On April 27, Resource Recycling reported on Intel’s latest earnings call. But the company’s results also contained a critical detail that was unreported and received little attention in the business press.
Intel’s Q1 non-GAAP gross margin came in at 41%, roughly 650 basis points above the company’s own guidance. Management attributed the beat to a combination of higher volumes, favorable mix, pricing, and better 18A yields. According to industry analyst Ben Bajarin, who posted on X following the earnings call, part of the lift came from yield salvage: selling marginal silicon, much of it edge-die that would normally be binned out or scrapped rather than shipped into a usable SKU. Intel is now capturing revenue from silicon that would previously have been written down or held in reserve.
The demand environment made this possible. With AI and data center buildout running against tight fab capacity, customers are willing to qualify hardware that would previously have been passed over. Intel has responded by pushing deeper into yield salvage, recovering commercial value from dies that would previously have fallen below the threshold for a salable SKU.Â
None of this is technically new. What changed is that the market now prices marginal silicon in a way that makes recovery economics work at the OEM level.
What this means for secondary markets
The competitive implication is indirect but real because primary manufacturers are reclaiming value from material that component brokers, refurbishers, and downstream resellers have historically accessed cheaply. The arbitrage around imperfect or off-spec parts, sourced at low cost and remarketed into lower-demand channels, contracts as OEMs move further down the yield curve themselves. OEM-backed certified refurbishment and take-back programs have been quietly chipping at the same territory for years. Yield salvage at scale extends that pattern further upstream, to silicon that never leaves the fab as scrap in the first place.
The strategic implication means that Intel’s success here rests on granular visibility into die characteristics and precise buyer matching. Secondary market operators have historically worked with considerably less precision — label-level SKUs, basic diagnostic data, limited failure-mode intelligence.
One discipline question sits at the center of this opportunity. Intel operates within strict qualification frameworks and remains acutely sensitive to field failure rates. In the secondary market specifically, the temptation to push marginal parts harder will be real, but a failure in a refurbished server at a cost-sensitive enterprise erodes intermediary trust far more quickly than a comparable failure rate erodes trust in an OEM. Disciplined testing and transparent grading are foundational, not optional.
Intel’s results show that the company turned a supply constraint into a yield opportunity, with the boundary between waste and viable product moving, and primary manufacturers pulling it upstream.
Stakeholders operating in the secondary and component markets should position their capabilities to hold ground at that boundary before more of it disappears.
