Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: Nvidia Strikes Deal with Competitor Groq to Sell Inference Technology in Simple Termsand what it means for users..
As the world winds down to celebrate the Christmas and New Year weekends, AI chipmaker Nvidia took another stab at continue market leadership through a non-exclusive agreement with its competitor Groq. Under the deal, the company will acquire assets worth $20 billion to enhance its position as the dominant chipmaker.
In a blog post, Groq further noted that founder Jonathan Ross, President Sunny Madra, and other members of the team will join Nvidia to help advance and scale the licensed technology. However, Groq would continue to operate as an independent company with Simon Edwards taking over as the CEO.
It appears likely that the deal was brokered by Disruptive, which led Groq’s latest round of funding in September when the company had raised $750 million at a $6.9 billion valuation. The news of the deal, shared by Disruptive CEO Alex Davis with CNBC, accentuates efforts by Nvidia to maintain market leadership in the AI-chip making space where Google, Amazon and others are now attempting to get their fingers into the pie.
Disruptive had invested over half a billion dollars in Groq since its founding back in 2016. In the recent fund raise, a slew of investors had committed funds to the company including Blackrock, Neuberger, Berman, Samsung, and Cisco. What is interesting is that Donald Trump Junior’s 1789 Capital is also one of those putting dollars into the venture.
Is it an acquisition or just a deal to acquire some assets?
Davis also clarified to CNBC that Nvidia would get all of Groq’s assets barring its cloud business. This represents Nvidia’s largest purchase ever, though the company claims that it was not actually an acquisition in the true sense. In the past, Nvidia had acquired Israeli chipmaker Mellanox for around $7 billion.
Nvidia boss Jensen Huang told staffers that the latest deal would expand Nvidia’s capabilities as they integrate Groq’s low-latency processors into the Nvidia AI architecture. This would extend the platform to serve as a broad range of AI inference and real-time workloads, Huang said in the email to employees that CNBC accessed.
One can safely say that by embracing Groq into its fold, Nvidia has managed to create a broader offering for companies seeking to grow their AI capabilities. While the Nvidia GPU emerged as the industry standard, Groq’s chip called the LPU claims to run LLMs at a pace ten times faster with just one-tenth of the energy requirement.
Incidentally, it was Jonathan Ross, working at Google then, that helmed that company’s chip innovation through the tensor processing unit or TPU, a custom AI accelerator chip. We had reported a month ago about the GPU versus TPU battle and why it was more likely that Nvidia might not be too worried of competition. This latest acquisition proves that Jensen Huang is keeping close watch of the industry and doing what’s right to retain the Nvidia edge.