Why the AI Boom Is Good for Microsoft, but Risky for Tech Hardware Makers
Text size
Victor J. Blue/Bloomberg
Corporations are shifting budget priorities to experimental artificial-intelligence projects. That may be good news for
Microsoft
,
but it could just as well turn out to be bad news for tech hardware vendors.
Earlier this month, Piper Sandler published its latest survey of 147 chief information officers, asking them about areas they intend to spend more or less on in the near future. The “results suggest that IT budgets are likely to moderate in 2023,” wrote Piper’s research team. The buyers revealed that their aggregate budget outlook for this year had declined to a 3.6% growth rate year over year, down by 1.3 percentage points compared with six months ago.
And there was another important trend buried in the numbers: Companies are prioritizing new AI applications above everything else. Generative AI rose nine spots in priority to become the top emerging technology trend for the next three years, according to the survey. The report also noted that three-quarters of CIOs are either testing or implementing AI projects already.
Microsoft (ticker: MSFT) may be the big winner from the shift to AI in underlying enterprise spending. The company can benefit in two primary ways: rising demand for its Azure cloud-computing service and new paid software features.
There is evidence that Microsoft is making progress on both fronts. Piper Sandler analyst Brent Bracelin says CIOs revealed intentions to use more of the company’s cloud services for AI. Then, in the past week, Microsoft announced a higher-than-expected pricing of $30 per user a month for its coming AI-enabled Microsoft 365 Copilot software subscriptions.
The price point was well above Oppenheimer’s estimate of $20 per user and sparked a wave of investor excitement. Following the news, Microsoft stock added $102 billion in market value in one trading session, with its shares surging 4%, to a record high this past Tuesday.
But while Microsoft basks in the excitement over AI, other vendors are facing more challenges. After looking at the CIO responses, Piper’s team predicted that hardware suppliers, such as computer server makers, would most likely face budget pressures later this year.
Taiwan Semiconductor Manufacturing (TSM) provided more evidence of an AI-driven budget-shift theme on Thursday. TSMC dominates the market for making high-end chips across tech, so it often gets an early read on market changes. According to TrendForce, the chip giant has about a 60% share of the third-party chip manufacturing business, followed by
Samsung Electronics
(005930.Korea), with 12%.
TSMC management reported an overall softening of semiconductor demand, adding that the macroeconomic environment, expectations for China’s recovery and end-market demand both looked weaker than they had expected only three months ago. With fading smartphone and PC categories accounting for a majority of TSMC’s business, the company significantly lowered its financial guidance for the year, forecasting a year-over-year decline of 10% in revenue.
In keeping with the CIO survey, TSMC also saw a significant rise in AI-related interest—though not enough to offset the overall drop in semiconductor demand.
Still, the biggest takeaway came when an analyst asked if the surge in AI orders was taking sales away from traditional server chips. “Will [AI] cannibalize data-center processors?” said TSMC Chairman Mark Liu on the second-quarter earnings call. “In the short term, when the capex [budgets] of the cloud service provider is fixed, yes, it will.”
With CIOs and the largest chip foundry in the world saying similar things, investors should be cognizant of the risks that AI may bring to makers of computer servers and server processors.
Write to Tae Kim at [email protected]
link