A Quarterback Weighs In on the Moon

Drake Maye is 24 and starts at quarterback for the New England Patriots. This week he went on a radio show and said he was "not really sure 200 percent if we went there or not", referring to the Moon. He added that he took a college class on conspiracies and feels a little skeptical.

Apollo put twelve astronauts on the Moon across six missions between 1969 and 1972. They brought back 2,196 samples weighing 842 pounds, which researchers still study. Observatories have bounced lasers off reflectors the astronauts left to measure the Earth-Moon distance to about 3 centimeters. In 1977, a Soviet radio telescope picked up the transmitters the astronauts left and found their positions matched NASA's reports. NASA's Lunar Reconnaissance Orbiter has imaged the Apollo 12, 14, and 17 sites, and India's Chandrayaan-2 orbiter photographed the Apollo 11 lander in 2021.

Maye can hold any opinion he likes. A 12-year-old who idolizes an NFL quarterback hears an adult they admire treating settled evidence as an open question. That same kid sits in a science class where a teacher is working to build the habit of checking evidence first.

Our country needs more engineers, technicians, and scientists. Students decide early whether math and science belong to them. Every public figure who shrugs at evidence gives a student one more reason to shrug. Celebrities have reach, and reach carries a duty to look something up before speaking.

Maye's conspiracy class presumably covered how to weigh evidence. The Moon landing makes a good practice problem, and one call to any physics department would settle it in ten minutes. The reflectors are still on the Moon, and the physicists still have the measurements.

Fermilab Study Ties Qubit Performance to Nanometer-Scale Fabrication Details

How Microscopic Manufacturing Features Impact Quantum Computing

Key Summary Takeaways

•       Averages Hide Weak Links: Quantum processors are bounded by their weakest qubits, making consistency across a chip essential.

•       Nanometer Precision Matters: Features as small as 1 nanometer in oxide thickness can dramatically impact qubit performance.

•       3 Microscopic Culprits: Surface oxide thickness, sidewall edge slopes, and substrate trench depth directly dictate energy loss.

•       Faster Screening: Advanced techniques like terahertz nanoimaging allow manufacturers to catch defect issues before cooling chips to absolute zero.

Overview

When tech companies talk about quantum computers, they usually share average performance metrics—such as two-qubit gate fidelities between 99.0 and 99.5 percent. However, averages hide a critical challenge in quantum engineering: on any given quantum chip, some qubits perform exceptionally well while others perform poorly. Because a quantum computer is only as strong as its weakest link, understanding why that variation occurs is essential for scaling up functional processors.

A groundbreaking study conducted at Fermilab's SQMS Center (Superconducting Quantum Materials and Systems Center) examined the spread behind these averages. The researchers traced performance differences directly to microscopic physical features measured in single nanometers—a scale tens of thousands of times smaller than a single strand of human hair. The peer-reviewed findings were published in Applied Physics Reviews based on collaborative research involving Northwestern University, Rigetti Computing, NIST, and Ames National Laboratory.

How a Transmon Qubit Works

The most common type of superconducting quantum bit is called a transmon. You can think of a transmon as a tiny electrical circuit made from a thin metal film (typically niobium) laid out on a base chip made of silicon or sapphire. The circuit consists of two primary components:

•       Capacitor: Stores electrical energy across metal pads patterned on the surface.

•       Josephson Junction: A razor-thin insulating barrier sandwiched between two superconductors. This junction gives the circuit non-linear, unevenly spaced energy levels, allowing engineers to isolate the two lowest energy states as the 0 and 1 of a qubit.

To function properly, the chip operates inside a dilution refrigerator at millikelvin temperatures (near absolute zero), where the metal loses all electrical resistance. Microwave pulses are then sent into the circuit to move the qubit between state 0 and state 1.

Understanding the T1 Lifetime & Energy Leaks

A fundamental metric of qubit quality is its T1 lifetime (energy relaxation time). This measures how long a qubit can remain in its excited state ('1') before releasing its energy and dropping back to the ground state ('0'). Because every quantum logic operation takes a set amount of time, a longer T1 allows a quantum computer to execute longer, more complex algorithms before stored information disappears.

A transmon's electric field is not completely confined inside the clean metal pads; it extends slightly into the chip surface, the underlying substrate, and the surrounding air. Energy stored in clean, defect-free material returns safely to the circuit. However, energy stored in surface materials with atomic defects gets absorbed and turned into heat or stray vibration. Engineers call this share of field energy sitting in lossy material the Energy Participation Ratio—and lower participation in lossy material means a longer-lasting qubit.

The Three Microscopic Culprits Behind Variance

By testing 22 transmon qubits across 7 multi-institution chips using a rigorous blinded protocol, the study identified three distinct manufacturing features that control how much energy leaks away:

1. Surface 'Rust' (Niobium Oxide Layer Thickness)

When niobium metal is exposed to ambient air during manufacturing, it naturally reacts with oxygen to form a thin surface oxide layer (primarily Nb₂O₅) a few nanometers thick. This oxide layer is naturally disordered and contains atomic defects that trap microwave energy. The study revealed that a variation of just 1 single nanometer in oxide thickness caused a noticeable difference in qubit lifetime, surprising even experienced chip fabricators.

2. Edge Slope (Sidewall Geometry & Footer Angle)

When chemical etching carves the metal film into qubit circuit patterns, it leaves an edge profile. The researchers compared sloped, tapered edges (about 30-degree angles) against near-vertical, sharper edges (10 to 15 degrees). Sloped edges increase the physical surface area where lossy oxide forms, exposing 20 to 30 percent more of the qubit's electric field to defective surface material and significantly lowering performance.

3. Moat Depth (Substrate Trenching Saturation)

To reduce electric field concentration at the chip surface, manufacturers etch shallow trenches into the substrate beside the metal pads. The study found that digging even a shallow trench under 20 nanometers produces an immediate, dramatic boost in qubit quality. However, once trench depth exceeds 70 nanometers, the benefit flattens out and saturates. Beyond 70 nm, deeper trenching no longer helps, and the sidewall edge angle takes over as the dominant factor driving energy loss.

Diagnostic Proxies: Testing Without the Deep Freeze

Cooling a quantum chip down to millikelvin temperatures inside a specialized dilution refrigerator takes days of setup and substantial operational expense. One of the study's key practical breakthroughs was identifying two room-temperature or mild-cold diagnostic tools as effective predictors of qubit quality:

•       Terahertz Near-Field Nanoimaging (s-SNOM): Maps localized dielectric loss and oxide non-uniformities at nanometer resolution.

•       Magneto-Optical Imaging: Screens local superconducting transition properties across the metal film.

By using these proxy techniques as a screening step on the cleanroom floor, fabrication teams can receive rapid feedback on each etch and oxidation step—catching bad chips before spending time and money on full cryogenic testing.

What This Means for Future Quantum Hardware

To scale up quantum processors from dozens of experimental qubits to thousands of fault-tolerant logical qubits, manufacturing consistency is everything. Eliminating nanometer-scale variations during chip fabrication will allow every qubit on a processor to perform like the best qubit on the chip.

An expanded section covering these device-to-device yield limits and nanometer-scale fabrication parameters will be included in Chapter 5 of the December 2026 third edition of Quantum from the Ground Up. You can download current book editions here.

The Trout Stream Under the Turnpike, and the Data Center Over the Barnes Aquifer

Image source https://tinyurl.com/2t3kumjt

When my Dad drove me, he let me out at East Mountain Golf Course. He said to be home for supper. I said I would. He drove off and I went down to the water.

I was ten when I first fished it. The stream came out of the woods below the golf course. It was small and cold and clear. You could step across it in three places. I fished it downstream, one pool at a time, and I kept low and quiet. The brook trout were native and they had been there longer than the golf course, the turnpike and the town. They were dark on the back and orange on the belly, with white edges on the fins. I kept a few. I put the rest back.

I never saw another person on it. There were no boot prints in the mud and no line in the branches. No beer cans, no bait tins. The golfers stayed on the fairways and the people in the cars on the turnpike never looked down. I think I was the only one who ever fished that section. I told no one. My friends did not know. I kept it to myself and the trout stayed where they were.

There was a pool in the woods where the stream bent against a root bank. It was deep on the outside of the bend and slow. A big brook trout lived there. I saw the fish the first time as a shadow that moved when I moved. I put a worm over the edge of the current and let it drift under the roots. The trout took it hard.

It was the biggest brook trout I had ever seen. Its belly was deep orange. Its back was dark green with pale spots, and the red spots had blue rings around them. I held the trout in the water and it stayed there a moment, then it went. I let it go. I could have kept it. I let it go anyway.

I caught the trout several times that way. I knew it by the size and by the pool. Each time I let it go. Nobody else knew it was there, so nobody else would have let it go.

The woods got thick below the pool and I followed the stream through them. Then the Massachusetts Turnpike came over the top and the stream went into a tunnel. Trucks ran overhead and the concrete shook a little. Inside it was dark and loud and the water was colder than outside. I waded in with the rod held low. The big brown trout lived under there. They held in the current, facing upstream, and waited for the water to bring them something to eat. Sometimes I could see them. They were always hard to catch. I caught a few over the years and let every one go. The one I remember best, I lost. I had hooked a big brown and I had no net. I never carried a net because it tangled in the brush. I had the trout in my hand. It flipped its tail and spat the hook. I was going to let it go anyway. Still. I stood in the dark a while after that.

On the other side the stream came out into the light and ran on to the railroad tracks. There I left it. I climbed the cinders and walked the rails home to 342 Holyoke Road. The ties were spaced wrong for my legs. I walked them anyway.

When my Dad did not drive me, I went the other way. I left the house in the morning and cut through the woods to the tracks. I followed the rails until the stream passed under them. I started there and fished upstream toward the tunnel, and then on to the pool. My Mom knew the time to expect me and that was enough.

The water was cold in July. It came out of the ground that cold and stayed that cold all summer. I never wondered why. A boy does not wonder about water that works. I know now where it comes from. It comes from the Barnes Aquifer, and I fished it for years without knowing its name.

The last time I fished the stream was (I think) the summer between my sophomore and junior years of college. That was about fifty years ago. By then I had switched to a 1 weight 7 foot 6 inch fly rod. The trout in that stream always ate up hellgrammite pattern flies. My parents' house sold last month, and I doubt I will fish it again. I wrote about the sale and the empty rooms.

Thousands of cars cross the stream every day on the turnpike. I doubt anyone knows it is there. I always look. I can see a small section of it. It looks like beavers have dammed part of it. They found the stream too.

I do not know if the brook trout are still there. The big one is gone. Brook trout seldom live past five years. I do not know if the water is still cold in July.

A data center campus is now proposed over the aquifer. In 2021, the Westfield City Council approved a $4 billion campus at 199 Servistar Industrial Way, 1.8 miles from my childhood house. The vote passed 9 to 3. The site sits partly on wetlands and directly over the Barnes Aquifer, the region's municipal drinking water source. In July 2026, the council voted unanimously for a one year moratorium on new data centers. The campus would draw 274 megawatts.

Servistar says it plans a closed-loop cooling system. The reporting I reviewed does not name the cooling vendor or design. NVIDIA's current liquid cooling design uses dry coolers and claims zero water consumption on site. Dry coolers can demand 10 to 35 percent more electricity than evaporative towers. Lawrence Berkeley National Laboratory found that 92.5 percent of a data center's water footprint comes from generating its electricity. Servistar's plans include natural gas generators. I covered the engineering in more detail in my post on NVIDIA's cooling design.

State certification standards will require the developer to disclose cooling method, power source, and noise mitigation before construction proceeds.

I do not know what the campus will do to the stream. Brook trout need cold water, and that stream ran cold because of the aquifer. I never told anyone about it. I am telling you now, so that when the disclosures come out, you know a stream runs under the turnpike on that water.

Inside NVIDIA's 113-Degree Cooling Design, and Westfield's Data Center Question

Given the strong debate and varied perspectives surrounding the proposed Westfield (my hometown) data center campus, this post focuses strictly on verifiable technical specifications and empirical data as I understand it. The goal is to provide an objective, fact-based overview of the project's engineering parameters, NVIDIA's recent high-temperature liquid cooling design, local climate limits, and broader infrastructure trade-offs. Read my earlier post on the topic here.

I grew up at 342 Holyoke Road in Westfield, Massachusetts. In 2021, the City Council approved a $4 billion data center campus at 199 Servistar Industrial Way, 1.8 miles from that house. The vote passed 9 to 3. The site sits partly on wetlands and directly over the Barnes Aquifer, the region's municipal drinking water source. In July 2026, the council voted unanimously for a one year moratorium on new data centers, and the campus would draw 274 megawatts. Servistar says it plans a closed-loop cooling system. The reporting I reviewed does not name the cooling vendor or design.

In June 2026, NVIDIA published a cooling design that addresses the same question. It appears in the DSX reference design, a guide for building the full AI factory infrastructure stack. NVIDIA's Ali Heydari says the design has zero water consumption and has eliminated most power use for cooling. NVIDIA's own description of the Vera Rubin NVL72 system calls it single-phase direct liquid cooling with a 113°F supply temperature.

Earlier liquid-cooled deployments supplied water in the 80°F to 90°F range. Those systems cooled the CPUs and GPUs with cold plates and left other components to air. Rubin cools every chip and networking component by liquid, eliminating fans inside the server chassis and compute racks. The coolant is a 75 percent water and 25 percent propylene glycol mix that enters the chip at 113°F and leaves near 131°F. Operators have traditionally recommended an ambient temperature of 64°F to 81°F. NVIDIA's summary also lists a 6U system now fitting in 2U and no hot or cold aisle management.

Inside the building, a Cooling Distribution Unit (CDU) uses a liquid-to-liquid heat exchanger to isolate the secondary coolant loop in the server racks from the primary facility loop. The primary loop then carries heat outside to dry coolers. NVIDIA says they reject heat efficiently for much of the year and that the loop avoids evaporative cooling 99 percent of the time. In favorable climates, NVIDIA says this cuts water use from roughly 2.6 million gallons per megawatt per year to near zero. Hotter climates such as Phoenix may still need chillers on peak summer days. The captured heat can also be reused to warm nearby buildings.

Heat path from chip to outdoor air. Coolant temperatures are NVIDIA figures.

Westfield's climate aligns with the thermal operating requirements for those dry coolers. The average July high is 83°F, and the average January high is 34°F. Barnes Municipal Airport recorded a high of 98.1°F in July 2026. Because fluid leaves the chips near 131°F and typical heat exchangers require a 5°F to 9°F approach margin, outdoor air only needs to remain below about 104°F to supply 113°F coolant back to the racks. As a result, dry coolers in Westfield can maintain a 113°F supply temperature year-round without requiring supplemental mechanical chillers. Nearby Springfield has gained 11 more above-average summer days since 1970, but local peaks remain within the system's thermal operating limit. A campus drawing 274 megawatts releases close to that much heat, and dry coolers send it to outdoor air.

Westfield outdoor temperatures compared with coolant temperatures. The 104°F limit is an estimate.

NVIDIA estimates that a 50MW facility saves over $4 million a year in cooling energy and water. Cooling has accounted for up to 40 percent of a data center's electricity. One reviewer notes the capital cost premium over air cooling remains unknown.

Environmental and energy analyses point out that on-site claims exclude water consumed by off-site power plants supplying the grid. Lawrence Berkeley National Laboratory found that 92.5 percent of a data center's water footprint comes from generating its electricity. Dry coolers can also demand 10 to 35 percent more electricity than evaporative towers. Servistar's plans include natural gas generators, and Westfield Gas & Electric estimates a gas pipeline at about $20 million.

State certification standards will require the developer to disclose cooling method, power source, and noise mitigation before construction proceeds. 

Every source is linked above for you to read and weigh.

IonQ Runs Real-Time Error Decoding on a Single CPU

On September 22, IonQ reported a real-time error decoder that runs on one standard off-the-shelf CPU. Decoding is the classical half of quantum error correction, and it sets a speed limit on the quantum half. This post covers what the decoder does, the architecture it serves, and what the test did and did not show.

What a Decoder Does

An error-correcting code measures parity checks on the qubits every cycle. The results, called syndromes, do not reveal the stored data. They show where errors likely occurred. A classical decoder reads the syndrome stream, infers the most probable errors, and hands corrections back to the machine. Some logical operations wait on that answer before the program can continue, so a slow decoder stalls the quantum computer. IonQ describes this as the failure of conventional approaches, where the classical side gets overwhelmed and the quantum system has to pause. Streaming decoders address it by processing syndromes faster than they accumulate.

The Architecture It Serves

IonQ published its Walking Cat architecture in April as a full blueprint for a trapped-ion fault-tolerant computer, covering the compiler, error-correction protocols, micro-architecture, and decoder. It builds entirely on low-density parity-check (LDPC) codes. In the notation [[n, k, d]], n is physical qubits, k is logical qubits, and d is the code distance. The paper introduces a [[70, 6, 9]] code for fast logical gates and a [[102, 22, 9]] code that packs 22 logical qubits into each memory block.

The hardware is a quantum charge-coupled device (QCCD) chip. Electric fields shuttle ions between storage and interaction zones, which lets the chip implement the non-local connections LDPC codes need. A cat factory produces cat states that travel through the machine and get consumed by logical operations. Reservoirs of fresh ions replace qubits lost during operation. The paper's dense design reaches 110 logical qubits and about one million T gates per day with 2,514 physical qubits. Speed is the tradeoff: IonQ estimates a 30-bit Shor factoring run takes about 23 hours.

Trapped-ion cycles run slower than superconducting ones, so the decoder works on a millisecond-scale budget. The April paper described a streaming beam decoder that works on syndrome data in sliding windows to fit that budget.

Testing the Decoder at 408 Logical Qubits

The new paper, Real-time decoder for a MegaQuOp quantum computer using a single CPU, evaluates a dual-decoder architecture on benchmark circuits. The circuits simulate up to 408 logical qubits across 88 memory blocks and magic state factories, and they run more than 31.5 million operations. Under standard operational noise, the decoder added as little as 0.02% stretch time, meaning extra run time spent waiting on decoding. The simulation exceeds the 110 logical qubits in the April dense design.

IonQ says the result shows classical hardware overhead does not have to grow exponentially with logical qubits or circuit depth. That claim matters for scaling, because a decoder that needs more classical hardware with every added qubit would cap the machine size. The company's roadmap runs past 256 physical qubits toward thousands.

Limits of the Result

Every circuit was simulated. The figures come from IonQ's own paper and press release, and they reflect a standard noise model. Hardware adds its own error mix, including the ion loss the April architecture handles with reservoirs. Whether the decoder holds its 0.02% overhead on a running machine is a hardware question, and no device with 408 logical qubits exists to answer it.

For readers of Quantum from the Ground Up: this adds a CPU-based decoder alongside the NVIDIA decoder in Chapter 12, and a classical-side entry to IonQ's coverage in Chapter 6. The two decoders report different metrics, so no ranking follows. The Q-Day range in Chapter 13 stays where it is until decoder results come from hardware. This post will be incorporated into the next edition. The current edition is at gordostuff.com/p/quantum-from-ground-up-hardware.html.

950 Claude Agents Found Something in Phage DNA

PCR (Polymerase Chain Reaction) was still something being figured out when I was an undergrad studying microbiology at UMass Amherst. ELISA (Enzyme-Linked Immunosorbent Assay) had been out for a few years but very few labs were running it. . Most of our work still ran on overnight cultures: streak a plate, incubate it overnight, read the colonies in the morning. A literature search meant pulling bound journals off library shelves.

On September 23, Anthropic announced that Claude agents found a previously uncharacterized enzyme system in bacteriophages, the viruses that infect bacteria. Anthropic scientists gave Claude one prompt: search a massive DNA sequence database for interesting reverse transcriptases, enzymes that copy RNA into DNA. About 950 agents ran for 21 hours and used 210 million tokens. The output:

•       200,000+ reverse transcriptases gathered

•       3,500 new candidate systems flagged

•       20 candidates written up as human-readable reports

One agent reading raw DNA beside an unusual enzyme spotted a tandem repeat array by eye. It counted the repeats, measured their spacing, compared the layout with known systems, and checked the literature before filing a report. Anthropic named the system Array-associated Reverse Transcriptases, or ART: the enzyme, a partner gene, and a repeat array. Lab work confirmed the array is transcribed into short RNAs.

Earlier studies had identified the enzyme; Claude was first to connect the array and partner protein. Nobody knows what ART does. The preprint has no peer review yet, and reruns struggled to reproduce the find. CRISPR pioneer Feng Zhang reviewed the preprint and said the result merits further investigation.

Anthropic says this kind of genome mining takes an expert weeks to months. Humans still run every experiment in its lab. The constraint now sits at the bench.

PCR took years to move from a new method into every teaching lab, and it needed Taq polymerase from a Yellowstone hot spring bacterium to get there. ART has 21 hours of agent time and one lab result behind it. Robots can pipette these days but so far anyways, they cannot dream up what to test next…. At least not very well.

950 agents turned a search that would take an expert months into a Tuesday afternoon, and handed a lab twenty leads worth checking.

DOE Puts $215 Million on 100 Logical Qubits

My quarterly-updated book, Quantum from the Ground Up, tracks the hardware race toward fault-tolerant quantum computers. The Department of Energy just put a date on it. On September 17, DOE launched the Quantum Genesis Q Competition with up to $215 million planned. Teams must deploy at least 100 logical qubits running hundreds of millions of fault-tolerant operations on chemistry, materials, physics, and applied mathematics problems. The evaluation is set for September 2028.

I've written prior - a logical qubit combines many physical qubits under error correction so the result survives a long calculation. Chapter 4 of the book walks through how that works. Some chips already carry more than 1,000 physical qubits, and those qubits still make too many errors to finish useful work. Here's the DOE competition breakdown.

Phase I awards up to $1.5 million per team for early milestones. Phase II holds $100 million for teams that reach 100 logical qubits, plus two $50 million bonus pools at 150 and 200. HPCwire reports that each pool splits equally among the teams that qualify, so a team at 200 logical qubits draws from all three.

       Structure of the DOE Quantum Genesis Q Competition. Source: U.S. DOE, HPCwire

Verification sits with the national labs. DOE plans a $45 million testbed to benchmark competitors' claims independently. Vendors submit results, and the labs check them before any pool pays out.

QuEra and Harvard have shown 96 logical qubits on neutral atoms, the figure cited in Chapter 8, so at least one platform sits near the count. The operation depth and outside verification set the higher bar. Microsoft's Majorana roadmap, covered in Chapter 10, targets 2029, a year after DOE's evaluation.

Two caveats. Only $2.5 million of the $215 million comes from FY2026 funds, and Congress controls the rest. And 100 logical qubits sits far below the roughly 1 million physical qubits Gidney estimated in 2025 for breaking RSA-2048, the baseline in Chapter 13. A 100-logical-qubit machine cannot break RSA-2048, so the schedule for moving to post-quantum encryption stays the same.

Applications close October 19. DOE holds an applicant webinar September 25 at 1:00 p.m. ET. This post goes into the next edition (December 2026) of Quantum from the Ground Up, available here.