For decades, Intel stood as the undisputed leader of the semiconductor industry, driving the digital revolution with its groundbreaking microprocessors and commanding dominance in chip manufacturing. The company’s legendary ‘Intel Inside’ campaign not only highlighted its technological prowess but also established Intel as a household name, synonymous with innovation and trust. However, in recent years, Intel’s once-mighty legacy has been shaken by relentless competition, production setbacks, and a rapidly shifting tech landscape driven by AI advancements. In a dramatic turn of events, just months after CEO Pat Gelsinger unveiled an ambitious restructuring plan in September, Intel announced his retirement and he stepped down from the board of directors, effective December 1, 2024.

Gelsinger’s appointment as CEO in 2021 was heralded as a potential turning point for Intel — a return to its engineering roots and a bold move to reclaim its industry leadership. It was the time when Intel’s annual revenue was inching to $80 billion, and the Intel board was positive about Gelsinger's appointment to ensure strong execution of Intel’s strategy to build on its product leadership and take advantage of the significant opportunities ahead as it continued to transform from a CPU to a multi-architecture XPU company.

However, things didn’t go as expected. The company’s revenue declined steadily to $54 billion after 4 years, including its core competency areas such as the x86 CPU segment. Also, in a shocking turn of events, in November, NVIDIA replaced Intel in the Dow Jones Industrial Average (DJIA).

“When Pat Gelsinger had taken over as Intel’s CEO, its market cap was about 30-35% lower than NVIDIA. However, presently NVIDIA’s Market Cap is about 35 times that of Intel and Intel’s stock price has fallen about 60-70% from its 2021 high. Recently, NVIDIA replaced Intel in the Dow Jones Industrial Average, which is a further testament to the fact that Intel is no longer seen as providing a representative exposure to the semiconductor industry. Intel’s revenue was around $79 billion in 2021, which had fallen to about $54 billion in 2023. So definitely Intel is facing severe headwinds over the past few years,” explains Devroop Dhar, Co-founder and MD at Primus Partners.

Big promises, bigger misses

The 30-year Intel veteran and the company’s first chief technology officer, Pat Gelsinger was brought back to reclaim the edge Intel had begun to lose in semiconductor manufacturing to its overseas rivals. Once the undisputed leader in the industry, Intel’s market dominance was eclipsed in 2021 when Samsung seized the crown, driven by its strong growth across both logic IC and memory segments, soon after Gelsinger took over.

Gelsinger attempted to transform Intel into a chip making powerhouse, which would take on the Taiwanese and Korean giants in this space. Under his leadership, Intel looked at investing about $100 billion to ramp up manufacturing of chips in the USA and to compete with the likes of Samsung and TSMC. He even considered using an internal Intel foundry for chip manufacturing with the expectation that the foundry would serve Intel as well as other customers.

Gelsinger did a great job selling a $100 billion leading-edge expansion plan in the EU for EU CHIPS Act subsidies and another $100 billion expansion plan in the US on the leading edge for US CHIPS Act subsidies. Intel was the poster boy for the US Government’s CHIPS Act. The government had also put its weight behind Intel as the US Department of Commerce awarded Intel $7.86 billion in direct funding to ramp up manufacturing and advanced packaging across various states such as Arizona, New Mexico, Ohio and Oregon.

“Gelsinger succeeded in getting the highest subsidies approved both in EU CHIPS Act and US CHIPS Act, however, both EU and US expansion projects are far away from being on time and rollbacks are already happening in the US,” says Danish Faruqui, CEO of Fab Economics – a US-based semiconductor Fab/OSAT Greenfield projects advisory and implementation consultancy. “While the idea was to make Intel self-sufficient in the complete life cycle of chip manufacturing, however, the Intel Foundry unit posted significant losses which are only expected to further mount in days to come, proving that the attempt has not worked out well for the firm,” adds Dhar.

Gelsinger also did an outstanding job in the roadmap setting for regaining the technology prowess of Intel by 5 nodes [Intel 7, 4, 3, 20 A, 18A] 4-year plan - which everyone bought as a turnaround plan from the Board to Investors. “However, a critical caveat in the 5N 4Y plan was the hidden rebranding of Intel nodes from 10nm to Intel 7, 7nm to Intel 4 to calibrate with equivalent transistor densities at rivals TSMC and Samsung,” explains Faruqui. The second half of 2023 was the show time for which everybody who bought Gelsinger's turnaround plan of 5N 4Y was waiting. It was the time when Intel’s most awaited first EUV node or Intel 4 High Volume Manufacturing (HVM) milestone was planned - closely followed by Intel 3. “However, Intel 4 and 3 did not achieve real HVM, neither at the yield level nor at the manufacturing capacity level, forcing Intel itself to outsource to TSMC 3nm for its leadership products,” he adds.

In parallel, there was immense branding and wait for Intel 20A and 1.8A, which would include both RibbonFET or Gate All Around architecture as front-end radical innovation and backside power delivery network (BPDN) radical innovation at the backend. Many tier-1 customers like Qualcomm, Amazon, and Broadcom were shown to be queuing up as with the Intel 1.8A paper plan. If this would have worked, Intel could have gotten ahead of TSMC. “However, recent reports on around 20% yielding Intel 1.8A has plummeted the confidence of the board, investors and customers alike. In summary, it was a great branding of a 4 four-year turnaround plan but after 4 years the execution outcome was dismal,” adds Faruqui.

Intel also fell behind competitors like NVIDIA in the AI chip race as it continued to focus on its core business of making processors for PCs and servers and not investing enough in R&D for AI-powered technologies. “Presently, NVIDIA has more than 90% market share for data centre GPUs and more than 80% market share for AI processors. Intel is attempting to give competition to NVIDIA in AI chips by positioning its GPU as an alternative to NVIDIA and has priced its GPU lower than that of NVIDIA,” adds Dhar. However, NVIDIA has been transforming itself into an integrated player with a comprehensive stack of chips, systems, software, and services for data centres, cloud, and edge computing, which has made it difficult for Intel to catch up.

Pat Gelsinger

Intel’s revival under Pat Gelsinger: Bold promises, unmet expectations