The Morning Download: Google Posts Up Against Nvidia

What's up: Meta in talks to use Google’s AI chips; HP to cut up to 10% of workforce as part of AI push; Walmart tests ads in its new AI agent.

Good morning. The AI Olympics are getting more interesting. Watching Nvidia for the last year has been like watching the Dream Team in a shootaround, fun but not a game. Now Google is on the court, putting pressure on Nvidia, and that’s a good thing for enterprise technology.

Nvidia has more than 90% share in the market for GPUs, which at this point are actually large and complex systems with many components. Nvidia’s margins exceed 70%, and the business earlier this month reported record sales and strong guidance to investors.

Just last month Nvidia became the first company to crack the $5 trillion market cap. It has been in the $4.4 trillion range in recent days and it is still the world’s most valuable company.

At this point, its success is its greatest weakness, given that it presents such a large and inviting target. Alphabet’s Google is the ideal competitor, given its deep financial resources, powerful research muscle and an expansive business model capable of supporting ambitious, long-term undertakings. With cultures that fuse science and commerce, they are more similar than dissimilar.

Good for you. All of this is good news for the enterprise technology market, which stands to benefit from a rivalry that lowers the initial price and total cost of ownership in AI infrastructure.

Nvidia’s and Google’s chips don’t overlap in the way that Ford and General Motors or Coke and Pepsi do. Nvidia’s general-purpose, run-anywhere GPUs are more expensive than Google’s TPUs, which are targeted toward specific jobs or customers.  The companies may argue about which solution offers the best total cost of ownership, but that’s bound to be case specific. In the end, the market is moving toward more choice in more situations, and that is a good thing. More on Google's growing TPU business below. 

The competition also reflects a shared conviction that the demand for AI chips is growing and worth fighting over.

Down the road, rising competition may help shield all players from the scrutiny of regulators concerned about the market’s highly concentrated structure.

Attention readers: The Morning Download will be taking a break for Thanksgiving. We will be back in your inbox on Monday.

Look out Nvidia? Meta Platforms is in talks to use chips made by Google in its artificial-intelligence efforts.

Google has been working for years to refine its chips, known as tensor processing units, or TPUs, and scale up that part of its business. The WSJ reports that a Meta deal could be worth billions of dollars, but the talks are continuing and may not result in one.

The Information reported on the talks between Google and Meta on Monday night.

Google said its cloud is experiencing “accelerating demand” for both its custom TPUs and Nvidia GPUs and that the company is “committed to supporting both, as we have for years.” Last month, Anthropic announced that starting next year it would spend tens of billions of dollars to buy up to one million Google TPUs.

Nvidia by far remains the dominant player in the space. Still, a significant Google-Meta deal would represent a potential crack in Nvidia’s market dominance, says the WSJ. Nvidia’s shares fell 2.6% Tuesday.

Nvidia on Tuesday morning posted a statement on X: “We’re delighted by Google’s success—they’ve made great advances in AI and we continue to supply to Google. NVIDIA is a generation ahead of the industry—it’s the only platform that runs every AI model and does it everywhere computing is done.”

Turning heads. Nvidia is the second AI juggernaut to comment recently on Google’s AI endeavors, once written off as playing catch-up. The Information reports that in the lead-up to the release of Google’s new Gemini AI model, OpenAI CEO Sam Altman warned colleagues of rough vibes ahead.

Already? Advertising could be coming to AI agents.

IT stalwarts turned in solid quarters driven by demand for AI-enabled software, underscoring how enterprise adoption of AI is beginning to translate into real revenue. But while AI is lifting some incumbents, it is also forcing others to overhaul their cost structures to keep pace.

Workday logged higher profit in its latest quarter as sales climbed, boosted by demand for its AI-driven products, said Chief Executive Carl Eschenbach. The human-resources software company posted a profit of $252 million, compared with $193 million a year earlier. For the current quarter, the company guided for subscription revenue of $2.36 billion, representing growth of 16%. 

Dell expects AI server shipments to more than double during the full year, hitting $25 billion, the WSJ reports. The company posted a profit of $1.55 billion in the third quarter, compared with $1.17 billion a year earlier. Dell said it expects full-year revenue to be $111.2 billion to $112.2 billion, up from a prior range of $105 billion to $109 billion.

HP plans to cut up to 10% of its workforce, some 4,000 to 6,000 employees, the WSJ reports. Chief Executive Enrique Lores said the restructuring plan is part of a push to adopt AI across the company to accelerate product and software development as well as automate customer support and some internal processes. 

The move also comes as HP faces higher costs in its personal computers business as a surge in AI demand from data centers drives up the price of memory chips.

McKinsey also cut about 200 global tech jobs in the past week. A representative in an emailed statement to Bloomberg links the cuts to AI: “We are continually working to make our professional support functions more efficient and effective, including by taking advantage of AI.”

A T-Day dinner discussion idea: Who will replace Tim Cook?