Nvidia and Intel: Old Rivals Build a New Chip While Musk Eyes China Warily

The technology industry is entering a phase in which yesterday's sworn enemies are sharing a bed, and market leaders openly acknowledge the threat from Eastern competitors. Jensen Huang has officially confirmed that Nvidia is working with Intel on a custom processor based on the X86 architecture. It is fair to call this a tectonic shift.

For a long time Nvidia relied on the ARM architecture or its own designs, but it seems that, for deep integration into data-center ecosystems and specific enterprise workloads, the capacity of Intel and the decades-proven X86 architecture is still not enough to be dispensed with. The partnership looks like an attempt by Nvidia to cement its dominance in server solutions by drawing on Intel's manufacturing know-how, which Intel now badly needs in order to survive.

While chipmakers negotiate alliances, Elon Musk delivered a dose of sobering analysis on Tesla's quarterly earnings call. According to him, the humanoid robot field has no serious players other than Tesla and Chinese companies. Musk stressed that China's manufacturing base has reached incredible heights, and that this is where he sees the main competitors for his Optimus.

Interestingly, the role of "brain" in this ecosystem goes to Grok from xAI. Musk calls it a "conductor" that will run not only the robots but the entire fleet of self-driving cars. This is an important signal: Tesla is finally turning from an automaker into a company where software and AI agents determine the value of the hardware.

Without advanced chips, as Musk noted, any robot is just an expensive and useless pile of metal.

However, the real world makes its own adjustments to ambitious plans. Efirt, a notable player in industrial robotics, announced price increases of 5–12% on its products. The reason reads like a dispatch from the front: copper is up 38%, aluminum 25%, and memory chips have soared a staggering 300%. This is a cold shower for those used to discussing AI only in terms of cloud computing and algorithms. Physical embodiment of intelligence requires resources, and those resources are becoming scarce and expensive. Inflation in high-tech hardware may become the main brake on robotization over the next two years.

Against this backdrop, ByteDance continues its expansion into consumer hardware. The Doubao smartphone project, whose first version was more of a trial balloon, is moving into a full-fledged stage. The second iteration of the device, developed jointly with ZTE, is expected in mid-2026. ByteDance is betting that an AI assistant should be integrated at the core of the system rather than exist as a standalone app. This is a direct attack on Apple's and Google's territory. If the company can create a seamless AI experience in the user's pocket, the rules of the smartphone market will change forever.

Alibaba is keeping pace, having decided to consolidate its efforts in autonomous transport. The merger of its autonomous driving unit with the company Zelos suggests that the era of scattered experiments is over. The giants want to pool resources to get commercial driverless vehicles onto roads faster. At the same time, Xpeng is patenting new mechanisms for robot manipulators, trying to reduce the risk of damage to the drives and make operation cheaper. The industry is clearly shifting its focus from prototypes that merely "work" to cost-effective, reliable machines.

The takeaway: Nvidia's readiness to work with Intel and Musk's acknowledgment of China's strength show that the old templates no longer work. We are heading into an era of hybrid technologies and fierce competition for physical resources. Can Western software beat Chinese manufacturing in the robot race?