At Globes’ TECH IL conference, entrepreneurs and investors debated how Israel’s tech sector will look after the AI revolution, what advantage the country still has, where the value will accumulate, and what the state must do to stay relevant. The panel was moderated by Globes high-tech reporter Meital Wiesberg.
Tomer Tsikurul of Phoenix Investment said he was optimistic, arguing that Israel can both build technology and attract capital seeking it. Raz Mangel of Greenfield agreed the startup and investment environment is strong, but said the government must do much more in education and infrastructure if Israel wants leadership in major AI categories. Moshe Shalev, co-founder and chief product officer at Decart, said Israel has strong military-tech units and academia, but entered AI late and needs more founders willing to take risks.
Several speakers said the biggest opportunities may not be in building the foundation models themselves. Maksim Bar Kogan of Onyx Security said chipmaking is already a strength and that Israeli boldness can help in data-center construction. Erik Kleinstin of Glilot Capital said Israel is weaker at infrastructure but stronger in layers above the cloud, such as security and optimization. He also argued that people coming out of the defense sector will become the next AI founders and investors.
On jobs, Bar Kogan predicted a sharp drop in the number of programmers, comparing it to the Industrial Revolution, and said AI is already better than more than half of developers, possibly 70% next year. He said that could create a split between very highly paid programmers, even earning 1 million shekels a month, and fewer routine roles. Kleinstin added that AI will create economic growth and new jobs, but with more demand for soft skills and a risk that heavy AI use weakens cognition, citing a Harvard study. He also warned that cyberattacks will become easier and that small businesses will be the biggest victims.
Asked whether Israel can still build another Anthropic, Mangel said yes, because the field is non-linear and long-term. The panel also discussed U.S. restrictions on exporting the most advanced Claude model, with Tsikurul saying countries will respond by building their own models. On state policy, Shalev called for much larger Innovation Authority-scale efforts, Bar Kogan said the safest bet is enabling energy and regulation for data centers, and Kleinstin said the new AI headquarters is a start, but Israel still lacks the hyperscale firms that can invest hundreds of billions in local data centers.
