Chasing artificial intelligence is becoming expensive for big tech companies, as some of them are also revising their forecasts of spending on the technology. Amazon is expanding its data center infrastructure to Saudi Arabia and Mexico, and it has also announced an allocation of $10 billion for two giant data centers in the US.
Tech giants are parting with billions of dollars for AI
Big tech names are also sensing the demand for AI among all levels of businesses, which is why they are increasing their spending on core AI infrastructures. Microsoft spent $14 billion in the first quarter of 2024, which it says will increase over the year. Meta investors freaked out when Meta CEO Mark Zuckerberg revealed the hefty amounts the company will be spending over the years on AI, which are estimated at between $35 billion and $40 billion.
Investors mostly look for early profits, so a dip in the company stock was expected, which fell by 19% during extended trading and erased $200 billion from the company’s market cap. But Zuckerberg was also prepared for it, as he mentioned that the company has been through tough times before. He was hinting at the Metaverse adventure, which still costs billions of dollars just to operate and doesn’t generate much revenue.
Seeing these figures, it becomes clear that AI infrastructure is not an easy bet for the bigger tech giants. For example, Alphabet, the parent company of Google, increased its spending by 91% YOY and spent $12 billion during the previous quarter. Considering the resources needed to train the new generation of large AI models, including computing power and energy requirements, it’s predictable.
But at a time when big firms are forking over billions, there are two main beneficiaries of the total spend, with another secondary as well. They are none other than smaller companies and semiconductor makers, along with a third sector added: those who hold huge amounts of human-generated content like newspapers and forums, for example, Reddit.
Smaller companies now have to spend less
Speaking of our first nominee, the small companies are benefiting from the increased infrastructure facilities and competition in the artificial intelligence sector as the choices of AI tools are increasing for them and the prices are going down. Katie Prescott gave a wonderful example in The Times. She mentioned Olio, a platform for those who want to give away unwanted food items. She said that back in 2023, Olio estimated that it would cost $80,000 for it to leverage AI for one year, but after inquiring just after three months, the cost quoted was less than 50% of the previous amount, which later came down to $4000.
OLio is not alone here; many small corporations have the same story, as most of them are now paying less than 50% of what it was originally. Though it’s still not the same for everyone, the competition between the tech giants and the availability of models like the Phi-3 from Microsoft, which can even run on a smartphone without internet access and performs equal to models that are 10 times their size, are definitely helping small businesses, and we expect the price to come further down.
Semiconductor makers like Nvidia are also benefiting from the AI wave, as along with new innovation, the allocation of billions by big tech firms is funneled down to a handful of semiconductor makers like Nvidia, AMD, and Taiwan Semiconductor, as demand for their silicon chips and processors has increased many fold due to data center expansions.
Last but not least, newspapers and forums are also making deals with AI firms to provide access to their content for model training through licensing. For example, Reddit recently said that they will prefer business conversations over legal action, and Financial Times is also said to have made a deal with OpenAI for content access and usage, according to media reports, but details are not known yet. These are just the ones benefiting directly from AI development; there are numerous businesses leveraging AI for every type of work one can dream of.
