Sources: SoftBank's Arm is targeting an IPO at a valuation of between $60B and $70B as soon as September 2023, a sign of bullish interest in AI chips
- Chip design firm may raise as much as $10 billion in debut — Nvidia, Intel are in discussions to be anchor investors
Context & Ripple Effects
Arm's path back to the public markets has been a slow re-rating. After the $60B valuation SoftBank sought when the Nvidia sale collapsed in 2022, expectations were cut to an $8B raise at a $50B+ valuation by March 2023 — and this report marks the climb back above $60B, with the raise size up to $10B.
The structure matters as much as the number: Nvidia and Intel are in talks to anchor the deal, and days later Apple, Samsung, Nvidia, and Intel were all reported as planned investors — meaning Arm's biggest licensees would become its shareholders. SoftBank meanwhile bought out Vision Fund's 25% at a $64B valuation just before the August 21 filing, consolidating its control ahead of the debut.
First-order effects
- SoftBank converts its most valuable asset into up to $10B of cash while retaining control, repairing Vision Fund-era losses without giving up the upside on the remaining ~75% stake.
- Nvidia and Intel, if they anchor, take equity positions in the company whose instruction set underpins much of their own competitive ecosystem — a direct financial tie between rivals.
Second-order effects
- Bringing Apple, Samsung, Nvidia, and Intel in as investors locks Arm's largest customers into its roadmap, raising the switching cost for any licensee tempted by RISC-V alternatives.
- A successful $60B–$70B listing resets the pricing benchmark for every AI-silicon story behind it, giving SoftBank leverage for follow-on moves and other chip designers a public comparable.
Third-order effects
- The customer-as-shareholder cap table points toward Arm evolving from neutral IP licensor into a strategic platform — a direction confirmed when it later moved to stand up its own AI chip division targeting a spring 2025 prototype.
- If the debut prices well, expect more AI-infrastructure assets to route through public markets rather than M&A, shifting how the sector is financed and valued.
The trend: AI infrastructure financialization: the companies supplying the compute layer are being repriced through public listings and cross-holdings rather than acquisitions.