SoftBank's Next Frontier: Intelligence, Not Just Infrastructure

The signal is clear: when a $100 billion fund hires a former spy chief, they're not betting on better algorithms—they're betting on control. SoftBank’s appointment of Yossi Cohen, former director of Mossad, as its AI strategic advisor is not a routine personnel move. It is a declaration of war. The market yawned, but the noise is the signal. Alpha found in the noise. SoftBank has long been the most aggressive bettor on AI infrastructure. Masayoshi Son’s obsession with AGI is no secret—he has publicly predicted the arrival of artificial general intelligence within a decade and positioned SoftBank as the “infrastructure provider for the AGI era.” The crown jewel is Arm, the chip architecture underlying nearly every mobile AI chip and an increasing share of server CPUs. With Vision Fund’s peak $100 billion pool, SoftBank commands capital at a scale few can match. Yet after the WeWork and Uber debacles, Son pivoted from spray-and-pray finance to concentrated, controlling bets. AI became the single focus. The Cohen appointment completes that pivot. Core insight: SoftBank is constructing a triad of capital, chip control, and intelligence. This is not about better models—it’s about controlling the environment in which models operate. Cohen’s background in intelligence and security gives SoftBank a capability no other major AI investor possesses: a built-in geopolitical risk assessment unit. From my years auditing Layer-1 tokenomics in 2018, I learned that the most dangerous blind spots are ignored until they collapse. SoftBank’s move suggests they see the collapse coming—not of code, but of trust. The AI industry is moving from a race for accuracy to a race for safety and sovereignty. The party that can verify that an AI system is not being exploited by state actors, or that its training data is not compromised, will hold the ultimate leverage. Cohen’s network—especially in Israel, where cybersecurity and AI talent are deeply intertwined with military intelligence units like Unit 8200—provides a channel to evaluate and invest in startups that operate at the intersection of AI and national security. That is a pipeline invisible to most Silicon Valley VCs. Collapse detected. Lessons extracted. But the contrarian view is sharp. This move carries significant risk. First, geopolitical compliance blowback: Cohen’s Mossad background will trigger extra scrutiny in China, where SoftBank has deep ties (Alibaba, Didi, ByteDance). It may also complicate relations with Middle Eastern LPs like Saudi Arabia’s PIF, which has no formal relations with Israel. Second, reputation risk: the AI safety community—which includes researchers at OpenAI, DeepMind, and Anthropic—is deeply wary of any association with intelligence agencies. Trust is the currency of open collaboration; Cohen’s appointment could poison that well. Third, the risk of overestimating the value of intelligence: SoftBank may believe Cohen’s network gives them a unique edge, but the actual investment outcomes will depend on execution, not access. Hiring a former spy does not guarantee superior due diligence—it may just add a layer of opacity that alienates founders and regulators. Bubble burst. Truth remains. From my experience designing a DeFi yield strategy in 2020, I saw how liquidity fragmentation was a manufactured narrative—but geopolitical fragmentation is real. SoftBank’s bet is that AI will become the most geo-strategic technology since nuclear weapons, and that the winners will be those who can navigate the security landscape. Whether that vision is correct depends on how quickly the rest of the market adopts the same mindset. If they are early, the payoff is enormous. If they are wrong, they become a target. Yield farming's new frontier. Takeaway: SoftBank is not just investing in AI—it is investing in the architecture of control. The next 12 months will reveal whether this strategy unlocks exclusive access to Israeli AI security startups or triggers a wave of regulatory backlash. The question every investor should ask: who holds the ultimate authority over AI safety—the technologists or the spooks? The answer will define the next cycle of crypto and AI investment.