NVIDIA's Meteoric Rise Threatens Google's Empire – Forbes

In this photo illustration, the logo of NVIDIA is seen displayed on a mobile phone screen with AI … [+] (artificial intelligence) written in the background. (Photo Illustration by Idrees Abbas/SOPA Images/LightRocket via Getty Images)

The unthinkable is happening; Google GOOG is being disrupted by a small software startup. Now the search giant finally has a plan to fight back. It might not matter.
The chief executive at Alphabet (GOOGL), Google’s parent company announced in April that two of its advanced research divisions are merging to build general artificial intelligence applications.
Investors should buy Nvidia (NVDA) shares into any weakness. Let me explain.
Artificial intelligence research has never been hotter. OpenAI, the small company behind ChatGPT is taking the computing world by storm with a next generation chatbot that appears to behave like a human. ChatGPT can do everything from writing music, poetry, and software code, to carrying out casual conversations about recipes and philosophy. This could be catastrophic to Google’s core business, internet search.
ChatGPT launched in December and added 1 million active users in the first week. By the end of January that number had swollen to 100 million. For context, this growth is orders of magnitude faster than Instagram, TikTok, and Google Search, that needed almost 1 year to reach 100 million users, according to a report from Reuters.
Sundar Pichai, Alphabet’s chief executive is rightfully concerned.
His plan is to merge Google Brain and DeepMind, its advanced research divisions. While the new subsidiary may lack a creative name, Google DeepMind, it combines a wealth of real-world AI accomplishments.
Pichai noted in early April in his blogpost that since 2016 the research teams helped refine Search, YouTube, Gmail, and build the computational photography that made Pixel smartphone cameras best in class. The academic triumphs include Transformers, AlphaFold, and TensorFlow and JAX, the foundational frameworks that facilitate large language modelling used for general AI system development.
It is undeniable that Google has the intellectual talent and LLM knowhow to fight back against OpenAI. This misses the larger point, though. This isn’t really a story about ChatGPT. AI is a disruptive gamechanger.
Research from Lambda Labs shows that ChatGPT could have been created in only 34 days, at a cost of a mere $5 million.
Analysts at Lambda figure that 355 years would have been required using traditional methods to train LLMs. Developers at OpenAI blew up the norm. They used 1,023 A100 graphics processing units scaled into a supercomputer. The Nvidia GPUs used massively parallel design to bust through the complex math in short order.
This feat of brute force pushed demand for the Nvidia gear through the roof. Two year old A100s are now selling on eBay for up to $45,000 apiece. The retail cost is $10,000, if you can find one.
Nvidia reported in November that Microsoft MSFT (MSFT) would acquire “tens of thousands” of these GPUs for its Azure cloud computing subsidiary. And a note last week at Business Insider indicated Elon Musk has purchased 10,000 units to bolster an internal AI project at Twitter.
It’s not hard to see where all of this is going. There is an AI gold rush.
Developers understand it is now possible to build extremely disruptive products with relatively few engineers and little money, using AI. ChatGPT is only the beginning of what is possible. An entire ecosystem is developing around staking claims and building businesses. And Nvidia has a monopoly on the shovels and pickaxes. It’s really an incredible story, with unprecedented upside from here.
Nvidia’s AI hardware is running at every major datacenter in the world, according to Jenson Huang, chief executive officer. Chieftains at Microsoft’s Azure, Amazon Web Services, and even Google Cloud are scrambling to buy gear ahead of a new wave of AI development. More importantly, Nvidia controls the direction of the entire AI stack through the governance of its CUDA open source software platform.
At a share price of $271.19 Nvidia stock trade at 44.8x forward earnings and 24.8x sales. The stock has also run up immensely since October when it touched $110 amid stories of semiconductor gluts and China sales bans.
Investors should buy pullbacks to the $245 area.
Discover the secrets to successful investing with our Strategic Advantage newsletter. Try it now for just $1!