In today’s Finshots, we unpack why the world’s hottest AI-chip company just wrote a $5 billion cheque to a wounded rival, and what this uneasy alliance really fixes (and doesn’t).

But here’s a quick sidenote before we begin. We’re gathering real reviews for Health & Term Insurance policies to feature on the Ditto Insurance website, and we’d love your help! It’ll only take a minute or two:

• Click here for Health Insurance

• Or here for Term Insurance

P.S. Your review might just make it to our policy pages.

Thanks for helping us out!

Now, onto today’s story.

The Story

Before we get into the actual story, here’s a little analogy that will help:

Did you know that white tea and green tea come from the same plant? However, white tea costs as much as ₹10,000 for a kilogram. It consists of the finest and the most tender of leaves, processed with the utmost care, and hand-plucked just before dawn. You miss this opportunity, and you will still get tea, just not the ‘best’ kind. And the price comes down from ₹10,000 per KG to, say, a few hundred rupees per kilogram.

Computer processors are just like that. Processors like Intel’s i3, i5, i7, or i9 often come from similar semiconductor wafers. What matters is yield: chips that fail to meet the specifications (for the number of cores, cache, speed, power, etc.) may be classified into lower‑tier models. Larger, more demanding designs have a higher chance of defect, so fewer make it as i9‑class chips and may be reduced to a lower class.

And that’s kind of where Intel’s troubles began. As rivals like TSMC (the Taiwan chip giant) had excellent yields from their wafers, Intel stumbled. Its most advanced manufacturing nodes delivered poor yields. Too many errors. Too much wastage. And slowly, what was once the gold standard of chipmaking began falling behind, and Intel’s leadership in foundry technology eroded by 2015. By 2021, Intel even outsourced GPU (graphic processing unit) production to TSMC due to ‘internal yield concerns’. Simply put, this is where a tech company faces manufacturing problems in the sense that the % of functional and defect-free products is lower than expected.

As the years passed, the cracks widened. Intel’s foundry business became a money drain. Its CPUs lost the performance crown to the US tech giant AMD. And its most ambitious chip designs faced delay after delay. So much so that in a world increasingly dependent on silicon, the OG chipmaking giant looked like it was being left behind. And eventually, investors lost confidence.

Until a couple of weeks ago.

That’s when the US government announced that it would acquire a 10% non-voting stake in Intel, backed by a funding commitment of $8.9 billion. A signal, perhaps, that Washington still sees Intel as a strategic asset worth saving. And now, Nvidia is making its move too – it has agreed to buy roughly a 4% stake in Intel, valued at around $5 billion. 

What’s interesting here is that Intel is going to manufacture custom chips for Nvidia’s data-centre platforms, as well as personal computer GPUs. Nvidia will knit the pieces together with NVLink, its high-speed interconnect, so the CPUs can feed GPUs faster and juggle AI workloads more efficiently.

But here’s the big question: Why would Nvidia, the market leader in GPUs, invest in a struggling rival like Intel?

The answer is basically less dependence on TSMC.

The vast majority of Nvidia’s advanced chips are made in Taiwan by TSMC. That’s terrific while geopolitics behaves, but terrifying when it doesn’t.

And Nvidia needs a Plan B. That’s where Intel steps in. Sure, Intel can’t match TSMC on cutting-edge yield or node maturity. But it still has world-class expertise in x86 chips (chips like Pentium and Core) manufacturing, packaging, and testing. It’s one of the most widely used chips in laptops today. So these are areas that Nvidia can tap into without compromising too much on quality. 

There’s also the political upside here. By working with Intel, Nvidia cuts some of its reliance on Taiwan and brings part of its supply chain closer to the US. That helps it dodge steep tariffs and soften the blow from China’s new GPU bans. In short, it’s Nvidia making sure it’s not putting all its eggs in one basket.

And that brings us to the real meat of the deal.

This partnership is two-pronged. For data centers, Intel will build Nvidia’s custom x86 CPUs that Nvidia can integrate into its AI infrastructure platforms and offer to the market.

For personal computing, Intel will build and offer to the market x86 chips (SoCs) with Nvidia’s RTX chiplets, powering a wide range of PCs that demand top-tier CPU–GPU integration.

Sure, these are not the highest-end chips that Nvidia makes, but it gets access to Intel’s vast x86 ecosystem, which Intel still has a strong hold on. 

But mind you, this is easier said than done.

Because Intel will still need to prove that it can deliver chips on time, at scale, and without yield disasters. And the chips it’s making for Nvidia aren’t going to be the bleeding-edge AI GPUs that TSMC currently churns out. So no, this isn’t a turnaround moment.

Add to this, Intel’s financial troubles haven’t exactly gone away. In FY24, the company posted massive losses of almost $19 billion! And things didn’t improve in Q1FY25 either. The company still bleeds money.

While rivals like TSMC have mastered advanced chip production with excellent yields and tight timelines, Intel is still grappling with consistency. Its manufacturing delays have become infamous, and its cost structure doesn’t quite allow for the kind of aggressive pricing the market now expects.

In the meantime, AMD has gained ground in the personal computer race, offering better value-for-money chips and snapping up market share.

That double whammy of foundry failure and market erosion has left Intel in a tough spot, which is why this Nvidia deal matters. It’s not a miracle fix. But it’s the first sign of confidence from Silicon Valley in a while.

And it’s not just Nvidia or the government betting on Intel. Investors seem to think this might be the beginning of a slow crawl back. The market reacted swiftly to the announcement, Intel’s stock jumped 22% in a day, the sharpest rally it’s seen in years. AMD, on the other hand, saw its stock fall by 5%. Maybe that’s a sign that investors believe this alliance could tilt the scales, even if slightly.

So, we’re unlikely to see Intel manufacturing Nvidia’s most powerful AI chips anytime soon, those still belong to TSMC’s cutting-edge fabs. However, this partnership gives both companies something they need. For Nvidia, it’s risk mitigation. For Intel, it’s relevance. 

And sometimes, that’s all a company needs to stay in the game.

But tell us what you think: Could this be the start of Intel clawing its way back to the glory days? Or does it simply add one more weak link to an already fragile supply chain?

Until then…

Sharing is caring. If you found this story useful, please do consider telling your friends about it. All you have to do is send it on WhatsApp, LinkedIn, or X and ask them to subscribe! :)