Bank of America's AI ranking puts Israel 12th. The binding constraints are energy, minerals, and data centers.

Bank of America's Global Research team has ranked 30 economies on their ability to compete in AI behind the United States and China. The result is a list that strips the usual "talent nation" framing down to plumbing. South Korea finishes first in both the short-term (under three years) and long-term (three to ten years) assessments. The United Arab Emirates leads the long-term category on the strength of a 70% corporate AI adoption rate, the highest in the world, and a stated national target to become the leading AI country by 2031. Israel, despite a workforce of AI engineers denser per capita than almost any other country's, lands 12th of 30. (Ynetnews, citing Calcalist)

The diagnostic behind those placements is not who has the most researchers. It is who can plug them into the physical inputs that large-scale AI requires: compute, energy, hardware supply chains, and the minerals that go into them. South Korea and the UAE have it. Israel, on Bank of America's reading, does not.

The ranking, produced by Bank of America's Global Research group and obtained by the Israeli business daily Calcalist, divides the 30 economies into two horizons. The short-term list weights the inputs needed to build AI infrastructure now: computing capacity, energy availability, hardware supply chains, minerals, labor stability, and investment. The long-term list weights the ability to turn AI adoption into productivity-driven growth. The two lists agree at the top. They diverge below: countries with industrial depth and cheap energy (South Korea, the UAE) climb in both; countries without them sit lower in one or both lists.

South Korea tops the table on both axes. The country combines a deep manufacturing base in memory chips and advanced electronics with sustained state and private investment in compute, and Bank of America's analysts credit that industrial depth for a finish no other mid-sized economy has matched. The United Arab Emirates sits a few places behind South Korea in the short-term list and climbs into the long-term lead on the back of a 70% corporate AI adoption rate and a government strategy backed by some of the cheapest energy in the world. UAE officials have publicly targeted 2031 as the year the country takes the lead.

Israel's profile splits cleanly. The country scores well on private AI investment, on the number of AI startups per capita, on its role in semiconductor and AI-adjacent supply chains, and on the depth of its AI workforce. It scores poorly on energy availability, on access to the minerals and metals that data-center and chip manufacturing require, and on the physical plant of large-scale data centers. Canada, Switzerland, the Netherlands, and Singapore outperform Israel in several categories. South Korea and the UAE outperform it across most of them. None of the 30 economies is close to the United States or China; the ranking is about who climbs the second tier.

The pattern is the same one that shows up in Israel's broader tech sector: strong innovation and talent, weak on government investment, on infrastructure costs, and on natural resources. Bank of America's ranking formalizes a constraint that has been a hand-wave in AI competitiveness coverage until now. Talent density without infrastructure density is a ceiling, not a story of decline.

The next three to ten years are where that ceiling starts to bind. Corporate strategy, capital allocation, and energy policy are already diverging across the second tier. South Korea's chipmakers and the UAE's state-backed compute build-outs are scaling. Israel's grid, mineral exposure, and data-center pipeline are not visibly closing the gap. The next Bank of America update will read like a scoreboard for the same race: which mid-sized AI economies pair their workforce with the kilowatt-hours, the racks, and the rare-earth supply chains that large-scale AI now demands.