The original Bitcoin Mining Singularity paper proposed a measurable milestone for Bitcoin mining hardware: the point at which an ASIC reaches 1 joule per terahash (1 J/TH). The original model used a 9.5 J/TH baseline and historical efficiency-improvement scenarios of 15%, 21% and 25% per year to place that milestone broadly inside a 2033–2039 window.

Now there is a new data point. Bitmain lists the Antminer S23 XP Hyd. at 600 TH/s, 5,340 watts and 8.9 J/TH, with futures shipments beginning in November 2026. The previous S23 Hyd. specification was 580 TH/s, 5,510 watts and 9.5 J/TH. The new machine therefore gives us an opportunity to revisit the model using a real announced hardware specification rather than simply extending the old curve.

What the Updated Chart Shows

Figure 1 — Bitcoin Mining Singularity: Updated ASIC Efficiency Paths. The updated chart begins with the 2025 S23 Hyd. baseline of 9.5 J/TH and adds Bitmain’s newly announced S23 XP Hyd. at 8.9 J/TH in 2026. The original Bitcoin Mining Singularity model anticipated that long-run ASIC efficiency gains could continue within roughly 15%, 21% and 25% annual improvement scenarios, producing a 1 J/TH crossing during the 2030s. The newest flagship step, however, improves rated efficiency from 9.5 to 8.9 J/TH—about 6.32%—while hashrate rises from 580 to 600 TH/s, about 3.45%. The chart therefore adds a slower dashed trajectory based on that latest 9.5→8.9 step. This line is not presented as a newly established industry growth rate; one product transition is insufficient to determine a long-term trend. Instead, it tests the original anticipation against new evidence. If future generations return to the historical 15–25% efficiency range, the original Singularity window remains plausible under the model. If improvements remain near the latest single-step pace for many generations, the 1 J/TH milestone would move much farther into the future. The emerging possibility is exactly what the original Thermoeconomic Capitulation framework was designed to examine: as ASICs become more efficient, each additional reduction in J/TH may become progressively harder, producing a flatter efficiency curve even while hardware continues to improve.

9.5 to 8.9 J/TH Is Still Progress

The difference between 9.5 and 8.9 can look small compared with earlier generations of Bitcoin miners, but that is precisely why the new result matters. The percentage reduction is:

(9.5 − 8.9) ÷ 9.5 × 100 = 6.32%

Lower J/TH is better. At the manufacturer’s rated specifications, the S23 XP Hyd. produces more hashes while consuming less total wall power than the S23 Hyd. specification used in the previous model. Bitmain’s own 2026 material described the 9.5 J/TH S23 Hyd. as its first miner below 10 J/TH. Moving the flagship specification to 8.9 J/TH pushes the industry farther below that threshold.

But the magnitude of the step is smaller than the 15–25% annual efficiency range emphasized in the original research. That does not invalidate the earlier historical average. It gives us a new observation that can be used to test whether the curve is beginning to flatten.

This May Be Thermoeconomic Capitulation in Real Time

The original paper did not argue that ASIC efficiency could improve exponentially forever. Its central idea was the opposite: technological progress encounters physical, thermal, manufacturing and economic resistance.

The Thermoeconomic Capitulation model represents efficiency as:

J(t) = J₀ × (1 − r)^(t − t₀)

where J₀ is the starting J/TH efficiency, r is the assumed improvement rate and t is time. The original paper used 9.5 J/TH as its modern baseline and modeled several rates rather than claiming that one fixed percentage must persist forever.

The updated evidence suggests that distinction was important. Early ASIC generations could make enormous efficiency jumps. Once hardware is already below 10 J/TH, removing another joule per terahash becomes a much more demanding engineering problem.

The Original Singularity Window Is Now a Testable Forecast

At 25% annual improvement, the model reaches approximately 1 J/TH in the early 2030s. At 21%, the crossing occurs later, while a 15% trajectory moves it toward the end of the original 2033–2039 Singularity Zone.

The 8.9 J/TH machine does not give us enough evidence to discard those trajectories. ASIC development is not smooth. Product launches occur at irregular intervals, process nodes change, architectures change, cooling systems change and manufacturers can make larger improvements in one generation than another.

What has changed is that we now have a useful competing scenario. If a roughly 6% step became representative of the long-term rate, the Singularity would be delayed substantially. If the next generation produces another double-digit improvement, the historical band may reassert itself.

That means every major ASIC generation from this point forward becomes a measurement against the forecast.

Hashrate Is Also Moving More Slowly in This Flagship Comparison

Efficiency is only half of the original Bitcoin Mining Singularity research. The paper also anticipated major increases in per-machine computational power.

The S23 Hyd. specification was 580 TH/s. The new S23 XP Hyd. is 600 TH/s. That is an increase of about 3.45% between these particular flagship specifications.

Again, two machines do not establish a permanent rate. But this is worth watching because the original framework anticipated that per-machine hashpower could continue climbing rapidly even as efficiency improvements became harder.

If both J/TH improvement and per-machine hashrate growth begin slowing at the same time, the model will need to place more weight on diminishing engineering returns. If future hydro generations make another large jump, this period may instead prove to be a temporary plateau.

The Most Important Number Is Now the Next Number

The 8.9 J/TH S23 XP Hyd. does not prove the Bitcoin Mining Singularity will arrive in 2033. It also does not prove that it has been pushed decades away.

It does something more scientifically useful: it gives the original forecast another point that can be tested.

The original research anticipated two forces operating simultaneously. Engineering would continue pushing J/TH downward, while thermodynamic and economic resistance would make each additional improvement more difficult. The move from 9.5 to 8.9 J/TH is consistent with continued progress, but its comparatively modest size is also consistent with the possibility of diminishing efficiency returns.

For now, the 2033–2039 Bitcoin Mining Singularity remains a model-derived window rather than an established future event. The correct response to the new machine is not to force it onto the old curve. It is to add it to the evidence and watch what happens next.

When the next flagship ASIC arrives, its J/TH figure will tell us whether the curve is bending back toward 15–25% improvement—or whether the industry has entered a slower phase of Thermoeconomic Capitulation.

BitcoinVersus.Tech Editor’s Note:
We volunteer daily to ensure the credibility of the information on this platform is Verifiably True. If you would like to support to help further secure the integrity of our research initiatives, please donate here: 3C9o19EH5HSiwEPyCTmEKzxhNCbo2X6TTb

https://x.com/1BitcoinVersus/status/1937006164555993338

BitcoinVersus.tech is not a financial advisor. This media platform reports on financial subjects purely for informational purposes.

Disclaimer: The underlying inputs, calculations, and mathematical relationships used in the Power Efficiency Index are intended to be transparent and verifiable. The math is verifiable; the interpretation is experimental. PEI is a research framework, not a prediction of future price, guaranteed fair value, or financial advice. Actual market prices are influenced by many variables outside the model, including supply, demand, liquidity, regulation, adoption, speculation, and broader economic conditions.