SoftBank's $100B Project Izanagi Bets Big on Controlling the AI Compute S-Curve

This new chip venture is a direct bet on the exponential growth of AI compute demand, aiming to capture the foundational infrastructure layer as adoption accelerates. The project, led by Katsunori Sago, a former SoftBank strategy officer and ex-Goldman banker, is part of a broader strategic offensive. Sago, who once seemed poised to succeed SoftBank founder Masayoshi Son, is now deploying his own capital into Japanese startups, hunting for long-term returns. His move signals a shift in Japan's venture landscape, attracting top talent away from traditional corporate careers.

The venture's true scale and ambition are defined by its parent: Masayoshi Son plans to raise $100 billion for a new chip venture called Project Izanagi. This colossal fund is a strategic bet to rival Nvidia and build the essential compute layer for the AI paradigm shift. Son envisions this initiative as contributing to the development of artificial general intelligence (AGI), a future where machines surpass human intelligence. It complements SoftBank's existing Arm Holdings, which designs processor architectures, aiming to create a vertically integrated stack for the next computing era.

Viewed through the lens of the technological S-curve, this is a classic play on the steep adoption phase. The venture targets the infrastructure layer where ownership captures disproportionate value as demand explodes. SoftBank's parallel moves-investing up to $40bn in OpenAI and launching a $500bn AI infrastructure plan in the US-show a consistent strategy of backing the entire stack. Project Izanagi represents the hardware foundation for that stack, positioning its backers to benefit from the exponential growth curve, not just the application layer.

Strategic Rationale: SoftBank's Vertical Integration Play

SoftBank's new chip venture is not a standalone bet. It is a deliberate piece in a much larger puzzle of vertical integration, aimed at controlling the entire AI infrastructure value chain. The company is assembling a stack from the ground up, starting with architectural IP and moving through manufacturing to the software layer.

The strategic logic is clear. SoftBank already owns Arm, Graphcore, and Ampere Computing, giving it deep control over processor design. Now, by investing $2 billion in Intel common stock, it is directly acquiring a stake in advanced semiconductor manufacturing capacity. This move complements its existing architectural holdings, creating a vertically aligned ecosystem. The goal is to own the fundamental rails of the AI paradigm, from the chip blueprint to the factory floor.

This chip venture directly complements SoftBank's other massive AI bets. The company has fully satisfied its $40 billion commitment to OpenAI, securing an 11% stake in the leading AI model developer. This creates a powerful feedback loop: SoftBank can use its chip designs and manufacturing capacity to produce hardware optimized for OpenAI's models, while OpenAI's software advancements drive demand for that hardware. It is a vertically aligned ecosystem from models to chips, a setup designed to capture value at every layer.

This concentrated infrastructure bet is a direct response to the exponential growth of AI compute demand. The venture's scale-$100 billion for Project Izanagi-is justified only if demand for compute power continues to outpace supply. By controlling the architectural IP, manufacturing, and software layers, SoftBank aims to position itself as the essential infrastructure provider for the next computing era. It is a classic S-curve play, betting that the steep adoption phase of AI will reward those who own the foundational infrastructure.

Risks and Catalysts: Execution on the Exponential Curve

The venture's massive scale is matched only by the monumental challenge of execution. The primary risk is simply catching up to entrenched leaders like Nvidia. The company has built a dominant position not just with hardware, but with a ubiquitous CUDA software stack that has evolved for over 16 years. Displacing that ecosystem requires not just competitive chips, but a compelling new software platform-a task of staggering complexity. The venture's success hinges on whether it can achieve a meaningful market share against this entrenched moat.

A major catalyst will be the launch and adoption rate of Project Izanagi's chips. The venture must demonstrate clear performance and cost advantages to gain traction. The project's name, after the Japanese god of creation, reflects its ambitious vision, but the real test is in the silicon. Early adoption by major cloud providers or enterprise customers will be critical proof of concept. Without rapid, widespread uptake, the $100 billion bet risks becoming a costly footnote in the AI infrastructure race.

Finally, investors must monitor SoftBank's financial health and its ability to fund these long-duration, high-risk bets. The company has a strong cash position, with ¥6.2 trillion ($41 billion) in cash and cash equivalents, and plans to invest $30 billion from its own coffers. Yet its history with the Vision Fund, including the massive bet on WeWork, shows the vulnerability of such concentrated, leveraged strategies. SoftBank has already sold off major assets, including a stake in Arm for up to $40 billion, to fund its current moves. The risk is that overextending on this chip venture could strain its balance sheet, diverting capital from other parts of its AI stack or leaving it exposed to economic downturns. The venture's fate is inextricably linked to SoftBank's financial discipline.