In March 2004, at the American Council for an Energy-Efficient Economy's (ACEEE) Market Transformation Symposium, Chris Calwell of Ecos Consulting unveiled an idea: award a certification badge to desktop PC power supply units (PSUs) that exceeded 80% efficiency. At the time, it was not uncommon for commercial PC power supplies to have efficiencies of only 45–65%. The rest of the power was lost as heat. Funded by the California Energy Commission, PG&E, and the U.S. Environmental Protection Agency (EPA), a methodology for measuring power conversion efficiency was established.
In February 2005, Seasonic's SS-400HT became the first product certified under this program, achieving roughly 83% efficiency at 400W output. Silent PC Review's own measurements showed it reaching 90%, but under the official test protocol the figures were 81.5%, 85.3%, and 83.2% at 20%, 50%, and 100% load respectively. Even so, this far exceeded the market average of the era.
This "80 PLUS" program subsequently added tiers—Bronze, Silver, Gold, Platinum, and Titanium—progressively raising the efficiency ceiling. Titanium, introduced in 2014, required 94% efficiency at 50% load and 90% at 100% load, and it held the top spot for more than a decade.
The Titanium Wall Was a "1% Wall"
Titanium's requirements for 115V non-redundant power supplies are 90% at 10% load, 92% at 20% load, 94% at 50% load, and 90% at 100% load. These figures were achieved through the adoption of new materials such as GaN (gallium nitride) and SiC (silicon carbide), more sophisticated synchronous rectification, and advances in control ICs.
However, Titanium had a structural blind spot: it did not measure efficiency at 5% load. For a 1600W power supply, a 5% load equals 80W output—roughly the power draw of a GPU-equipped workstation sitting idle. Real-world desktops and workstations spend most of their operating hours in this low-load range. Yet the certification system had never looked at this region.
Jeremy Townsend, Senior Vice President at CLEAResult (the organization that runs the 80 PLUS program), explained the significance of Ruby certification this way: "Ruby sets a new standard for what high-performance power supplies can achieve, including efficiency at low load, which is where most systems spend the majority of their time."
Ruby Is a New Standard That Fills the "Low-Load Gap"
On March 12, 2025, CLEAResult announced 80 PLUS Ruby. Initially, it targeted redundant power supplies for data centers (230V, 277V, 480V AC, and 380V DC), setting a stringent bar of 96.5% at 50% load and above 90% at 5% load. Data centers currently account for roughly 0.4% of U.S. electricity consumption, but this figure is projected to surge to 6.7–12% by 2028, making the reduction of power conversion losses an urgent priority.
Then in 2026, Ruby was extended to cover 115V non-redundant power supplies (for desktops and workstations). The requirements differ from those for data centers, with a somewhat relaxed standard.
| Load | Titanium Requirement (115V Non-Redundant) | Ruby Requirement (115V Non-Redundant) | Ruby Requirement (Data Center 230V Redundant) |
|---|---|---|---|
| 5% | Not measured | 85% or higher | 90% or higher |
| 10% | 90% | 91% | 91% |
| 20% | 92% (PF≥0.95) | 93% (PF≥0.95) | 95% (PF≥0.96) |
| 50% | 94% | 95% | 96.5% |
| 100% | 90% | 91% | 92% |
The gap from Titanium is one point at each load level. On paper this may look small, but as Seasonic CEO Walter Sun points out, a 1% improvement from 94% to 95% efficiency is technically harder to achieve than a 1% improvement from 90% to 91%. This is because in a region where absolute losses are already small, squeezing out further reductions becomes increasingly difficult.
What the Measured Values Reveal About the "Margin"
On July 30, 2026, a third-party testing organization measured the Seasonic PRIME RX-1600 at 115V, 60Hz. The results were as follows.
| Load | Ruby Requirement | Measured Efficiency | Margin Over Requirement |
|---|---|---|---|
| 5% (80W output) | 85% | 88.88% | +3.88 points |
| 10% (160W output) | 91% | 93.35% | +2.35 points |
| 20% (320W output) | 93% | 95.38% | +2.38 points |
| 50% (800W output) | 95% | 95.27% | +0.27 points |
| 100% (1600W output) | 91% | 92.81% | +1.81 points |
Power factor reached 0.9831 at 20% load (requirement: 0.95) and 0.9944 at 100% load. Average efficiency came in at 90.06%.
What stands out is the size of the margin at 5% load. At 88.88%, it exceeds the requirement by 3.88 points. For a 1600W power supply, a 5% load corresponds to 80W of output—the typical power draw of a system with a high-end GPU sitting idle. Maintaining efficiency in this range requires optimized switching control, circuit topology switching at light loads, and a comprehensive overhaul of component selection.
At 50% load (800W output), on the other hand, the margin over the requirement is a mere 0.27 points. This is the design's limiting point—the region where Seasonic notes that "the next 1% is difficult."
From 83% to 95% Over 21 Years, But the Weight of "the Last 1%" Differs
Looking back at Seasonic's 80 PLUS history, efficiency has not progressed in a straight line.
| Year | Milestone | Efficiency (Near 50% Load) |
|---|---|---|
| 2005 | SS-400HT, first 80 PLUS certification (Standard) | 85.3% |
| 2008 | First 80 PLUS Gold certification | Approx. 90% |
| 2014 | Titanium standard introduced (industry-wide) | 94% |
| 2026 | PRIME RX-1600, first 115V Ruby certification | 95.27% |
While efficiency improved by roughly 9 points over the nine years from 2005 to 2014, the improvement over the twelve years from 2014 to 2026 amounts to only about 1.3 points. A power supply at 94% efficiency discards roughly 51W as heat out of 854W input; at 95%, that figure drops to about 42W. The difference is just 9W, but Seasonic explains that shaving off this 9W required redesigning the entire platform—switching control, thermal design, component selection, and the verification process.
Why does this "last few percent" matter so much? Walter Sun puts it this way: "We are at a point where every additional percent of efficiency is difficult to achieve, but the effort is worthwhile because these small gains translate into enormous savings on a global scale."
To put this at scale: a 2024 study from ETH Zurich (a master's thesis on power conversion losses in network equipment) estimates that raising a PSU's 80 PLUS tier from Bronze to Titanium alone can reduce overall system energy consumption by 3.2% to 10.7%. If Ruby delivers efficiency beyond this, the reduction margin would expand further still.
What the AI Era Demands from Power Supplies
It is no coincidence that the PRIME ENTERPRISE RX-1600 bears the "Enterprise" name and targets AI systems. Modern AI workstations can draw over 600W from a single high-end NVIDIA GPU alone, with total system consumption reaching 1600W—yet the system drops into idle between training jobs. Handling these sharp load transients while maintaining high efficiency even at low loads: that is the essential requirement Ruby certification imposes.
Compliant with the ATX 3.1 standard, this product features a 12V-2x6 connector (an improved version of the former 12VHPWR) and supports power delivery compliant with the PCIe 5.1 CEM specification. ATX 3.1, published by Intel in September 2023, includes relaxed hold-up time requirements and changes to sense pin configuration.
CLEAResult's Townsend suggests that Ruby will become the industry benchmark going forward: "We expect Ruby to help set the pace for what manufacturers will be working toward heading into 2027. Congratulations to Sea Sonic for this achievement."
Questions That Remain
The retail price and release timing for the PRIME ENTERPRISE RX-1600 have not yet been announced. Given that Seasonic's previous Prime series (Titanium-certified products) were priced in the $300–500 range, the Ruby-certified product is likely to exceed that range.
On the technical side, the thin 0.27-point margin at 50% load is a point of concern. Whether this margin proves sufficient once accounting for unit-to-unit variation and aging in mass production remains to be verified after market launch. Additionally, Ruby certification for 230V environments (such as in Europe) has not yet been achieved. Because the optimal circuit topology differs between 115V and 230V, whether the same platform can cover both remains a separate question.
And the biggest question of all is how much this 1% efficiency improvement can serve as a deterrent against the rising power demands of the AI era. Against the projection (announced by CLEAResult in March 2025) that data center power consumption will reach 6.7–12% of total U.S. electricity by 2028, the contribution a 1% improvement in power supply efficiency can make is limited. That said, improvements in power supply efficiency also translate into reduced cooling loads, so the actual energy-saving effect may be larger than the numbers alone suggest. Whether Seasonic's claim of "enormous savings on a global scale" turns out to be more than rhetoric will depend on just how widely Ruby certification becomes adopted as an industry standard.