Skip to main content

The Rise of NVIDIA, Part 14: The $40 Billion That Got Away

The $40 Billion That Got Away

🎮 The Rise of NVIDIA — a 20-part series. See all parts »  |  « Part 13: Buying the Plumbing

By the summer of 2020, Jensen Huang had spent two decades building NVIDIA into the company that supplied the world's compute. But there was one thing NVIDIA didn't own: the instruction set that ran inside almost every phone, tablet, and embedded chip on Earth. That belonged to a quiet company in Cambridge, England. So NVIDIA tried to buy it — and set off the most contentious semiconductor deal in history.

The $40 Billion Handshake

On September 13, 2020, NVIDIA announced it had signed a definitive agreement to acquire Arm Limited from Japan's SoftBank Group for a staggering $40 billion, paid in a mix of cash and NVIDIA stock. It was, and remains, the largest deal the chip industry had ever attempted.

The logic was seductive. Arm doesn't manufacture chips — it designs the processor blueprints that other companies license, then license again. Arm technology sits inside Apple's iPhones, Qualcomm's Snapdragons, Amazon's data-center servers, and billions of tiny embedded controllers. Huang framed the combination in his usual grand terms: uniting NVIDIA's leadership in AI with Arm's "vast ecosystem" to create "the premier computing company for the age of artificial intelligence." He promised to keep Arm headquartered in Cambridge, build a world-class AI research center there, and even erect an Arm-powered supercomputer on the site.

SoftBank's Masayoshi Son — who had bought Arm for $32 billion just four years earlier — called NVIDIA "the perfect partner." Arm's own CEO, Simon Segars, said his team was "excited to be joining NVIDIA." On paper, everyone was aligned. The problem was everyone else.

Switzerland Is Not For Sale

Arm had a nickname in the industry: the "Switzerland" of semiconductors. Because it licensed its designs neutrally to everyone — including companies that competed fiercely with each other — Arm was trusted precisely because it took no sides. Apple, Qualcomm, Samsung, Amazon, Google, and yes, NVIDIA itself, all built on the same foundation.

Now one of those licensees wanted to own the foundation. Rivals were alarmed. If NVIDIA controlled Arm, what stopped it from seeing competitors' confidential roadmaps, raising their licensing costs, or steering Arm's future designs toward NVIDIA's own advantage? Qualcomm, Microsoft, Google, and others reportedly voiced objections to regulators around the world.

The watchdogs listened. The European Commission opened an in-depth investigation in October 2021, warning the deal could harm competition in chips for data centers, automotive, and the Internet of Things. Britain's Competition and Markets Authority raised both competition and national-security concerns — Arm was a crown jewel of UK technology. China's regulators were widely reported to be in no hurry to approve anything either.

The FTC Drops the Hammer

The decisive blow came from Washington. On December 2, 2021, the U.S. Federal Trade Commission sued to block the merger outright. The FTC called it "the largest semiconductor chip merger in history" and argued that the combined company "would have the means and incentive to stifle innovative next-generation technologies," from data-center chips to driver-assistance systems in cars.

"Tomorrow's technologies depend on preserving today's competitive, cutting-edge chip markets," said Holly Vedova, director of the FTC's Bureau of Competition. The complaint leaned on exactly the fear the industry had voiced: that NVIDIA would distort Arm's incentives and "unfairly undermine" its rivals. This was a vertical merger — supplier meeting customer — and antitrust enforcers had decided to make an example of it.

With three major jurisdictions lined up against the deal and no realistic path to approval, the math became impossible. On February 7, 2022, NVIDIA and SoftBank jointly announced the termination, citing "significant regulatory challenges preventing the consummation of the transaction, despite good faith efforts by the parties."

Losing, and Winning Anyway

The failure wasn't free. SoftBank kept the $1.25 billion that NVIDIA had prepaid, booking it as pure profit. NVIDIA walked away with nothing to show except a 20-year license to keep using Arm technology — the same license it already had.

Yet Huang's parting words were oddly generous. "Arm has a bright future, and we'll continue to support them as a proud licensee for decades to come," he said, predicting Arm would be "the most important CPU architecture of the next decade." SoftBank, denied its buyer, did the obvious thing: it took Arm public. In September 2023, Arm listed on the Nasdaq under the ticker ARM in one of the year's biggest IPOs, with SoftBank retaining the controlling stake.

Here is the twist history handed NVIDIA. The deal collapsed in early 2022 — months before ChatGPT launched and the AI boom sent NVIDIA's own valuation past a trillion, and then multiple trillions, of dollars. Had the merger gone through, NVIDIA would have paid $40 billion in cash and stock that soon became worth vastly more. Losing Arm may have been the most expensive-looking defeat that quietly saved NVIDIA a fortune. Sometimes the deal you don't get is the luckiest break of all.

Next: the software moat almost nobody saw — how CUDA quietly locked in a generation of developers and became NVIDIA's real fortress.


🔗 Explore more from Syncster

Comments

Popular posts from this blog

Cursor AI Review: Is the AI Code Editor Worth It?

I've been using Cursor as my main code editor for a while now, and enough people have asked whether it's worth switching to that a proper review felt overdue. Short version: for me, yes — but with caveats. What is Cursor? Cursor is an AI-first code editor built as a fork of VS Code. That means every extension, theme, and keybinding you already use in VS Code works here, but with AI woven directly into the editing experience instead of bolted on as a plugin. It's made by Anysphere and can run models from OpenAI and Anthropic under the hood. What I like Tab completion is uncanny. Cursor predicts your next edit — not just the rest of the line, but the next change across the file. Once you get used to hitting Tab, going back to a plain editor feels slow. The Composer / Agent mode. You describe a change in plain language and it edits multiple files at once, showing you a diff to accept or reject. For refactors and boilerplate, this saves real time. It unde...

How I used Google Sheets and Apps Script

Google Sheet is one of the most powerful spreadsheet application that exists online, rivaling with Microsoft's Excel. One of the main strengths is its strong support for collaboration with other users, much easier and popular than collaboration tools with Microsoft Office. Aside from plain spreadsheet, it also supports extensions such as macro. If you are familiar with macros on other office tools, they work almost the same. However, the most extension I use and tinker with is the Apps Scipt . Apps Script Extension One of the challenges I faced recently is how do I track or monitor reports in our department if they are submitted on time or worst, forgotten due to lack of better monitoring tools. So I thought if there can be simple applications that can be deployed or use by a more general user to allow reminding periodically what reports are approaching due dates or those that are past dues. Then I looked for a way, instead of creating a full blown app from scratc...

MacBook Pro M5 vs M5 Pro: Which One Should You Actually Buy?

Apple's latest 14-inch MacBook Pro comes in two very different flavors: the base M5 and the step-up M5 Pro . On paper they look similar — same gorgeous Liquid Retina XDR display, same design — but under the hood the gap is bigger than the names suggest. Here's a clear, no-hype breakdown, with concrete use cases so you can match the chip to your work. Quick spec comparison Spec M5 M5 Pro CPU 10-core (4 performance + 6 efficiency) Up to 18-core (6 performance + 12 efficiency) GPU 10-core Up to 20-core Neural Engine 16-core 16-core Memory bandwidth 153 GB/s 307 GB/s (roughly double) Unified memory 16 / 24 / 32 GB 24 / 48 / 64 GB Max storage Up to 4 TB SSD Up to 8 TB SSD Battery (video playback) Up to 24 hours Up to 22 hours Media engines Single encode/ProRes engine More encode/ProRes engines (higher configs) What actually changes between them More cores — the M5 Pro nearly doubles CPU cores and adds GPU cores, so sustained, multi-threaded work finishe...