Groq Raises $650M After Nvidia’s Leadership Raid
AI chipmaker Groq just raised $650 million and started rebuilding its leadership, roughly six months after Nvidia gutted the company in one of the year’s strangest deals. According to TechCrunch AI, Groq confirmed the round on Monday, led by Dallas late-stage firm Disruptive (whose founder Alex Davis also chairs Groq) and Fort Lauderdale hedge fund Infinitum. The company didn’t disclose a new valuation. It was last worth $6.9 billion after a $750 million round in September.
What stands out here is the timing. This money comes right after Nvidia signed a non-exclusive license for Groq’s core technology in December and hired away founder and CEO Jonathan Ross, president Sunny Madra, and other staff. That’s the now-familiar “not-acqui-hire” pattern: a rival pays investors a fat IP licensing fee, poaches the critical talent, and leaves the company technically intact but stripped for parts.
What actually happened
- The raid: Nvidia licensed Groq’s tech and pulled out Ross (who helped build Google’s Tensor Processing Unit) and other leaders. Co-founder Doug Wightman stayed and stepped up as CEO.
- The fallout: Groq built a chip called a language processing unit, or LPU, for inference. With Nvidia now holding that IP, Nvidia rolled out its own Groq 3 LPX inference hardware system at GTC in March. Groq’s former crown jewel is now powering a competitor’s product.
- The pivot: Groq has shifted its weight to its “neocloud” business, the inference cloud that Madra ran after Groq bought his analytics startup Definitive Intelligence in 2024.
- The rebuild: New hires include COO Alan Rice (ex-xAI and Meta), CTO Sinclair Schuller, and CPO Rakesh Malhotra, who previously worked together at Apprenda and Nuvalence.
Why it matters
The cloud business is doing real volume. TechCrunch AI reports it now runs 13 data centers across North America, Europe, the Middle East, and APAC, serves more than five million developers and thousands of AI companies, and processes trillions of tokens a week. That’s a credible foundation, not a press-release fantasy.
But here’s the hard part. Groq’s whole edge was custom hardware built for fast, cheap inference. Now that hardware IP is shared with Nvidia, the most powerful player in the entire AI stack. So the question is no longer “can Groq build a better inference chip?” It’s “can Groq run a better inference cloud than everyone else, using technology its biggest rival also owns?” That’s a much steeper climb.
Inference is exactly where the demand is heating up. As models move from training to serving billions of real queries, the cost of running them is becoming the main event, and VC money is pouring into anything that makes inference faster or cheaper. That demand is real. So is the competition. Groq is now fighting in a crowded lane without the hardware moat that made it special.
The bigger pattern
This is the third high-profile not-acqui-hire in about a year. Microsoft did one with Inflection, Meta did a $14.3 billion deal with Scale AI, and Nvidia did this one with Groq. The structure lets a giant grab talent and tech while sidestepping the antitrust scrutiny a full acquisition would trigger. Regulators are watching, but the deals keep happening.
The encouraging signal for Groq: companies seem to survive these raids. Scale AI’s CEO Jason Droege told Forbes the business rebounded after Meta’s deal and is on track for $1 billion in revenue. So a gutted company can come back. Whether it comes back as a leader or a survivor is the open question.
What to watch next
- Valuation clues: Groq stayed quiet on its new number. If it raised at a flat or down round from $6.9 billion, that tells you how investors really price the post-Nvidia business.
- Neocloud growth: Token volume and developer count are the metrics to track. If they keep climbing, the pivot is working.
- Margin pressure: Without exclusive hardware, Groq competes on price and reliability against Nvidia’s own clusters. Watch whether it can hold its cost advantage.
Groq has a shot, and $650 million buys a lot of runway. But it’s now playing offense in a market where it used to set the rules. You can find the full details at the original TechCrunch AI report.