Connect with us

Bussiness

After 25 years, Intel’s Dow days look numbered. Nvidia is the top dog in AI race

Published

on

After 25 years, Intel’s Dow days look numbered. Nvidia is the top dog in AI race

In November, 1999 the selection committee for the Dow Jones Industrial Average caved in and allowed two high-fliers on New York’s tech-focused Nasdaq market, Intel and Microsoft, into the grandfather of stock market indices.

Size doesn’t overly matter for this five-member team, comprised of two representatives from the Wall Street Journal and three from S&P Dow Jones Indices.

Acceptance into the Dow club of just 30 stocks is more down to vague criteria like a company having an “excellent reputation”; demonstrated an ability for “sustained growth”; and whether it is “of interest to a large number of investors”.

On all counts, it was arguable at the time that the Dow compilers were already late in adding the two and booting out “smokestack” stocks like oil group Chevron and Goodyear Tire & Rubber Company.

The Wintel partnership – using Microsoft’s Windows software and Intel’s x86 computer processors – had dominated the personal computers (PC) market during a period of large growth through the 1990s. Intel had soared close to 3,500 per cent over the decade. Microsoft had jumped at twice the pace, while the Dow had only risen 275 per cent.

Unfortunately for the index committee, the tech stock inclusions took place just months before the dotcom bubble burst and tech stocks plummeted. And twenty-five years later, it finds itself, once again, well behind the curve.

While Microsoft has grown to become the most valuable public company in the world – overtaking Apple this year as its market capitalisation breached $3 trillion (€2.77 trillion) – Intel’s shares have lost more than half their value in the past three years as its name has long become a byword for old tech.

It’s surely only a matter of time before it is replaced by Nvidia, the new darling of Wall Street as the global leader in artificial intelligence (AI) computing (the Dow may not be the significant benchmark it once was. The S&P 500 is now the most widely followed index in the US, and indeed, globally. But inclusion in it still carries symbolic weight).

Nvidia is now valued at $2.72 trillion. The $220 billion or so that was added to Nvidia’s market cap in just one day last week, when it delivered a blowout quarterly earnings, dwarfed Intel’s entire $130 billion value.

Intel may be the largest maker of chips known as central processing units (CPUs) in the US, but its focus remains on the PC and laptop industry, which has been in decline for years.

It conceded defeat eight years ago in is efforts to infiltrate the smartphone market with its own chips. Still, it has become a sizeable manufacturer of phone chips for other semiconductor companies, including an outsourcing deal reached last autumn with Arm, a large supplier of intellectual property to many chip companies, especially in mobile phones.

In the data-hungry space of AI, Intel is struggling to play catch up with the wider industry.

Conor Pope’s guide to getting the most bang for your summer holiday buck

Here, it is dealing with the ignominy of its Santa Clara, California neighbour, Advanced Micro Devices (AMD) – for decades its fiercest, if distant competitor, in the PC chips market – becoming the main threat so far to Nvidia in designing and selling graphic processing units (GPUs) used for machine learning and generative AI like ChatGPT.

It didn’t have to be this way. The company started working on a project back in 2008 to develop general-purpose GPUs. It ended up being axed more than a decade ago – ultimately putting it on the back foot in the current AI arms race.

Intel acquired an AI accelerator chip programme, called Gaudi, five years ago with its purchase of Habana Labs of Israel. It unveiled a third-generation chip, Gaudi3, last month and expects to generate as much as $500 million of Gaudi revenue this year.

By contrast, tech market research firm Omdia expects Nvidia’s AI revenue to soar to $87 billion this year from $34 billion in 2023. The big spend on AI chips for computer systems and networks has eaten into corporate budgets for Intel’s own data centre chips business.

The AI boom in the past 18 months masked a wider downturn in semiconductors in 2023, following something of a dream run during the peak of the pandemic when there was an unprecedented surge in demand for laptops and tech gadgets.

The wider sector has rebounded this year. Intel – one of the largest foreign multinational employers in the Republic, with about 6,500 working in its chip wafer fabrication facilities in Leixlip, Co Kildare – saw its revenue climb 9 per cent in the first quarter due to a recovery in the PC market.

Chief executive Pat Gelsinger – an electrical engineer by background, who returned to run the company three years ago, having previously worked for it for 30 years until 2009 – has been trying to turn around Intel’s fortunes in recent years by becoming a much bigger player in the contract manufacturing market for other companies’ chips.

It has split its business in two this year: a chip development, or products, unit; and a manufacturing, or foundry, business. But in doing so, it revealed that the foundry division made a $7 billion loss last year, up from about $5 billion for each of the two previous years. Intel’s forecast that this business will only break even in 2027 has disappointed investors.

Analyst at Deutsche Bank said in a recent report that Gelsinger’s focus on improvement manufacturing, costs and winning foundry customers is correct. “However, the financial benefits from the company’s necessary but costly transformation efforts remain many years into the future,” they said.

Hardly likely to save it from being booted out of the Dow in the meantime.

Continue Reading