Right place, right time: The story behind Intel’s amazing rebound

A year ago, Intel was widely viewed as on its way to becoming a Silicon Valley relic: bleeding cash, shedding jobs and struggling to keep pace with the AI revolution upending the semiconductor industry. Today, the mood could hardly be more different.

Intel’s stock has surged almost 500 per cent over the past year as investors bet that new boss Tan Lip-Bu can engineer one of the most remarkable turnarounds in modern tech history.

But Intel’s revival says as much about the AI infrastructure boom and geopolitical panic over semiconductor supply chains as it does about the company itself. The chipmaker has made smart moves, sure, but it has also been rescued by a market suddenly desperate for more AI computing power plus US chipmaking capacity.

When the Muar-born, Singapore-raised Mr Tan took over from Mr Pat Gelsinger in March 2025, Intel was in deep trouble, slashing thousands of jobs, freezing major manufacturing projects and offloading non-core assets after its chipmaking business haemorrhaged billions of dollars. Investors were openly asking whether it should be broken up.

What changed at Intel was not just its strategy, but also the economics of the industry around it. As demand for AI hardware exploded, the boom in data centre spending suddenly created a far more forgiving environment for a company that, only months earlier, had looked dangerously exposed. 

For years, Intel relied on a “vertically integrated” model in which it both designed and manufactured its own chips. The strategy proved enormously successful in the PC era, when the company dominated the market. At its peak, Intel controlled more than 90 per cent of the global x86 processor market, giving it an extraordinary grip over the chips powering most of the world’s PCs. 

But the industry moved on, and a series of shifts combined to knock Intel from its dominant position. The first blow came in 2020, when Apple began replacing Intel processors in Macs with its own chips. It shattered the assumption that the company’s architecture would remain central to the future of computing.

The second blow came when Intel began losing ground on two critical fronts: manufacturing advanced chips and designing processors for AI systems. Taiwan’s TSMC had pulled ahead in cutting-edge chip production, while US rivals Nvidia and AMD raced ahead in the data centre chips powering AI.

Demand booms

Intel largely missed the shift because it remained too anchored to its traditional processor business even as the semiconductor industry moved towards greater specialisation. Nvidia came to dominate AI chips, while TSMC entrenched its lead in advanced manufacturing. Intel was squeezed between them. The failure was managerial as much as technological.

Intel’s rebound has so far been fuelled mainly by booming demand for its existing chip business, aggressive cost cutting and rising investor confidence in Mr Tan’s turnaround plan. The data centre boom has flooded the semiconductor industry with spending, buying the company time as it restructures and fights to reposition itself.

At the same time, Intel has begun chasing external customers for its factories as it tries to build a foundry operation capable of rivalling TSMC. The strategy has so far come at a huge cost: Intel’s foundry division burned through US$7 billion (S$8.9 billion) in 2023 alone. Under Mr Gelsinger, the company ploughed billions into expanding manufacturing capacity with little immediate payoff before the former CEO was pushed out in 2025.

Yet, the AI boom may now be rewriting the economics of the foundry business. Demand for advanced chips has become so ferocious that the market may finally be able to sustain another major manufacturing player alongside TSMC, which dominates the market for cutting-edge chip production.

And Mr Tan appears to have concluded that the tightly integrated model which powered the company for decades could no longer sustain it on its own. Intel is now courting external customers for its factories even as it pushes deeper into its own chip business.

That strategy has become far more viable because geopolitics has made semiconductor manufacturing a strategic issue. Intel still commands something few Western companies can replicate: advanced chipmaking capacity on US soil. 

As tensions between China and Taiwan have escalated, the US has become increasingly alarmed by its dependence on TSMC for many of the world’s most advanced chips. That geopolitical shift has sharply strengthened Intel’s hand.

The US has already thrown substantial weight behind that strategy through the CHIPS and Science Act, which was designed to rebuild US chipmaking and cut reliance on Asian supply chains. In 2024, Intel was earmarked for up to US$8.5 billion in direct funding under the programme, the largest proposed award to a single company.

Mr Tan has leaned aggressively into those advantages: engineering expertise and US chipmaking facilities. But the company still trails Nvidia in data centre chips and continues to face pressure from large tech groups increasingly designing their own chips, like Apple.

Still, much of Intel’s recovery remains tied to forces largely outside its control: surging investment in AI, flaring geopolitical tensions and massive state-backed spending on semiconductor infrastructure. Intel did not win the AI race; geopolitics and the AI boom revived it.

Risks Intel faces

However, three risks loom large for the company. The first is that the current AI infrastructure splurge could eventually replay the pattern of earlier tech investment cycles, from railways to telecoms, where waves of spending fuelled massive overcapacity before collapsing into sharp reversals.

Some analysts now estimate that investment in data centres could swell to as much as $9 trillion by 2030, even as fears mount that supply could eventually overwhelm demand. If the cycle turns, Intel would be badly exposed.

The second is customer concentration. Intel’s recovery is increasingly tied to spending on AI infrastructure by a small group of so-called “hyperscalers” including Microsoft, Amazon and Meta. Any pullback in data centre investment would send shockwaves through the semiconductor industry.

The third is political risk. Intel has benefited from a powerful geopolitical tailwind as the US races to rebuild domestic semiconductor manufacturing. But that support comes with strings attached. In 2025, President Donald Trump publicly called for Mr Tan to resign amid concerns over his ties to China, underscoring how exposed Intel has become to shifts in US politics.

Yet, Intel’s troubles began long before the current frenzy over AI and semiconductor supply chains; the company stayed rich long enough to miss the future. Intel remained highly profitable in PCs and servers even as the semiconductor industry’s centre of gravity shifted towards graphics processing units, or GPUs — the chips that became critical to AI systems — during the late 2010s.

In 2019, Intel generated US$21 billion in profits, about 40 per cent more than the combined earnings of TSMC, Nvidia and AMD. Within five years, the balance of power in semiconductors had been transformed. By 2024, Intel was posting a net loss of US$18.8 billion even as Nvidia’s valuation surged into the trillions. The reversal was brutal.

The case shows how easily established companies can squander their own strengths. Intel retained deep engineering expertise and advanced manufacturing capabilities but stayed wedded to the business model that had made it successful in the first place.

It remains unclear whether Intel is undergoing a genuine reinvention or simply riding the current AI boom. But the company is at least trying to reinvent itself around advanced chipmaking and US manufacturing — two areas that have suddenly become strategically important again.