Jon Peddie Research, a market analyst, said that if Intel’s finances continue to deteriorate and the accelerated computing and graphics (AXG) division does not provide competitive products and cannot turn a profit, it may become the next division to be abandoned.
Recently, the American multinational corporation released its second-quarter 2022 earnings report, which was lower than expected in both revenue and profit and estimated some losses.
To make progress in the discrete graphics card and data center industries, Intel established a GPU team in 2016. It spent roughly 3.5 billion US dollars to create products in 2017, following the discovery of Raja Koduri from AMD to reveal the discrete graphics card development plan.
Even though the investment must be repaid, figures show that since the accelerated computing and graphics business unit was founded in the first quarter of 2021, there has been a cumulative loss of US$2.1 billion.
Only two entry-level GPUs for entry-level graphics cards and a few data center requirements have been launched by Intel in the five years since the department was founded. Several applications, including artificial intelligence, deep learning, and machine learning, are currently expanding along with the growth of data centers.
The GPU market is crucial for Intel’s future growth, but the GPU program has fallen short of expectations. As a result, Intel may need to change its approach and explore using the Habana Gaudi 2 processor to handle workloads that require a more explicit parallel computing architecture.
The development of GPUs is not over just because it has stopped. For the foreseeable future, Intel will need to continue investing in costs while also maintaining current devices, updating drivers, etc. Since Intel started preparing for the GPU market six years ago, rivals Huida and AMD have updated three generations of products and created GPUs with faster performance and more complete functions.
At the same time, numerous new GPU design firms have appeared. The laws governing GPU development and design have also changed significantly, and Intel’s investment looks to be failing to effectively grow the departmental business.
Meanwhile, according to Jon Peddie Research, Intel will have to make a decision, and the GPU company may continue to lose money for several years as it attempts to enter an extremely competitive and unfavorable market.
Discussion is no longer centered on performance. Although Intel has consistently maintained and promoted the GPU brand’s popularity, the most crucial thing to accomplish right now is to introduce fully functional GPU products as soon as possible while assuring a steady supply.