Upstart Threatens NVIDIA (NASDAQ: NVDA) With 20x Chip Improvement

This post includes affiliate links. If you purchase anything through these affiliated links, 247wallst.com may earn a commission.

Key Points:

• Etched, led by Gavin Alberti, claims its Soho chip could outperform NVIDIA's by 20 times.
• The startup raised $120 million and attracted top talent from major tech firms.
• The impact on the AI chip market will depend on further testing and real-world performance.
• With the competition heating up for Nvidia, investors are already starting to look ahead to 2025’s best investments. See what the hype is all about here .

Doug and Lee discuss a new competitor in the AI chip market, a company called Etched, founded by Gavin Alberti, a young entrepreneur with a Harvard background. Etched claims that its new chip, named Soho, could be 20 times faster than NVIDIA's (NASDAQ: NVDA) , focusing on embedding deep learning architecture directly into the chip. While this ASIC-based approach could revolutionize the AI chip industry, there are risks if the technology doesn't gain acceptance. Etched has raised $120 million, attracting top talent from major semiconductor companies, indicating potential in the tech world. Doug and Lee agree to follow this closely as more details emerge about its commercial viability.

Watch the Video

https://videos.247wallst.com/247wallst.com/2024/09/Upstart-Threatens-Nvidia-With-20x-Chip-Improvement.mp4

So NVIDIA has some competition.

To me, it's unexpected because it's competition that's really technology.

Tell me about it.

Boy, and you knew it was coming.

There's a company that is reasonably new called Etched.

And Etched is run by a former Harvard dropout, kind of like your friend Bill Gates.

And his name is Gavin Alberti.

And they've created a chip that they feel could be 20 times or more faster than NVIDIA's chip.

And they intend to do this by embedding, and I'm reading to some degree on this, their chip is called the Soho, or Sohoo, depending on how you pronounce it.

And their chips are hard-coded for individual architecture.

And they only support transformers, which was a technology Google put together years ago.

And the whole point is here, they're literally embedding that into the chip, and they're saying that it'll be much, much faster.

Does that sound right?

Well, I don't know.

I mean, they're using an ASIC chip versus a GPU chip.

And some of the detractors say, well, if the ASIC model isn't accepted, then it doesn't matter.

They're toast.

You know, they're a donut.

But if it does work, and again, they insist, and I'm reading here right now, they can run AI models faster and cheaper than GPUs.

And you remember the transformer is, it's a deep learning architecture that Google came up with that if you could put this in, and remember when chips are etched, that's literally the carving on it that Lam Research makes tools for the semiconductor industry that etches chips in silicon.

So it's kind of a unique process.

And again, I'm no super tech guy.

I'm sure there's super tech guys that will come in and say, oh, I don't know if he knows that much about it, which I'll admit I don't.

But I thought it was interesting because this guy kind of came out of nowhere and he's not very old.

And somehow he has garnered a former chief technology officer from Cypress Semiconductor.

And I'm sure you remember them, Doug, right?

They were big.

They were bought for $9 billion.

And one of their engineering guys was one of the top engineering guys at Intel.

And I don't think either of those guys goes there unless they think they at least have something on the platter.

Yeah.

So it would be unusual if they did.

I mean, it would almost revolutionize this part of the AI business.

But you know, disruptive technology at this point in our lives seems to happen every other day.

So I guess it's possible that there's certainly enough smart money going into this thing that, you know, there's, there's, they've raised 120 million.

And again, you don't raise that kind of money out of the chute unless, unless there's something there that's extremely intriguing.

Now granted private equity and Uber rich people have very deep pockets, but, and so, you know, if you give them $5 million, well, to all of us, that would be a lifetime worth of earnings.

For them, it's, you know, it's on their money clip.

So it'll be, this is a story we're really going to have to follow closely.

And I'm going to get more versed in actually how this chip works.

But we wanted to let our viewers and our readers know that we got an eye on these guys.

And again, it's etched, E-T-C-H-E-D, just like the process that, you know, like I said, Lam research tools do.

And it's going to be interesting to see how Gavin and his crew, and again, they're all young.

They all look like they're 28 or 27.

But hey, God bless them.

That's how technology has thrust the United States to the top of the totem pole.

Listen, Zuckerberg, Gates, Ballmer, all these people.

All of them.

were in their 20s when they started these companies.

I mean, I think Bezos.

Oh, Larry Ellison.

We could name you 20 that were all that young.

And, you know, they fought their way through it.

And I can't remember, was it Hewlett or Packard that Steve Jobs went to early on?

It was one of those guys, you know.

And you're right.

You can't discount their age.

That's for darn sure.

Well, and this is a little bit off topic, but the amount of intellectual property, technology intellectual property that is created in the United States is probably 10x of the whole rest of the world.

Yeah, come on.

It tends to come out of a dozen universities.

I mean, really, the place where these things are incubated, the schools where these are incubated and the people who are related to those schools, it's a tiny universe that IP in the United States and IP worldwide is controlled by a tiny universe of schools that produces these people.

Yep.

Yep.

And, you know, it's something Americans can always be very proud of.

And that's why, you know, it's such a big issue in Washington because our technology, we have brilliant young people and there are men, women, you know, people that have come here from out of the country to go to school.

And then the Chinese just, you know, regularly rip it off.

And that's why, that's our biggest challenge is not intellectuals from other countries.

It's people stealing our intellectual property in China.

It's very true.

Well, look, once all the funds are raised, at some point, there'll be a judgment about whether or not this is commercially viable and scalable.

So let's come back to this when there's some sign that they've actually road tested this stuff.

And people are saying, you know, there's a chance that they're right.

Yeah.

And, and, and they're in the chips, you know, being worked on now and it will be interesting.

I'm going to keep a close eye on it.

I wish I had more detail.

And to be frank with you, I wish I understood it a little bit better.

I mean, I was, I was there through all the technology in the nineties when I was, you know, a wall street sales guy.

And that was a little easier to understand than the depths of the AI chip.

But, you know, we're going to keep a close eye on it and we're going to kind of kind of keep an eye on who else is keeping an eye on them.

And then we'll get back to our viewers and let them know.

Good.

Want to Retire Early? Start Here (Sponsor)

Want retirement to come a few years earlier than you’d planned? Or are you ready to retire now, but want an extra set of eyes on your finances?

Now you can speak with up to 3 financial experts in your area for FREE . By simply clicking here you can begin to match with financial professionals who can help you build your plan to retire early. And the best part? The first conversation with them is free.

Click here to match with up to 3 financial pros who would be excited to help you make financial decisions.