Ex-Intel Exec Raja Koduri Gives a Sneak Peek into his New Startup

I don’t think people should do computer science: Raja Koduri
Listen to this story

Intel Corporation made a significant move in December 2022, splitting its Accelerated Computing Systems and Graphics (AXG) business unit into two distinct entities, namely, Data Center and AI (DCAI) business unit and AXG business unit.

The decision was aimed at aligning the company’s graphics efforts to better compete with industry heavyweights—NVIDIA and AMD. The AXG business unit’s consumer-oriented portion was slated to merge with Intel’s Client Computing Group (CCG), which specialises in developing platforms around the company’s central processing unit (CPU) products. Meanwhile, the teams responsible for managing data centre and supercomputing graphics processing units (GPUs), including the highly anticipated Ponte Vecchio and Rialto Bridge products, will be transferred to the DCAI business unit.

Raja Kodouri, the former head of the AXG division, was to assume the role of Chief Architect, leading the charge in expanding the company’s efforts in the CPU, GPU, and AI domains as well as accelerating critical technical programmes.

In a surprising turn of events, Raja Koduri announced his resignation from Intel Corporation in March 2023, only three months after being appointed as the Chief Architect. This announcement comes at a crucial time for Intel as it prepares to face off against its primary competitor, NVIDIA

Koduri’s departure has caught many industry experts off guard, as he was given a significant position at Intel merely months before his resignation. To gain insight into his decision and what lays ahead, Analytics India Magazine got in touch with him.

Challenges Intel faced 

Koduri has been with Intel since 2017, the year when Intel decided to get into GPUs for the first time in two decades. When asked what were the various challenges he faced competing with NVIDIA while being at Intel, he said that the biggest challenge lies in their (NVIDIA’s) proprietary software ecosystem and the reach it has. 

“Their products are ubiquitous, and they do an excellent job of hooking customers,” says Koduri, “Furthermore, the other major challenge is execution, and specifically, proper execution”.

As per Koduri, Intel experienced major issues with the 10-nanometer chip production that brought the entire product pipeline to a standstill. He says that “the situation was akin to a clogged drainage system”. The most significant hurdle was that it took a considerable amount of time to resolve this. 

“The process of clearing out the old products to make way for the new ones, along with innovative ideas from pioneers like me and Jim Keller, and new architectures, was all hindered by these drainage execution problems. Nevertheless, you’ll start to notice these issues gradually clearing up. Intel has already launched new products, and I anticipate some remarkable developments from them in the next three to four years,” shares Koduri. 

Why he left

However, when all things were sorted out, why did he choose to leave? As per Koduri, he had a major surgery in December and had to take a medical leave due to which Intel decided to give the responsibility of heading the business to someone else. 

“When I took the break, and you know, when you have that space in mind to think about, I made a decision that I want to do something outside, particularly, in the open ecosystem,” says Koduri. 

Koduri added that he wanted to get back hands on to do more software for himself. It’s been a long time since he was actually proactively coding anything. So, he wanted to learn everything since he feels like he hasn’t learned a lot about what’s happening on software. 

However, Koduri believes that he can leverage his knowledge of hardware. As per him, he can utilise his knowledge on hardware to help other companies. “I don’t need to be building hardware myself,” says Koduri. 

He shared that he is now advising companies Tenstorrent, Bodhi Computing, and others. “I’ll be advising several hardware startups as well, on their roadmaps and all,” says Koduri. 

What’s next?

After leaving Intel, there was news that Koduri is going to work on a generative AI gaming startup. Upon inquiring what it was and how he plans to move ahead with it, Koduri says that he had some ideas way back in 2017 as well. 

“But now with the whole progress on large language models, diffusion models, transformers and all that, the ideas that I’ve had in 2017 are very much a reality now,” says Koduri. 

Koduri says he wants to build a platform that connects all these tools, with the best hardware there. He believes that today, the models are not working in real-time. “So, if you want a real-time gaming use case, you want to deliver AI-generated content at 60 frames per second. Like how would you go about doing that?” explains Koduri, “The problem hasn’t yet been solved”.

According to reports, Raja Koduri is in negotiations with Hiranandani-backed data centre operator ‘Yotta’ for a deal for his generative artificial intelligence startup, which he claims will have a significant presence in India.

As a startup, Koduri accepts that he doesn’t know how to solve these problems, “But, I have a few smart people who will figure it out. We may fail, but we will learn from it,” believes Koduri. 

‘Don’t do computer science’: Raja Koduri

When asked why innovation is not happening at the rate humanity anticipated in the 1980s, Koduri believes that the sole reason why we couldn’t deliver the expectations is computer science

“Let’s talk about flying cars,” says Koduri, “You have to solve fundamental physics problems, mechanical engineering problems. They’re real hard problems.”

But Koduri believes when there is so much easy money to be made, writing an app for getting food delivered to my home can make millions of dollars, why would anyone put brainpower into the hard problems? 

“I would like to see people coming back to fundamental problems,” says Koduri.  “In fact, I advise, nobody should join a computer science course. It’s a waste of time, in my opinion,” shares Koduri. 

Koduri believes that if someone is going to university, they should go to electrical engineering, mechanical engineering, fundamental chemical engineering, where they get access to machines, get access to equipment that allows one to experiment with the physical things that one can’t otherwise sit at home and learn. 

“I strongly believe in encouraging fundamental research in India and applaud those who are already dedicated to this pursuit. While I appreciate software and hardware startups that focus on innovative products, such as a smaller and longer-lasting battery, I also want to prioritise supporting startups that aim to develop cutting-edge technology like body computing chips. It’s crucial that we foster a culture of innovation in India, particularly in areas that tackle complex issues like heart disease,” shares Koduri. 

Download our Mobile App

Lokesh Choudhary
Tech-savvy storyteller with a knack for uncovering AI's hidden gems and dodging its potential pitfalls. 'Navigating the world of tech', one story at a time. You can reach me at: lokesh.choudhary@analyticsindiamag.com.

Subscribe to our newsletter

Join our editors every weekday evening as they steer you through the most significant news of the day.
Your newsletter subscriptions are subject to AIM Privacy Policy and Terms and Conditions.

Our Recent Stories

Our Upcoming Events

3 Ways to Join our Community

Telegram group

Discover special offers, top stories, upcoming events, and more.

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Subscribe to our Daily newsletter

Get our daily awesome stories & videos in your inbox
How Generative AI is Revolutionising Data Science Tools

How Generative AI is Revolutionising Data Science Tools

Einblick Prompt enables users to create complete data workflows using natural language, accelerating various stages of data science and analytics. Einblick has effectively combined the capabilities of a Jupyter notebook with the user-friendliness of ChatGPT.

Intel Goes All in on AI

Pat Gelsinger said, there are three types of chip manufacturers, “you’re big, you’re niche or you’re dead”