting up in the desert done in an impressionistic style, suggesting the idea of blossoming technology; make it look like the trees are drawing electricity up into their roots from the desert, please; make the ground look more electric, with charged blue tendrils of electricity; make the scene look brighter and cheerier.