Google vs Nvidia: Chipmaker buys Groq for $20B; reason is $250B hit in market value

Nvidia has acquired the assets of AI spot startup Groq successful a $20 billion. The woody represents a 3x premium implicit Groq’s valuation conscionable months agone and it is being viewed arsenic a antiaircraft play to support Nvidia’s dominance against a rising inclination of tech companies preferring customized silicon, specifically Google TPUs.

The determination is said to beryllium driven by 2 cardinal factors: the request for outgo ratio and the request for higher processing speeds. Nvidia is presently struggling to conscionable the overwhelming planetary request for its GPUs, and tech companies are looking for specialised alternatives arsenic AI contention hots up.Nvidia mislaid billions aft Meta reportedly struck woody with Google for TPUsEarlier this year, Nvidia issued a connection defending its marketplace presumption aft a study claimed that 1 of its biggest customers, Facebook genitor institution Meta, is successful precocious talks to walk billions connected Google's competing AI chips.

This deed Nvidia's stock, erasing astir $250 cardinal successful marketplace value.

Difference betwixt Nvidia GPU vs Google TPU

Before we leap connected to Nvidia vs Google debate, it indispensable beryllium understood the quality betwixt the products offered by some the companies. Nvidia GPUs person are ‘General Purpose’, which means they tin beryllium utilised for AI, gaming, crypto mining and technological simulations. Google TPUs, however, are specialised chips designed for speeding up the “tensor” mathematics that powers instrumentality learning.

Apart from flexibility, different differences see velocity and efficiency. While Nvidia’s GPUs connection fantabulous speed, it has overhead for wide tasks and are costly. Google’s TPUs are ultra-fast for circumstantial AI training/inference (running an AI exemplary and getting an reply with little lag), and tin beryllium mixed-and-matched with Nvidia GPUs to summation ratio and little cost.

Why Nvidia struck woody with Groq

Groq offers a exertion akin to Google’s TPUs. According to the company’s website:Groq builds accelerated AI inference. Groq LPU AI inference exertion delivers exceptional AI compute speed, quality, and affordability astatine scale. Groq AI inference infrastructure, specifically GroqCloud, is powered by the Language Processing Unit (LPU), a caller class of processor. Groq created and built the LPU from the crushed up to conscionable the unsocial needs of AI. LPUs tally Large Language Models (LLMs) and different starring models astatine substantially faster speeds and, connected an architectural level, up to 10x much efficiently from an vigor position compared to GPUs.By acquiring Groq’s accelerated and businesslike LPU exertion to complement its almighty GPUs, Nvidia is trying to neutralise the hazard of a rival offering a low-cost alternate that could person ‘shrunk’ the company’s dominance to the “training-only” market. The strategical woody volition let Nvidia to seizure the full lawsuit lifecycle done a tiered offering – directing premium clients toward high-end GPUs for heavy-duty exemplary grooming and steering price-sensitive users toward LPUs for high-speed inference.