Groq vs Thinking Machines Lab

Hardware Specialization Returns For AI Inference Era

The a16z newsletter article revisits the historic debate between general-purpose and specialized chips, arguing that the pendulum has swung back in favor of specialized hardware for AI inference. It details how general-purpose CPUs won in the 1990s due to volume-driven performance gains, but shifts to GPUs and the end of Dennard scaling changed the landscape. The article highlights that specialized chips like Nvidia's LPU (formerly Groq), Cerebras, and d-Matrix are now complementary to GPUs, delivering better performance per watt for heterogeneous inference workloads. It reports that GPUs have surpassed CPUs in revenue, and that inference demand is expected to grow 24x while power availability only doubles by 2030. The author, from Gimlet, advocates for orchestrating diverse accelerators to maximize efficiency.

• GPUs generated more revenue than CPUs for chip companies.
• Nvidia acquired Groq, now known as Nvidia's LPU.
• Inference demand is projected to increase 24x from 2026 to 2030, while power availability only doubles.

Nvidia's $20B Groq Deal Faces Stockholder Lawsuit

A lawsuit filed by former Groq engineers Joshua Rubin and Benjamin Serebrin alleges that Nvidia's $20 billion deal with Groq "squeezed out" stockholders. The complaint, filed on October 2 at the Delaware Court of Chancery, claims the deal offered a "lowball" price and that Groq's board approved the transaction without a required stockholder vote. The lawsuit argues that a conflicted board cost stockholders billions of dollars, pointing to $17 billion allocated to a non-exclusive license and $3 billion in Nvidia RSUs for employees. Groq dismissed the lawsuit as "meritless," stating the licensing agreement with Nvidia delivered exceptional value. Nvidia has not commented. The deal, announced in December, involved a licensing agreement for Groq's inference technology, with CEO Jonathan Ross and other leaders joining Nvidia, but Groq remains an independent company.

• Former Groq engineers Joshua Rubin and Benjamin Serebrin filed a lawsuit on Oct. 2 alleging Nvidia's $20B deal with Groq shortchanged stockholders.
• The lawsuit claims Groq's board approved the transaction without a required stockholder vote and that the deal cost stockholders 'billions of dollars.'
• The deal allocated $17 billion for a non-exclusive license and $3 billion in Nvidia RSUs for moving employees.

Groq Founder: Poor Leadership Cost Company Years

Jonathan Ross, founder of AI chip maker Groq, admitted in a podcast that his management failures initially cost the company three to four years. A former Google engineer, Ross founded Groq in 2016 to develop Language Processing Units (LPUs) as alternatives to Nvidia GPUs. He acknowledged being 'a terrible leader' at the start, citing mistakes in hiring employees who were not self-sufficient and giving them too much responsibility, leading to stagnation. He learned to focus on people management and became more selective in hiring. In December 2025, Nvidia signed a $20 billion licensing and talent deal with Groq, after which Ross and other key engineers moved to Nvidia, where Ross now serves as chief architect for software. Groq continues as an independent company led by Adam Winter. Similar leadership challenges were shared by Figma CEO Dylan Field and Duolingo CEO Luis von Ahn.

• Jonathan Ross founded Groq in 2016 to develop AI inference chips (LPUs) as alternatives to Nvidia GPUs.
• Ross admitted that his poor leadership set back Groq by three to four years.
• In December 2025, Nvidia signed a $20 billion license and talent deal with Groq.

Putting Task Expertise into RL Achieves State-of-the-Art Performance on Text-to-SQL

Thinking Machines Lab announced two new updates: a research breakthrough in RL for text-to-SQL and new safety research grants.

Barret Zoph Joins Google as VP of Research

Barret Zoph, co‑founder of AI startup Thinking Machines Lab and a former OpenAI employee, has taken a role as vice president of research at Google. Zoph previously co‑founded Thinking Machines with Mira Murati after leaving OpenAI in October 2024, briefly returned to OpenAI in January 2026 to lead AI enterprise sales, and departed in June 2026 after a five‑month stint. Tech reporting notes Zoph was fired from Thinking Machines earlier this year. Google said it expects Zoph to contribute reinforcement learning and post‑training expertise to its Gemini efforts. The move is one of several high‑profile executive shifts in the AI industry this year.

• Barret Zoph co‑founded Thinking Machines Lab and later rejoined OpenAI.
• Zoph spent five months at OpenAI in 2026 heading AI enterprise sales and left in June 2026.
• Zoph was reported to have been fired from Thinking Machines earlier in 2026.