Meta platforms today
- 52-week range
- $ 414.50
▼
$ 740.91
- Dividend profitability
- 0.34%
- P/e ratio.
- 25.90
- Value is valuable
- $ 718.31
New and exciting report on magnificent seven drains, metabolism NASDAQ: metRecently came out. Thomson Reuters NYSE: TRRI He said that Meta is testing its own semiconductors for teaching artificial intelligence models.
This is a direct attempt to reduce dependence on the manufacturer of graphic processing (GPU) NVIDIA NASDAQ: NVDAField
So, what are META internal chips and why is the company taking this step?
How can this help the company limit the costs of AI, and can this development potentially benefit the action in the long run?
Meta-Fish: what are they and why Meta repels nvidia
Currently, Meta checks its first internal chip specifically for teaching artificial intelligence. Teaching AI refers to the usually one -time and initial cost of teaching the AI model, how to think and make forecasts. META, reportedly, has already used the internal chips for the output. The conclusion refers to the process of the AI model actually answering a unique question or circumstances. This is what happens when someone sets, for example, ChatGPT. Companies should train models before they can conclude.
Both of these tasks require extensive costs for computing power and energy, but occur at different times. Training requires much more initial costs for calculation and energy. As soon as the company teaches the model, the costs of each separate conclusion are low. However, since the conclusion continues, these small costs make up, which potentially creates large costs over time.
In general, Meta looks at both ends of the spectrum, trying to reduce costs both in training and in conclusion. Reducing costs in each is key. Meta wants to create the best models over time through training. They also want billions of users to interact with their models through the conclusion, which also makes their key.
The tendency to do this, creating your own chips, makes sense, taking into account the huge cost of NVIDIA graphic processors. NVIDIA gross profit amounted to almost 74% in the last quarter, which shows its huge price power. NVIDIA chips are mainly used for training and output. The construction of internal chips implicitly introduces more competition for NVIDIA. This means that companies such as META should not accept and pay any price that NVIDIA simply charges.
Energy plays a huge role in promoting the inner chip Meta
User chips obtained from AI, such as what META began to use, can also offer both higher performance and lower energy consumption compared to graphic processors. This is true for specialized output tasks, although graphic processors are still conducted as a result of the general purpose of Genai. Amazon NASDAQ: Amzn According to its user chip, Tranium2, offers 30% to 40% the best performance price than the NVIDIA H100 graphic processors.
The MIT Sloan Management School expects to increase energy consumption from data processing centers. Data processing centers are currently from 1% to 2% of global energy demand. MIT says that the number can increase to 21% by 2030, given the costs of AI. Thus, a decrease in the use of the energy of its chips is very important for meta. This can reduce the costs that he must pay for energy agreements to power his mass data centers.
The meta is at different stages, when it comes to using its home chips, depending on the task. Currently, META uses its SHIP, known as Artemis, in scale to provide recommendations. This includes the provision of advertisements and short videos, such as coils and recommendations of other content on Facebook and Instagram. However, Meta has not yet accepted Artemis for the purposes of Genai.
According to Reuters, the company is just starting to test its educational chip. If everything goes well, the company hopes to start using a chip for training by 2026. The company will first consider how it can use a chip to teach recommendations, and then, as it will do for GEN AI.
The cost of META chips can help reduce the cost, it is possible to benefit shares in the long run
Meta Platforms, Inc. (META) Monday price card, March 24, 2025
The West Meta to create its own chips can be an important shift wind for reserve. The prospect of significantly reduced costs can be added to the long term of the company. Nevertheless, interested parties will have to wait and see if the tests of his training chip will be successful. In addition, the increase in the production of these chips will have significant costs related to it, possibly placing an almost average broadcast on the fields.
Further events relating to the stage of testing the training chip and the future implementation of both chips will be the key to viewing.
Before considering a meta platform, you will want to hear it.
Marketbeat monitors the highest and most effective analysts with the most effective Wall Street analysts and promotions that they recommend to their customers daily. Marketbeat has identified five shares that leading analysts quietly whisper to their customers to buy now before the wider market wins … and the meta platforms were not on the list.
While the meta-platforms currently have a moderate purchase rating among analysts, analysts with the highest rating believe that these five promotions are better buying.
View five shares here
Trade of options not only for Wall Street Elite; This is an affordable strategy for those who are armed with proper knowledge. Think of options as strategic tools, with each tool intended for a specific financial task. Get this report to find out how trading in the options can help you use market volatility in your interests.
Get this free report