The boom in generative AI is making the fortunes of Nvidia, Broadcom and more
The ecosystem of companies revolving around generative AI is literally exploding: Nvidia, Broadcom and Tsmc are just some of the companies that are making huge profits
The boom of AI and in particular of the generative one is exploding the value of the entire technological ecosystem involved in the activation of increasingly popular applications such as ChatGPT. This ecosystem, which ranges from chip makers to software to network equipment, was recently photographed by an Economist article, which names and lists the colossal profits of its protagonists such as Nvidia , a company that has just grazed a capitalization of nearly a trillion dollars.
The AI Gold Rush
What the Economist calls the "gold rush" of AI triggered by the advent of generative systems such as ChatGPT promises rich profits for the companies that produce the necessary technology.
The most striking case is of course that of Nvidia, a company that makes semiconductors for AI servers. On May 24, Nvidia beat analyst forecasts for Q1 2023 revenue, posting $11 billion in revenue. The next day its stock price soared 30% and the company's total market value approached a trillion dollars.
An index of the thirty leading companies in this revolution has just been introduced on Wall Street: since the launch of ChatGPT last November, this index has jumped 40%, compared to 13% for the more famous Nasdaq of stocks. technological.
Boom for chip manufacturers and designers
The primary beneficiaries of the AI boon are chip manufacturers and designers such as AMD and TSMC, as well as producers of the computing infrastructure needed to support AI computational processes.
Behind the success of this sector is the continuous growth of the computing capabilities of the latest AI systems including generative ones. The latest version of ChatGPT, GPT-4, analyzes data using approximately one trillion parameters, five times more than its predecessor.
As Amin Vahdat, head of AI infrastructure at Google Cloud Platform, explains to the Economist , the size of AI appliances has grown tenfold in the past six years.
The most obvious winners in this race are chipmakers, whose products have been affected by a surge in demand. The latter in turn affects the profits of chip designers. Bank UBS estimates that the value of new demand for specialized chips will grow by at least $10-15 billion within the next two years.
In something of a virtuous circle, the entry of more powerful chips onto the scene benefits related manufacturers such as TSMC, whose share price jumped 10% after news of Nvidia's ultra-profits broke.
Swirling increase of network kits…
The manufacturers of network equipment such as routers and switches also benefit from this whirlwind of wealth and profits. According to the estimates of 650 Group, a research firm, the market for these kits will grow annually by 40% in the coming years, reaching a value of nine billion dollars in 2027.
Not only companies such as Nvidia, which has a 78% market share in this sector, are taking advantage of this, but also competitors such as the Californian Arista Networks, whose stock market value rose by 70% last year.
Another manufacturer like Broadcom saw its annual sales of specialty chips quadruple in 2023 to $800 million.
… and servers
The AI boom is also blowing a good wind on the companies that assemble the servers. According to the estimate of the market analysis company Dell'Oro Group, the various computer centers of the world will increase the share of specialized servers for AI from the current 10% to 20% within the next five years.
Server makers like Taiwanese Wistron and Inventec, which power Amazon and Microsoft's massive clouds, are also performing well. Another Taiwanese company like Wiwynn recently reported that its sales in the AI sector account for more than half of the orders received.
The role of software
All this hardware naturally needs dedicated software to function. Some of these programs are made by the same companies that produce the hardware: this is the case, for example, with Nvidia's Cuda platform.
But the market is now teeming with companies that produce management software (such as Datagen, Pinecone, Scale AI) or applications that implement linguistic models (such as HuggingFace and Replicate). According to PitchBook's calculation, there are at least eighty startups operating in this sector and many of these can have access to the almost unlimited resources of venture capital.
Just like with hardware, the main customers for AI software are the cloud giants. According to the Economist, Amazon, Alphabet and Microsoft will allocate resources equal to 120 billion to this sector in 2023 alone, much more than the 78 billion invested in 2022.
Computing centers
The last category of subjects that rejoices in the turbo-expansion of AI are the owners of the spaces that house data centres.
Specialist companies such as Equinox and Digital Realty are rapidly acquiring new properties by competing with the largest asset managers. Two years ago Blackstone bought Realty Trust, one of the largest computer center operators in America, for ten billion dollars.
The startups
Such is the buzz with AI that it creates huge opportunities for new players. AI startups are literally proliferating and raking in huge sums from interested investors.
One of them, Lambda, raised a whopping $44 million in March from investors like Gradient Ventures, an arm of Google, and OpenAI cofounder Greg Brockman.
This is a machine translation from Italian language of a post published on Start Magazine at the URL https://www.startmag.it/innovazione/intelligenza-artificiale-generativa-ecosistema/ on Sat, 03 Jun 2023 05:17:49 +0000.