Intel says organic substrate materials are approaching their scaling limits by the end of this decade and points to glass as the replacement, citing up to 10x higher interconnect density. Absolics separately says its glass substrate fits up to four times more chips in the same space, while Nippon Electric Glass says it has solved glass's crack-prone processing problem.
Why Is the Organic Substrate Approaching Its Limits, and Why Is Glass Emerging as the Alternative?
Intel said the semiconductor industry will likely reach the limits of scaling transistors on a silicon package using organic materials by the end of this decadeCITE:E1. That projection is presented as an industry-wide expectation rather than a measured production milestone, but it frames the shift toward glass as a response to an approaching ceiling rather than an incremental upgrade.
What Performance Advantages Does Glass Offer Over Organic Substrates?
Intel said glass offers ultra-low flatness along with better thermal and mechanical stability than today's organic substratesCITE:E2. Intel said these properties result in much higher interconnect density in a substrate compared with organic materialsCITE:E2.
How Much of a Performance Gain Can Glass Substrates Deliver?
Intel said a 10x increase in interconnect density is possible on glass substratesCITE:E4. Intel described this as a technical capability it is pursuing rather than a specification already shipping in a production part.
How Does Glass Change Chip Architecture and Chiplet Integration?
Intel said the technology gives chip architects the ability to pack more tiles, also called chiplets, into a smaller footprint on a single packageCITE:E5. That framing ties the interconnect-density gains described above directly to how many separate chiplets a designer can fit onto one glass-based package.
What Progress Have Intel and Absolics Made on Glass Substrates?
Intel said it is on track to deliver complete glass substrate solutions to the market in the second half of this decadeCITE:E3. Separately, Absolics, a unit of SK Group, said its glass substrate product allows larger processors and supports mounting up to four times more chips in the same spaceCITE:E6, a claim it made when breaking ground on its manufacturing site.
| Metric | Value | Entity | Date |
|---|
| Organic material scaling limit (projected) | Reached by end of this decade | Intel | 2023-09-18 |
| Interconnect density gain (glass vs. organic) | Up to 10x | Intel | 2023-09-18 |
| Chip capacity in same footprint (glass substrate) | Up to 4x more chips | Absolics (SK Group) | 2022-11-01 |
| Target for complete glass substrate solutions | Second half of this decade | Intel | 2023-09-18 |
What Are the Manufacturing Challenges for Glass Substrates, and How Are They Being Solved?
Nippon Electric Glass said glass's inherent brittleness and its tendency to crack during processing had been an obstacle to productionCITE:E7. Nippon Electric Glass said it has achieved crack-free CO2 laser processing of glass substrates by collaborating with Via Mechanics, a company it described as highly regarded for its laser processing technologyCITE:E7.
Taken together, the evidence shows glass substrates addressing two separate constraints at once: a density ceiling that Intel says organic materials will hit by the end of this decadeCITE:E1, and a manufacturing defect — crack-prone processing — that Nippon Electric Glass says it has now resolvedCITE:E7. Intel's 10x interconnect-density figureCITE:E4 and Absolics' 4x chip-capacity figureCITE:E6 both describe capability gains rather than shipped production volumes, and Intel's own target for a complete market solution is the second half of this decadeCITE:E3.
Author's Take・Nathan
The two headline figures in this shift — Intel's 10x interconnect-density claim and Absolics' 4x chip-capacity claim — are both capability statements, not shipped specifications, and that gap matters for anyone assessing timing. Intel itself frames delivery of a complete solution as a second-half-of-decade target, not a near-term one. What changes the calculus is the manufacturing side: Nippon Electric Glass's claim of crack-free CO2 laser processing addresses the brittleness problem that has kept glass out of volume production, which is arguably the more load-bearing development here since density gains are meaningless without a repeatable way to cut the material without breaking it. The metric worth watching is whether Intel's second-half-of-decade delivery target holds, since that is the point at which the 10x density figure would need to convert into an actual shipping package rather than a demonstrated capability.