TLDR
- Nvidia stock dropped more than 4% Thursday despite Big Tech signaling massive AI spending increases
- Google announced plans to sell its custom TPU chips to outside customers, raising competition concerns
- Amazon flagged its in-house chip business is growing fast
- The four major hyperscalers combined now plan to spend up to $725 billion on AI infrastructure in 2026
- Nvidia’s B300 servers in China have nearly doubled in price to ~$1 million due to chip smuggling crackdowns
Nvidia stock fell more than 4% on Thursday, even as the biggest names in tech announced eye-watering increases to their AI spending plans. The drop reflects a growing concern among investors: what happens to Nvidia’s dominance when its best customers start building their own chips?
The sell-off came the morning after Meta, Alphabet, Microsoft, and Amazon all reported earnings. Every one of them raised their capital expenditure guidance for 2026. Meta bumped its outlook by $10 billion, now targeting between $125 billion and $145 billion. Alphabet raised its guidance by $5 billion to as much as $190 billion. Microsoft said its Q4 capex alone would top $40 billion.
Add it all up, and the four hyperscalers are now expected to spend as much as $725 billion on AI infrastructure this year. Nvidia captures an estimated 90% of AI accelerator demand, so on paper, that should be great news for the chipmaker.
But markets don’t always move on good news alone.
Google’s TPU Move Spooks Investors
What rattled investors was a specific announcement from Alphabet. The company said it would begin selling its custom Tensor Processing Units — TPUs — directly to select outside customers who can install them in their own data centers.
Until now, TPUs were primarily used internally by Google. Offering them externally positions them as a direct, if limited, alternative to Nvidia GPUs. TPUs are generally seen as less flexible than Nvidia’s chips, but they can be more cost-effective for certain AI workloads.
Amazon also highlighted growth in its own in-house chip operations during its earnings call. Both companies are still major Nvidia customers, but the direction of travel is clear.
Nvidia has previously brushed off concerns about custom chip competition, arguing its GPUs offer broader flexibility for AI developers. That argument is getting harder to make without some skepticism from the market.
China Supply Crunch Pushes B300 Prices Toward $1 Million
On the supply side, prices for Nvidia’s advanced B300 servers in China have surged to roughly 7 million yuan, up from around 4 million yuan late last year. That’s nearly $1 million per unit.
A crackdown on chip smuggling in China — which had previously kept a grey market for restricted hardware going — has tightened supply considerably. The B300 is one of Nvidia’s most capable AI server systems, and it remains restricted in China under US export controls.
Meanwhile, Thursday was a mixed day for the broader chip sector. Qualcomm jumped 9% as it pushed further into data center markets. Memory players Sandisk, Western Digital, and Seagate also climbed, after Microsoft and Meta flagged rising costs for storage and memory components.
On the investment side, Nvidia’s venture arm NVentures participated in a $50 million extension of Swedish AI legal tech firm Legora’s Series D round, which valued the company at $5.6 billion and brought total funding to $600 million.
Nvidia stock was trading around $200.84 as of Thursday afternoon, down roughly $8.41 on the day.
🚨 Our April Stock Picks Are Live!
A new month means new opportunities. Our analysts have just released their top stock picks for April, highlighting companies with strong momentum that rank highly on our KO Score algorithm. We’re also now sharing trade ideas for both long-term and short-term investors, giving you more ways to spot potential opportunities in the market.
Sign up to Knockout Stocks today and get 50% off to unlock the full list and see which stocks made the cut.
Use coupon code Special50 for your exclusive discount!
