Nvidia still owns the AI chip race, but AMD and Intel are making Wall Street reconsider

Nvidia still owns the AI chip race, but AMD and Intel are making Wall Street reconsider Nvidia commands up to 81% of AI accelerator revenue, yet its stock barely moved while AMD and Intel surged triple digits in the first half of 2026 Share Add us on Google by Editorial Team Jul. 25, 2026 Nvidia’s grip on the AI chip market remains firm, with the company holding between 75% and 81% of AI accelerator revenue share in the first half of 2026. Its data center sales are projected to exceed $150 billion for the year.

Nvidia stock gained roughly 15% in H1 2026. AMD? Up about 114%.

Intel? A face-melting 200% to 220%. The numbers behind the AI chip pecking order AMD sits at roughly 5% to 7% of AI accelerator revenue, translating to somewhere in the $7 billion to $15 billion range.

Intel trails badly at around 1%. Advertisement AMD locked down a significant partnership with OpenAI for its Instinct MI450 chips, with deliveries expected to start in late 2026. Intel is readying its Crescent Island AI GPU for sampling in the second half of 2026.

The chip is designed for high memory efficiency, offering up to 480GB. Why Wall Street is rotating away from Nvidia Hyperscalers like Google, Amazon, and others have been building custom silicon, think Google’s TPU and AWS’s Trainium, that collectively captures a growing share of the AI compute market. These custom ASIC solutions are outpacing the growth of traditional chip competitors, which means AMD and Intel aren’t just fighting Nvidia.

They’re fighting their own customers. What this means for crypto and decentralized compute A more competitive chip market, where AMD and Intel offer viable alternatives, could meaningfully reduce costs for node operators on networks like Render, Akash, and io.net.

AMD’s OpenAI deal is particularly interesting through this lens. If the MI450 proves competitive with Nvidia’s offerings for inference workloads, crypto-native compute networks could adopt AMD hardware at scale, reducing the single-vendor dependency that currently makes decentralized GPU markets essentially a Nvidia proxy. Intel’s Crescent Island, with its 480GB memory capacity, could also matter for large language model inference, where memory bandwidth often constrains performance more than raw compute.

If Google, Amazon, and Microsoft keep building proprietary chips that stay in-house, the total addressable market for GPU-based compute networks shrinks. Custom ASICs locked inside walled gardens are the opposite of commoditized, widely available hardware. Disclosure: This article was edited by Editorial Team.

For more information on how we create and review content, see our Editorial Policy. AI Nvidia still owns the AI chip race, but AMD and Intel are making Wall Street reconsider Nvidia commands up to 81% of AI accelerator revenue, yet its stock barely moved while AMD and Intel surged triple digits in the first half of 2026 by Editorial Team Jul. 25, 2026 Share Add us on Google Nvidia’s grip on the AI chip market remains firm, with the company holding between 75% and 81% of AI accelerator revenue share in the first half of 2026.

Its data center sales are projected to exceed $150 billion for the year. Nvidia stock gained roughly 15% in H1 2026. AMD?

Up about 114%. Intel? A face-melting 200% to 220%.

The numbers behind the AI chip pecking order AMD sits at roughly 5% to 7% of AI accelerator revenue, translating to somewhere in the $7 billion to $15 billion range. Intel trails badly at around 1%. Advertisement AMD locked down a significant partnership with OpenAI for its Instinct MI450 chips, with deliveries expected to start in late 2026.

Intel is readying its Crescent Island AI GPU for sampling in the second half of 2026. The chip is designed for high memory efficiency, offering up to 480GB. Why Wall Street is rotating away from Nvidia Hyperscalers like Google, Amazon, and others have been building custom silicon, think Google’s TPU and AWS’s Trainium, that collectively captures a growing share of the AI compute market.

These custom ASIC solutions are outpacing the growth of traditional chip competitors, which means AMD and Intel aren’t just fighting Nvidia. They’re fighting their own customers. What this means for crypto and decentralized compute A more competitive chip market, where AMD and Intel offer viable alternatives, could meaningfully reduce costs for node operators on networks like Render, Akash, and io.

net. AMD’s OpenAI deal is particularly interesting through this lens. If the MI450 proves competitive with Nvidia’s offerings for inference workloads, crypto-native compute networks could adopt AMD hardware at scale, reducing the single-vendor dependency that currently makes decentralized GPU markets essentially a Nvidia proxy.

Intel’s Crescent Island, with its 480GB memory capacity, could also matter for large language model inference, where memory bandwidth often constrains performance more than raw compute. If Google, Amazon, and Microsoft keep building proprietary chip