The battle to develop next generation artificial intelligence chips is now luring in some of the biggest technology companies in the world.
What’s happening:
- Meta (NASDAQ: META) has announced they are developing their own chips for artificial intelligence computing power in order to reduce their dependence on Nvidia (NASDAQ: NVDA) GPUs
- Meta’s new custom designed artificial intelligence processors are known as Artemis
The big idea:
- Meta is planning to deploy their own purpose built chips across their data centres in an effort to reduce their dependence on other hardware providers and lower the costs of computing power for training their generative artificial intelligence workloads and large language models
- Because Meta is building their own chips that are custom designed to their own artificial intelligence needs, they can reduce the amount of total energy that would otherwise be required as a result of training their artificial intelligence workloads on other hardware
Why it matters:
- Meta is one of the latest large technology companies to venture into developing their own chips and silicon semiconductors in an effort to capture the boom in artificial intelligence workloads and computing capabilities
Who is making moves:
- Hailo recently raised a new large round of financing to advance their chips capable of running artificial intelligence workloads on devices without any connection to the Internet
- Microsoft, Google and Amazon Web Services are also all developing their own custom processors and artificial intelligence chips to capture the enormous opportunity in providing cloud computing to venture backed artificial intelligence companies
Going deeper:
- Meta has also just launched their Llama 3 next generation, open source large language model
- Many believe this is the single most advanced open source large language model in existence and it further deepens Meta’s position as a leader in developing technology for developers of new artificial intelligence applications
