After decades of secretive testing, with the occasional competition between a human and computer at a high-level board game, artificial intelligence (AI) entered into the mainstream in 2023 for everyone to use as a tool for text creation, photo and video editing, and add-on automation connected to popular software programs.
At the consumer-level, a lot of this is being provided for free at the moment, with premium plans for additional features and priority access. Various startups and tech companies operating these AI services are hemorrhaging money, with many spending hundreds of millions on cloud service costs each month and giddy investors footing the bill.
It is also placing a lot of strain on network operators, with more complex workloads requiring faster speeds and more bandwidth to provide adequate service. In the data center world, operators have started to offer AI-focused servers to meet the demand, outfitting them with GPUs and other high-performance computing (HPC) specifically designed for AI workloads. Everything, down to the networking protocol preference, is being debated for which is superior for AI.
Network service operators have AI at the forefront of their minds when it comes to future use, with AI-upgraded applications expected to be responsible for most network application traffic by 2025, according to Omdia’s AI network traffic forecast, 2022-30 report. This surge in network pressure will come from applications not used as extensively as text-based tools at the moment but which are about to come online, such as real-time video and high-resolution image creation. These require even more compute power and connectivity.
Network service providers may not be prepared for surge in traffic and have been warned about the need to upgrade networking infrastructure to be ready for AI. According to Omdia, the industry should be ready to move from ExaBytes to ZettaBytes (1021) to support the proliferation of AI, with 25% increases in traffic year-on-year between 2023 and 2030.
As AI matures, Omdia projects that extended reality (XR) and new media formats will be two of the main contributors to growth in traffic. Consumer application traffic is expected to outpace business, distribution of content will overtake collection, and network operators will look for ways to push AI-driven traffic from data centers to the edge.
Investment by some of the key players in the AI world could be seen as an investment into the next generation of networking and computing infrastructure. NVIDIA has seen its stock price rise by over 200% in the past year alone due to its position as a leader in the GPU market. Through this, NVIDIA has been able to invest heavily in AI startups and promote the use of networking hardware and tools for AI development. We are seeing heavy investment from most of the rest of the "Magnificent Seven," with Amazon, Microsoft, Alphabet, and Meta spending billions on home-grown solutions and on startups.
This entanglement of the tech giants and the rest of the AI ecosystem is helping the growth of the industry, as startups are able to cut deals for access to cloud services. Microsoft’s investment in OpenAI came with the added benefit of cheaper access to its cloud service platform, in return OpenAI was added to Microsoft 365. We expect to see similar deals like this by Amazon and Alphabet in the future.
However, this may come with a long list of problems for the ecosystem in the future, as the only way for a startup to truly push its technology forward is by having the backing of one of the Magnificent Seven. Without it, most independent startups would simply run out of money before their product generates enough to keep it afloat. Even OpenAI, which hit a $2 billion revenue run in December 2023, is looking to run another funding round, valuing the company at $100 billion in 2024.
Related articles:
Modernization of the data domain to include new practices, tools, and technologies is the third phase of digital transformation.
Think of Hypershield as an AI-enabled security fabric where every network port becomes a high-performance enforcement point.
With the rise of AI and its massive impact on networking, Cisco has pivoted swiftly to enable customers to utilize its converged networks to run taxing AI apps.
