Opinion
While the world builds more memory chips, Israel could reinvent them
The global memory-chip shortage is usually described as a supply problem. That is only part of the story. Behind the shortage lies a deeper technological constraint – one that could create an important opportunity for Israel.
Since late 2025, shortages of DRAM, the working memory used by computers and servers, and NAND flash, which stores data, have pushed prices sharply higher. Artificial intelligence is a major reason why. AI systems need to move enormous amounts of data at very high speeds, creating extraordinary demand for High Bandwidth Memory, or HBM – a high-performance form of memory designed primarily for data centres and used alongside advanced processors.
IDC projects that DRAM revenues will nearly triple in 2026 to $418.6 billion, driven in large part by soaring demand for HBM. Yet HBM is difficult to manufacture and package. Memory producers have shifted substantial capacity towards it, but are still struggling to keep pace. The result is less capacity for conventional memory and shortages felt well beyond data centres, from personal computing and cars to industrial manufacturing. The impact is already reaching consumers: in late June, Apple raised prices on a range of Macs and iPads by around 20%, citing sharply higher memory and storage costs.
1 View gallery
From left: Avner Golov, Dan Yachin, Or Tzofi,
(Alan Chapelski, Ella Faust, Kobi Kuenkas)
The natural response is to build more factories. More production capacity is certainly needed. But it addresses the immediate shortage, not the underlying technological problem.
For roughly a decade, memory technology has advanced much more slowly than the processors that rely on it. While leading-edge logic chips have become dramatically smaller and more powerful, DRAM has made far less progress. In practical terms, computing power has continued to accelerate while the memory needed to feed it has struggled to keep up. The industry refers to this growing imbalance as the “memory wall”.
There is an economic reason for this stagnation as well. For many years, memory was viewed as a relatively low-margin, highly cyclical business. That made the long-term, high-risk research required for a fundamental breakthrough less attractive. Alternative technologies have been developed, but none has yet been able to replace DRAM at scale while matching the necessary combination of cost and performance.
AI has now turned what was once an industry constraint into a strategic bottleneck. The industry is responding by increasing production and finding more sophisticated ways to connect memory to processors. These solutions can improve performance, but they do not solve the underlying problem: even HBM ultimately relies on the same DRAM technology.
This matters strategically because the memory industry is dominated by a small group of companies based mainly in South Korea, the United States and Japan. Samsung, SK Hynix and Micron account for most of the DRAM market, while Samsung, SK Hynix, Kioxia, Micron and SanDisk are major players in NAND. HBM is more concentrated still: according to Counterpoint Research, South Korea’s Samsung and SK Hynix together account for roughly 80% of the global market by revenue.
That concentration is already creating vulnerabilities. As supply tightens and prices rise, technology companies are looking beyond the traditional suppliers. Apple is testing memory chips from China’s CXMT, while some PC makers have already begun using its products. This could deepen reliance on Chinese suppliers at a time when US policy is seeking to reduce such dependencies in critical technologies.
There is also a longer-term risk. If a competing power achieves a breakthrough in the way memory is designed, it could bypass today’s bottlenecks and shift the technological balance.
This is where Israel could have a distinctive role.
Israel is unlikely to compete with South Korea, the United States or Japan in building large-scale memory-chip factories. Nor should that be the goal. Israel’s comparative advantage has never primarily been scale. It lies in advanced engineering, chip design, algorithms and the ability of highly skilled teams to solve complex technological problems.
The foundations are already in place. Israel’s semiconductor ecosystem includes roughly 200 companies employing about 45,000 people, and the country ranks second only to the United States in the number of chip start-ups. Intel, Nvidia, Apple, Google and Samsung all maintain significant development activity in Israel. Acquisitions such as Mellanox, Mobileye, Habana Labs and Annapurna Labs also demonstrate how technologies developed locally can become important assets for the world’s largest technology companies.
The strategic opportunity, therefore, is not for Israel to manufacture more memory than the global giants. It is to help develop better memory.
That means treating memory technology as a focused national R&D priority. The central effort should be directed towards new ways of designing memory and potential alternatives to DRAM. The objective would be to build knowledge, intellectual property and technological capabilities in areas where Israel can realistically achieve global relevance – much as it has done in chip design and algorithms.
Such research is inherently risky. That is precisely why it requires a deliberate framework linking universities, Israeli technology companies, multinational R&D centres and government funding. A breakthrough may not come quickly, and many research paths will fail. But the potential strategic and economic return is correspondingly large.
Israel should also strengthen work in complementary fields, including advanced chip integration and photonics. These technologies will not replace the need for innovation in memory itself, but they can improve how processors and memory work together. In photonics in particular, Israel already has academic and industrial capabilities that could be developed into a strategic advantage.
Finally, this should not be a purely national effort. Israel should use its participation in the US-led Pax Silica initiative to deepen partnerships with the United States, South Korea and Japan. Such cooperation could give Israeli researchers and companies access to laboratories, government research programmes, pilot facilities and advanced manufacturing infrastructure, while strengthening Israel’s role in global semiconductor supply chains.
The world is racing to build more memory. Israel should focus on a different question: Can memory itself be made better?
For a country whose technological strength has long come from doing more with less, that may be the more important race to enter.
Avner Golov is VP Research and Alliances at MIND Israel. Dan Yachin is a Senior Researcher for Critical and Emerging Technologies at MIND Israel. Or Tzofi is an AI and Geopolitics Researcher at MIND Israel.
