Former Intel CEO Admits Mistakes in AI Betting — Company Misses Out on Multi-Billion Dollar Industry
Former CEO of Intel Pat Gelsinger admits company's AI strategy miscalculations, stating that the impact of this technology had been underestimated. According to the former head, the focus on inference (inference) as a key area turned out to be a mistake, while competitors were already actively mastering model training and creating specialized accelerators.
With the AI segment growing rapidly, Intel Left Out of the Race, offering only a limited number of solutions to the market - in particular, the Gaudi accelerator line, which was not widely adopted in data centers. Attempts to scale solutions, such as the Falcon Shores project, were eventually curtailed.
At the same time, Competitors like NVIDIA and AMD are actively increasing their share of the AI market, releasing increasingly powerful and energy-efficient solutions. Intel, however, continues to rely on classic server CPU Xeon, failing to offer a competitive AI accelerator.
Intel's new CEO, Lip-Bu Tan, began to revise the IDM 2.0 strategy, which involved combining design and in-house manufacturing. Instead, the company will focus on chip architecture and design rather than developing a foundry business, which could change the balance of power in the future.
According to analysts, Lost Time in AI Could Cost Intel Tens of Billions of Dollars, as the industry already brings in huge profits for other manufacturers. Against the backdrop of this recognition, one can expect a radical change in the company's course in the coming years.
