Monday, December 30, 2024

The Hidden Cost of Innovation: How Tech Giants Are Funding the Backbone of AI

With more usage of AI tech and new technology, many tech giants are investing billions of dollars for AI data centers so they can meet the demand for computational energy that can handle the latest AI infrastructure and workload. Data centers have hefty operating costs but they make the running of AI data centers possible. Data from JP Morgan and New Street Research shows how much big techs are spending on AI data centers, including software, operating expenses, depreciation and electricity.

The most investment on AI data centers is being done by Microsoft, with most spent on GPUs($20 billion) and other AI spend ($20 billion) which makes total capex of $40 billion. The cost spent on training and R&D by Microsoft is $3 billion, cost on inference is $3 billion which makes total operating cost of $6 billion. This makes a total of $46 billion that Microsoft is spending on AI data centers. The second most spendings on AI data centers is by Google, with a total of $33 billion. The total capital expenditure by Google is $29 billion while the operating costs are $4 billion.
Followed by Google, Meta is spending the third most on AI data centers, with $23 billion total capex, $11 billion on GPUs and other chips while $12 billion on other AI spend. The total operating cost of AI data centers that Meta is spending on is $4 billion and this makes a total of $27 billion that Meta is spending on AI data centers. Amazon is also investing a lot on its AI data centers, with a total of $19 billion. $8 billion are spent on GPUs and other chips, while $8 billion is the other AI spend. The total capital expenditure on AI data centers by Amazon is $16 billion while $3 billion is the total operating cost. Google and Amazon are spending a lot on training their AI models and the training cost for AI models is getting more and more expensive.
The most data centers are owned by Microsoft (300), followed by 215 data centers of Amazon. Microsoft has partnered up with BlackRock to start a $100 billion plan through Global Artificial Intelligence Infrastructure that will focus on developing AI data centers and provide energy which will support those AI data centers.

Tech giants are investing billions in AI data centers to meet the growing demand for computational power.

Company Spending on GPUs and Other Chips Other AI-related Expenses Total Capital Expenditures Spending on Training and Research Spending on Inference Total Operating Costs
Microsoft $20 billion $20 billion $40 billion $3 billion $3 billion $6 billion
Meta $11 billion $12 billion $23 billion $2 billion $2 billion $4 billion
Google $14 billion $15 billion $29 billion $3 billion $1 billion $4 billion
Amazon $8 billion $8 billion $16 billion $2 billion $1 billion $3 billion

Read next:

• New Research Shows ChatGPT Uses Too Much Water to Keep its Data Centers Cool Which Can Have Harmful Effects to Environment

Mobile Addiction Soars, Americans Check Phones 205 Times Daily, Survey Reveals
by Arooj Ahmed via Digital Information World

No comments:

Post a Comment