If you have read Israeli tech news in recent weeks, you probably came away with a sense of crisis. Atera announced layoffs of about 10% of staff. SolarEdge began another round of about 200 cuts. Fireblocks is laying off about 7%. XTEND is cutting about 10%, two months after listing. After CrowdStrike acquired XM Cyber's technology, most of its Israeli employees were let go. And Qualcomm begins a global round of performance-based layoffs this month.
But had you read other news in the same weeks, you would have got the opposite picture. Tower said it will add about 150 employees in Israel. Nova and Camtek posted record quarters. Nvidia Israel has topped the Dunstech ranking of the best large tech companies to work for three years running and keeps hiring. Israeli cyber companies raised billions this year. So which is true? Both. And understanding the gap between them may be the most important thing for anyone working in tech or considering it.
The great split
An analysis by Dror Ben in Calcalist summed up the situation in one word: split. According to the data he cited, the layoff rate in software jumped to 6.6% in the first half, more than double the industry average. On the other side, chip, compute infrastructure and deep-tech companies are expanding. Overall tech numbers look stable, but beneath them the composition of the workforce is changing deeply.
The main reason is AI, and it works in two directions. On one side, tools like coding agents let software companies do more with fewer developers. Tasks that once took a whole team, such as writing tests, fixing bugs or building a simple interface, are done faster with AI. On the other, AI creates enormous demand for chips, data centers, security and infrastructure, and that is where companies are hiring.
Fiverr may be the clearest example of the first side. In its latest results, active buyers fell 25.8% in a year and marketplace revenue fell 24.7%. People who once ordered a logo, a translation or marketing copy from a freelancer now do it themselves with ChatGPT or Claude. It is not just one company's crisis; it is a sign of what is happening to any kind of work that can be described as a defined, repeatable task.
Not everything is AI
But be careful not to blame everything on AI. Some layoffs have entirely different causes. SolarEdge has struggled for years with a slowdown in residential solar, and this is its fourth or fifth round of cuts since 2024. Fireblocks is still adjusting to the crypto market after raising at an $8 billion valuation at the 2022 peak. XTEND is dealing with a stock that lost most of its value after listing.
And even where companies cite AI, some experts warn of 'AI washing': using AI as a convenient explanation for layoffs that really stem from other causes, such as slowing growth, investor pressure or the strong shekel, which sharply raises payroll costs for companies that earn in dollars and pay salaries in shekels. AI is a real factor, but not the only one.
Who is under pressure
So who is most at risk? Based on what is happening in the market, 'general' development roles at software companies, especially where AI is already good: simple interface work, manual software testing, basic technical support, content writing and translation. Middle-management roles focused mainly on coordination and reporting may also be hit, as companies try to flatten their structures.
The group causing the most concern is juniors: people finishing computer science degrees or retraining courses and looking for a first job. Companies used to hire young developers to handle simple tasks and learn on the job. Today some of those tasks are done by AI, and some companies simply hire fewer beginners. That is a problem for the industry too: if juniors are not trained today, where will the seniors come from in five years?
Who is in demand
On the other side, some fields see demand holding or even growing. The first is chips and hardware: VLSI engineers, optics engineers, physicists, image-processing experts and production staff. This week's results from Nova, Camtek and Tower show the field growing by tens of percent a year. The second is cyber, especially data and identity security, the field of companies like Cyera, Sentra and Orion.
The third field is AI itself, but not only people who build models. There is great demand for people who can deploy AI in organizations: connecting models to existing systems, building agents that actually work, and checking they do not make mistakes or expose data. The fourth is infrastructure: cloud, data centers, networks and energy. Anyone who understands how to make large systems run efficiently is in more demand than ever.
Another trend to watch is the return of 'performance culture'. Qualcomm, for example, said its layoffs will be based on annual performance reviews, and CEO Cristiano Amon wrote to employees that the aim is to build 'a stronger performance culture'. Other global giants have made similar moves in the past two years. For employees, it means job security at big companies is no longer guaranteed even when the company is profitable, and performance reviews matter more.
Employers carry responsibility here too. Companies that stop hiring juniors save money in the short term but may find in a few years that they have no one to grow into the next generation of managers and experts. Some are already trying other models: internship programs, teams in which a junior works alongside a senior developer and AI tools, and internal training that moves existing staff into more in-demand fields. Whoever builds such a model will have a real edge in the competition for talent.
What to do
For those working at a software company today, the first recommendation is to make AI a work tool rather than a threat. Developers who use coding agents well get more done, faster, and they are the ones companies will want to keep. The second is to go deep: someone who deeply understands a system, a business domain or a specific technology is harder to replace than someone who knows 'a bit of everything'.
For those seeking a first job, the market is harder but not closed. Consider fields with high demand even if they are less glamorous than a Tel Aviv startup: chip companies in the north and the Shephelah, infrastructure companies, security roles. Build real projects that show you can work with AI tools, not just grades. And anyone considering retraining should check exactly which role a course prepares for, and whether that role is in demand today.
And the best companies to work for? This year's Dunstech ranking offers a hint: Nvidia Israel at the top, followed by Island, the cyber company that built a secure enterprise browser, Microsoft's Israeli development center and Palo Alto Networks. Almost all the top companies work in fields that AI strengthens rather than threatens. That is no coincidence.
Bottom line
Israeli tech is not collapsing, but it is changing fast, perhaps faster than ever. An industry built on thousands of developers writing code for apps and cloud services is entering an era in which some of that work is done by machines, while other work, in chips, infrastructure and security, becomes more important than ever. Those who understand this shift in time and adapt to it will find themselves on the right side of the split.
