Google and Nvidia Consider Intel as Backup Chip Manufacturer, Why TSMC’s Loss Is Intel’s Gain
Let’s be honest. For the last few years, if you wanted a bleeding-edge AI chip, you didn’t really have a choice. You went to TSMC. That was it.
But on June 8, 2026, the script flipped.
In a report that sent Intel stock soaring nearly 13%, The Information dropped a bombshell: Google and Nvidia consider Intel as a backup chip manufacturer. Google has reportedly placed an order for over 3 million of its in-house TPUs (Tensor Processing Units) to be made by Intel in 2028. Meanwhile, Nvidia is quietly testing Intel’s latest 18A process for its next-gen “Feynman” GPU architecture.
It’s a stunning vote of confidence for a company that many had written off just a few years ago. It’s also a pretty clear sign that the AI supply chain is hitting a wall, and finding a second supplier isn't just nice to have anymore. It’s a survival strategy.
So, what exactly is going on? Why is Google, one of TSMC’s best customers, knocking on Intel’s door? And can Intel actually deliver?
Let’s break it all down.
The TSMC Crunch: How a Capacity Crisis Created Intel’s Opening
The short answer to “Why now?” is actually pretty simple: TSMC is bursting at the seams. Four people with direct knowledge of the discussions described the Taiwanese chipmaking giant struggling to meet surging demand for its manufacturing capacity, with constraints felt most acutely in advanced packaging lines.
You see, it’s not just about making the tiny transistors anymore. It’s about putting them together. TSMC’s CoWoS (Chip-on-Wafer-on-Substrate) packaging technology is the gold standard for stitching together high-bandwidth memory (HBM) and logic chips. But demand has exploded so fast that TSMC literally can’t keep up, creating a massive bottleneck in the AI supply chain.
This is the vulnerability Intel has been waiting for.
Side note: It’s a pretty wild reversal of fortune. For years, TSMC was the nimble giant eating Intel’s lunch. Now, Intel might actually be in the right place at the right time.
Inside the Deals: What Google and Nvidia Are Actually Testing
The headlines talk about “backup manufacturing.” But if you dig into the details, Google and Nvidia are taking two very different approaches to this Intel experiment.
Google’s 3 Million TPU Bet (Firm Handshake)
Google isn’t just window shopping. After months of testing Intel’s advanced packaging technology, Alphabet reportedly pulled the trigger. They’ve placed a formal order for more than three million of their custom Tensor Processing Units to be produced by Intel in 2028.
Why this matters: Google is already selling TPU compute capacity to giants like Apple and Meta. By spreading its production across both TSMC and Intel, Google insulates its Cloud business from a catastrophic supply disruption. Morgan Stanley estimates Google could produce over 6 million TPUs across 2027 and 2028, meaning Intel could end up handling almost half of that volume.
Nvidia’s 18A and “Feynman” Architecture Gamble (Testing the Waters)
Now, Nvidia is playing it cooler.
While Google has signed on the dotted line, Nvidia hasn’t placed a firm order yet. But they are definitely in the room. According to the report, Nvidia is running early trials, including multiproject wafer runs, on Intel’s most advanced manufacturing process, the 18A node. The specific target? Nvidia’s upcoming “Feynman” GPU architecture, scheduled for a 2028 release. They’re reportedly testing whether Intel can build a processor that fuses four graphics chips into a single, powerful unit.
For Nvidia, this feels less about immediate volume and more about leverage. If Intel can prove itself, Nvidia suddenly has a strong negotiating position with TSMC, which currently holds all the cards.
Beyond Chips: Why Advanced Packaging (EMIB) Is the Real Game-Changer
Here’s a detail most casual observers miss. This isn’t just about Intel’s 18A chips. It’s about how you connect them.
Google is reportedly looking closely at Intel’s EMIB (Embedded Multi-die Interconnect Bridge) technology. Think of EMIB as a super-fast, tiny subway system that shuttles data between different chiplets inside a processor. It’s Intel’s direct competitor to TSMC’s CoWoS.
Why does this matter? Because even if Intel’s chip-making isn't perfect yet, their packaging tech might be good enough to win deals. MediaTek has already signed on to use Intel’s EMIB-T for Google’s next-gen TPU custom AI chips. And SK Hynix is testing the compatibility of its HBM memory modules with Intel’s packaging technology.
This is like a restaurant that can’t cook the steak perfectly but makes the world’s best sauce. For now, packaging is the sauce that gets Intel a seat at the table.
Is Intel Ready? 18A Yields, Capacity, and the Foundry Reality Check
Alright, let’s pump the brakes for a second. We need to talk about the elephant in the room. Can Intel actually execute at scale?
The company has a long history of over-promising on process nodes and under-delivering on yields (the percentage of chips on a wafer that actually work). While TSMC is the undisputed volume king, Intel’s foundry division posted significant losses and is still burning cash at roughly $5 billion per quarter to build out capacity.
The question isn’t whether Intel’s 18A looks good in a lab. It’s whether they can manufacture millions of chips for Google and Nvidia without the defect rates spoiling the economics. That’s the million-dollar question, and we probably won’t know the real answer until 2027.
Geopolitics & Diversification: The ‘Made in USA’ Incentive
There’s also a political angle here that you can’t ignore.
You might have heard of the “CHIPS Act.” The US government is pouring billions into bringing semiconductor manufacturing back to American soil. Intel is arguably the biggest beneficiary of that push. One analyst noted that beyond the standard need to diversify supply chains, Google and Nvidia have a unique motivation: keeping the US administration happy.
Supporting Intel means supporting US-based manufacturing. At a time when the Taiwan Strait remains a geopolitical flashpoint, having a stateside backup isn't just a nice business hedge. It's becoming a strategic necessity.
Intel Stock Soars: What This Means for Investors
Disclaimer: I am not a financial advisor. But market reactions don't lie.
On the back of this report, INTC shares surged between 9% and 13% in pre-market and early trading. Investors finally saw something they’ve been waiting for: major customer validation. Google’s order suggests that Intel’s turnaround story under CEO Lip-Bu Tan is starting to generate actual cash flows from external customers, not just internal Intel chips.
If Nvidia converts its tests into a full commitment, the Intel stock narrative changes completely, shifting from a “comeback gamble” to a legitimate “second act” in the AI hardware story.
The news isn’t that Intel is “replacing” TSMC. Let's be clear, that’s not happening overnight. The real headline is resilience. By having a credible second source for manufacturing, the AI industry just got a whole lot safer.
Google and Nvidia consider Intel as a backup chip manufacturer precisely because the stakes have never been higher. A single outage, a single geopolitical event shutting down a TSMC fab, would cripple the global AI economy. By quietly bringing Intel into the fold, these tech giants are building a moat around their own futures.