Intel's CEO scraps forecast of selling $500M+ worth of Gaudi AI chips in 2024, blaming slower uptake on Gaudi 2 to Gaudi 3 transition and software ease of use
Intel's $500 million AI goal unachievable as competitors rake in billions Larry Dignan / Constellation Research : Intel defends Gaudi 3 as it misses 2024 sales targets X: @constellationr : .@Intel defends Gaudi 3 as it misses 2024 sales targets https://www.constellationr.com/ ... Intel CEO Pat Gelsinger argued that its Gaudi 3 AI accelerator will be able to deliver strong total cost of ownership even as it falls short of the $500 million revenue target for 2024. @ldignan @technology : Intel CEO Pat Gelsinger tells @EdLudlow the company plans to continue investing in their AI platform “Gaudi,” despite it missing its sales target this year https://www.bloomberg.com/... [video] LinkedIn: Frederick Chen : At their most vulnerable, Intel and Samsung foundries cut back on 7nm deployment. — $3.1 billion of charges related to …
Context & Ripple Effects
Intel introduced Gaudi 3 as a planned competitor to Nvidia’s H100 and AMD’s MI300X, then moved it toward mass production with performance and efficiency claims. The company also positioned the family through separately priced Gaudi 2 and Gaudi 3 accelerator kits, making commercial uptake—not just chip specifications—the key test.
The withdrawn target exposes friction in that rollout: Intel says customers are navigating a Gaudi 2-to-Gaudi 3 transition while software usability remains an obstacle. That qualifies the earlier Gaudi 3 performance claims with evidence that deployment readiness is shaping demand.
First-order effects
- Intel resets expectations for Gaudi revenue in 2024 and must continue funding the platform without the sales contribution it had forecast.
- Prospective Gaudi buyers face a less settled product transition, while Intel’s cited software ease-of-use issues become an immediate adoption hurdle.
Second-order effects
- Nvidia and AMD benefit competitively insofar as Intel’s alternative takes longer to convert technical positioning into deployed AI capacity.
- Intel’s Gaudi sales effort will need to put more weight on software tooling, migration support, and total-cost-of-ownership proof rather than accelerator specifications alone.
Third-order effects
- The result reinforces that AI accelerator competition is increasingly an integrated hardware-and-software contest, where a new chip generation can disrupt demand for its predecessor before the replacement is broadly usable.
- If similar transition problems persist, buyers may favor platforms with lower migration and software risk, raising the barrier for challengers even when their hardware economics are attractive.
The trend: AI infrastructure competition is shifting from peak chip performance toward the operational maturity of the full developer and deployment stack.