Loading...
Loading...
SoftBank is intensifying its bet on AI hardware and compute infrastructure through investments, acquisition interest, and bold expansion plans. The group and its Arm unit reportedly approached Cerebras before its blockbuster IPO but were rebuffed, underscoring competition for high-performance AI accelerators. SoftBank has also backed Graphcore with a $450 million injection and is negotiating a potential $100 billion plan to build wafer fabs and low‑carbon data centers in France to create a vertically integrated AI stack. These moves show SoftBank pursuing scale across chip design, manufacturing, and data‑center capacity to challenge incumbents and secure long‑term AI compute supply.
SoftBank's push into AI chips and infrastructure could reshape supply chains and competitive dynamics for AI hardware, affecting partners and rivals. Tech professionals should track potential shifts in chip design, data center capacity, and funding flows that influence platform choices and procurement.
Dossier last updated: 2026-05-13 16:20:12
Cerebras 筹集 55 亿美元，为 2026 年的 IPO 季拉开帷幕 - TechCrunch
Jordan Novet / CNBC : Cerebras opens at $350, valuing the chipmaker at $100B+, after raising $5.5B by selling 30M shares at $185, the largest US tech IPO since Uber's debut in 2019 — Cerebras Systems soared in its Nasdaq debut on Thursday, opening at $350 after selling shares at $185, well above the company's expected range.
Cerebras, the maker of wafer-scale AI inference accelerators, rebuffed preliminary acquisition interest from Arm and its majority owner SoftBank in the weeks before the company’s expected IPO, Bloomberg reported. Cerebras is pursuing a large U.S. public offering and has raised its share price twice to $185, implying it would raise about $5.55 billion with a market value near $40 billion (fully diluted ~$49 billion). The rejection signals Cerebras’s confidence in public-market valuation and preserves its independence amid intense competition and consolidation in AI chip supply for cloud providers and AI companies.
Arm and its majority owner SoftBank reportedly approached AI compute chipmaker Cerebras Systems in the weeks before Cerebras’s planned IPO with an acquisition bid, but Cerebras declined the overtures. The talks surfaced as Cerebras prepared to go public, highlighting strategic interest from Arm and SoftBank in expanding into high-performance AI hardware. If pursued, such an acquisition would have accelerated consolidation in the AI accelerator market and strengthened Arm’s positioning in AI compute ecosystems. The report underscores ongoing M&A activity around specialized AI chip vendors as tech incumbents and investors seek to secure hardware advantages for AI workloads.
Arm和软银曾试图在最后关头收购Cerebras - Bloomberg.com
2026年迄今为止规模最大的IPO：人工智能芯片制造商Cerebras计划筹集48亿美元，但预测市场如何看？ - Yahoo Finance
在Cerebras上市前，Arm和软银曾试图收购该公司
Arm和软银曾试图在最后关头收购Cerebras
Bloomberg : Sources: Arm and SoftBank expressed preliminary interest to acquire Cerebras weeks before its expected IPO; Cerebras rebuffed them — Arm Holdings Plc and its majority owner SoftBank Group Corp. made an approach to acquire Cerebras Systems Inc., the AI computing firm, weeks …
SoftBank CEO Masayoshi Son is reportedly negotiating a plan to invest up to $100 billion to build AI-focused wafer fabs and data centers in France as part of his broader "Izanagi" initiative. The project would leverage Arm Holdings' chip designs (SoftBank owns ~90% of Arm) and France’s nuclear-powered grid to supply low-carbon electricity for large-scale AI compute. SoftBank may contribute about $30 billion, with roughly $70 billion expected from Middle Eastern sovereign funds like Abu Dhabi’s MGX. The move aims to create a vertically integrated AI semiconductor and infrastructure stack to rival Nvidia, but raises concerns about SoftBank’s financial risk after recent large loans and heavy exposure to OpenAI. It marks a potential major shift in global AI hardware investment.
SoftBank has invested $450 million in UK semiconductor company Graphcore, according to First Financial. The strategic funding boost targets Graphcore’s AI accelerator chips and support for scaling its business amid rising demand for specialized AI hardware. The injection from a major tech investor matters because it strengthens Graphcore’s position against rivals like NVIDIA and other AI-chip startups, helps accelerate product development and go-to-market efforts, and signals continued investor appetite for AI-focused semiconductor plays. For the broader tech and cloud industries, the deal highlights ongoing capital flows into AI hardware that underpin large language models, generative AI workloads, and data-center acceleration.