AI "pick and shovel sellers"! Global chip companies' sales will exceed $400 billion in 2025, and Goldman Sachs predicts Nvidia alone will reach $383 billion in revenue next year.

AI "pick and shovel sellers"! Global chip companies' sales will exceed $400 billion in 2025, and Goldman Sachs predicts Nvidia alone will reach $383 billion in revenue next year.

Driven by explosive growth in artificial intelligence, the combined global sales of chip companies in 2025 will surpass $400 billion, setting a historic record for the industry, with that figure expected to climb even further in 2026.

According to Goldman Sachs estimates, Nvidia alone will reach $383 billion in GPU and other hardware sales in 2026, a 78% increase from the previous year.

Analysts surveyed by FactSet expect the combined sales of Nvidia, Intel, Broadcom, AMD, and Qualcomm will exceed $538 billion next year, not including Google’s TPU business or Amazon’s custom chip sales.

Last week Nvidia signed a $20 billion licensing deal with chip startup Groq, which specializes in AI inference acceleration. This marks a shift in the AI race from the training stage to the inference stage, as tech giants now compete to offer the fastest and most cost-effective inference capabilities.

These hardware designers are playing the role of “shovel sellers” in this digital gold rush, but industry growth faces major constraints, including shortages of data center components such as transformers and gas turbines, electricity supply issues, and uncertainty about whether AI companies can continuously secure enough funding to sustain the pace of chip purchases.

Intensifying Competition

Nvidia achieved more than double year-on-year revenue growth in 2025, but the competitive landscape is changing.

Data center operators, AI labs, and enterprise customers have strong demand for Nvidia’s advanced H200 and B200 GPUs, but Google’s increasingly sophisticated custom TPU chips and Amazon’s Trainium and Inferentia chips are also vying for customers.

Nvidia’s $20 billion licensing agreement with Groq last week reflects the industry’s shift from AI training to inference. Inference refers to the process where a trained AI model responds to prompts and provides answers.

Software developers like OpenAI are partnering with custom designers like Broadcom to develop their own chips. AMD, a chip manufacturer with a half-century history, plans to launch its first GPU to truly challenge Nvidia’s AI processors in 2026.

In October, Microsoft announced it would double its data center footprint within two years, suggesting chipmakers could earn even more revenue in 2026.

Supply Chain Bottlenecks Emerging

2026 could bring unprecedented challenges. Shortages of components like transformers and gas turbines are hindering data center construction, and operators are struggling to acquire the massive amounts of electricity needed to run computing clusters.

Another major challenge is the global shortage of components needed for AI data center servers.

Products in tight supply include ultrathin silicon substrate layers required by certain chips, as well as memory chips that deliver data to AI processors and help store computed results.

As data center construction accelerates and inference demand rises, demand for high-bandwidth memory chips is surging. Micron Technology’s Chief Business Officer Sumit Sadana said:

We are far from meeting customer demand, and this situation will persist for some time.

Micron is one of the largest manufacturers of high-bandwidth memory chips used in AI, with its share price up 229% so far this year.

Micron and competitors such as Samsung and SK Hynix are major beneficiaries of the supply squeeze, enabling them to raise prices and increase capital spending to expand production. But building the large-scale cleanrooms and factories needed to meet major chip company demand takes time.

Sustainability of Financing in Doubt

There are serious questions about the sustainability of financing behind data center construction; it’s unclear whether major clients like OpenAI can quickly raise enough funds to keep up rapid chip purchases.

Investors have grown accustomed to extraordinary quarterly revenue growth, and any signs of slowdown can easily trigger panic. This autumn, investors sold off AI stocks, including major chip designers, fearing that the financing supporting purchases of AI infrastructure products may not be as robust as previously thought.

A large portion of data center construction is driven by OpenAI, which has signed multi-billion dollar compute agreements with Amazon, Microsoft, Oracle, and others. Tech giants like Microsoft have committed to expanding data centers in 2026, but some analysts believe the boom may slow in 2027.

DA Davidson analyst Gil Luria said:

2026 could be the peak. If by the end of March we haven’t heard that OpenAI raised $100 billion, the market may start to hit the brakes.

Concerns are also mounting that as more chip companies launch AI products, profit margins will come under pressure.

After Broadcom announced record quarterly revenue in December, its share price still fell, partly because investors worried that sales growth for its high-margin product lines would slow. However, some in the industry take a more optimistic view, believing demand will remain strong.

Brad Gastwirth, Global Research Director at Massachusetts hardware distributor Circular Technologies, said:

I don’t think this is the peak. The race for general artificial intelligence continues to drive massive demand for computing power across all types of customers.

Risk Warning and DisclaimerThe market involves risk; investments should be made cautiously. This article does not constitute personal investment advice and does not take into account individual users’ specific investment goals, financial status, or needs. Users should consider whether any opinions, views, or conclusions in this article suit their particular circumstances. Investing based on this is at your own risk.