Strawman · Personal Prep
IonQ Partnership Visit Brief
Wednesday, May 13, 2026 · CASE Building, CU Boulder · Business casual
Where & when
Address
1725 Euclid Ave
CASE Building, Boulder, CO 80309
Parking
Euclid Parking Garage
1725 Euclid Ave
Main room
E422
East Wing, 4th floor
Your breakout
Room W311
11:00 AM – 12:00 PM
Date
Wed, May 13, 2026
8:15 AM start (optional breakfast)
01 · The TL;DR
What you actually need to know
You're in the Quantum Computing breakout from 11:00 AM to 12:00 PM in Room W311, with IonQ leads Chris Ballance (President) and David Allcock (VP Science, Boulder). Your CU peers in that room are Ramin Ayanzadeh (QUASAR Lab) and Scott Sternberg (CUbit, facilitator).
You don't owe a talk. You owe a sharp answer when someone asks what Research Computing brings, plus 2–3 concrete partnership ideas you're ready to put on the table. The breakout reports out 2–3 collaboration next steps in the 12:30 plenary.
The breakout is computing-focused. Sensing, networking, materials, and applications are running in parallel elsewhere — so assume technical fluency in your room.
02 · Logistics
The day at a glance
Where: Center for Academic Success & Engagement (CASE), 1725 Euclid Ave. Park at the Euclid Parking Garage. Main room is E422 (East Wing, 4th floor); your breakout is in W311.
-
8:15–9:00
Optional networking breakfastCASE E422
-
9:00–9:30
Welcome & collaboration vision (Ruzzene, Sternberg, Ballance)CASE E422
-
9:30–10:15
IonQ Strategic Priorities — Compute, Sensing, Interconnects, Networking, SLED/FederalCASE E422
-
10:15–10:25
Break
-
10:25–11:00
CU Research Showcase lightning talks — Diddams, Sternberg, Kitching, ShalmCASE E422
-
11:00–12:00
Quantum Computing breakoutCASE W311 · Ballance + Allcock (IonQ); Sternberg (CU facilitator); Ramsey, Ayanzadeh
-
12:00–12:30
Lunch & be seatedCASE E422
-
12:30–1:20
Collaboration roadmap report-outs — 2–3 next steps per breakoutCASE E422
-
1:20–1:30
Session close & next steps
-
1:30–3:00
Optional lab tour — Diddams Lab (ECNW 1B25), federal funding breakout
Note: the prep brief lists your breakout as running until 12:15. The official agenda says 11:00–12:00 with report-outs in the 12:30 slot. Going with the agenda.
03 · Context
Who is IonQ
IonQ (NYSE: IONQ) is a publicly traded, full-stack quantum platform company headquartered in College Park, MD. It was founded in 2015 by Chris Monroe (UMD) and Jungsang Kim (Duke) as a spinout from foundational trapped-ion research at both universities. Trapped-ion is their bet: highly stable atoms held in magnetic/electric fields as qubits, prized for fidelity and connectivity.
Market cap
~$19B (May 2026)
Q1 2026 revenue
$64.7M · +755% YoY
What they sell
Five strategic areas, per their agenda framing for tomorrow: Compute, Sensing, Interconnects, Networking, and SLED/Federal. Q1 2026 revenue split was roughly 60% commercial, 35% international, 35% multi-product — meaning they've successfully moved past pure-compute revenue into networking, sensing, and post-quantum security.
Recent moves that matter for tomorrow
IonQ has been on an acquisition tear. They closed the $1.08B Oxford Ionics acquisition in January 2026, which is why Chris Ballance is in the room as President (he was Oxford Ionics' co-founder and CEO). They opened a new Boulder office coinciding with that close, where David Allcock leads science. They also closed ID Quantique, Lightsynq, and Capella Space, with intent to acquire Vector Atomic for sensing.
In Q1 2026 they sold their first 256-qubit chip-based system to the University of Cambridge, published a fault-tolerance architectural blueprint, and were selected for DARPA HARQ, MDA SHIELD, and a $39M SDA HALO contract. Their roadmap claims 10,000+ qubits at 99.99999% logical accuracy by 2027 and 2M qubits by 2030.
Their university partnership pattern
Their flagship is the $7.5M expanded QLab partnership with UMD (April 2026, on top of an earlier $9M), which bundles compute access, networking testbed, quantum memory node deployment, and workforce development. They have similar (smaller) arrangements with Cambridge, University of Chicago, AFRL, and state partnerships in Tennessee/EPB and South Carolina. The pattern: hands-on student access to commercial systems, joint research, talent pipeline.
So what for you: They know what a university partnership looks like, and they've done bigger ones. The question for Boulder is what makes us a useful complement, not a duplicate of Maryland.
04 · Decoder
Quantum vocabulary, fast
Enough to follow the room without flinching. Skim once before you walk in.
- Qubit quantum bit
- The quantum version of a classical bit. A classical bit is 0 or 1. A qubit can be in a weighted mix of both at once (superposition) while it's computing — that's where the parallel power comes from. When you measure it, it collapses back to a 0 or 1.
- Trapped-ion IonQ's bet
- Single atoms (ions) held in midair by electromagnetic fields, manipulated with lasers. The atom is the qubit. Strengths: very high accuracy per operation, every qubit can talk to every other qubit (full connectivity), qubits stay coherent a long time. Trade-off: operations are slower than the rival approach.
- Superconducting the rival approach
- What IBM, Google, and Rigetti use. Tiny circuits cooled to near absolute zero. Faster operations, but lower fidelity and limited connectivity. Not what IonQ does — IonQ's whole pitch is that fidelity wins over speed.
- Fidelity
- How accurate one quantum operation is. 99.99% sounds great, but it compounds — over a thousand operations, errors stack up fast. The whole race is adding more nines. IonQ holds records here.
- Coherence
- How long a qubit stays useful before environmental noise scrambles it. Trapped ions have much longer coherence than superconducting qubits. Another IonQ talking point.
- Gate
- A quantum operation — the qubit equivalent of an AND/OR gate. Quantum algorithms are sequences of gates.
- Fault tolerance
- The holy grail. Bundling many physical qubits together with error correction to make one reliable "logical qubit." IonQ recently published a roadmap claiming 2 million physical qubits → useful logical qubits by 2030.
- QEC Quantum Error Correction
- The techniques that make fault tolerance work. Detect and correct errors during the computation, not after.
- AQ Algorithmic Qubits — IonQ's preferred benchmark
- Instead of just counting physical qubits, AQ measures how big a real problem the machine can actually solve well. IonQ talks about AQ35, AQ36, and so on — bigger number is better.
- VQE Variational Quantum Eigensolver
- A hybrid quantum-classical algorithm. The quantum machine prepares trial states; a classical computer optimizes the parameters in a loop. Used for chemistry, materials science, ground-state energy problems.
- QAOA Quantum Approximate Optimization Algorithm
- Similar hybrid idea, but for combinatorial optimization — logistics, scheduling, portfolios. Quantum proposes candidate solutions; classical refines.
Why hybrid is your pitch
Both VQE and QAOA need a classical compute partner running the optimization loop. That is literally your value proposition — Alpine is the classical side. Every time you hear "hybrid algorithm," your hand should go up.
- Simulator vs. hardware
- A "simulator" is a classical program (often GPU-accelerated) that pretends to be a quantum machine. Researchers debug and benchmark on simulators because real quantum hardware is expensive and time-shared. cuQuantum, Qiskit Aer GPU, and Pennylane-Lightning-GPU are all GPU-accelerated simulators that run on hardware like Alpine.
- SDKs what researchers actually code in
- Qiskit (IBM, dominant), Cirq (Google), Pennylane (Xanadu, hardware-agnostic, strong on quantum ML), cuQuantum (NVIDIA's GPU acceleration library), IonQ SDK (targets IonQ hardware). When you offer a "hosted SDK environment," you mean CURC installs and maintains these so researchers don't fight setup.
05 · People
Who's in your room
IonQ · Lead
Chris Ballance
President, Quantum Computing (Oxford)
Oxford Ionics co-founder and former CEO. Now runs IonQ's compute business post-acquisition. Physicist, Cavendish/Oxford background. Will care about technical credibility.
IonQ · Lead
David Allcock
VP Science (Boulder)
Runs IonQ's Boulder science office (opened alongside the Oxford Ionics close). Local presence is the strongest argument for closer integration than a typical remote partnership.
CU · Facilitator
Scott Sternberg
Executive Director, CUbit Quantum Initiative
Runs the room. Former Vaisala Inc. President; runs CU's quantum strategy. He'll set the framing; you complement, not compete.
CU · Faculty
Ramin Ayanzadeh
Director, QUASAR Lab · CS & ECEE
2026 NSF CAREER awardee on trustworthy quantum optimization. Has prior IonQ engagement. Quantum architecture/systems and quantum ML focus. Strong natural pairing with Alpine for hybrid workflows.
06 · Your opener
The 30-second pitch
Use this when Ballance or Allcock asks what Research Computing does:
Say it out loud once before you walk in
"CURC runs the classical compute, GPU, and storage infrastructure that supports research across CU Boulder. For a quantum partnership, that means we're the classical side of any hybrid workflow our faculty run with you — simulation, benchmarking, data, the SDK environment researchers actually log into. We're also where the workforce-development pipeline lives operationally: user support, training, access pathways. We're not a quantum hardware shop. We're the infrastructure that lets your hardware be productive at CU."
If they want a number: Alpine has roughly 21,000 CPU cores and a growing GPU footprint including new RTX Pro 6000 Blackwell and H200 NVL coming online. South Pod expansion lands mid-June. PetaLibrary handles research data at petabyte scale.
07 · Capability
What CURC concretely brings
Five capabilities, in priority order for this conversation.
-
Classical compute for hybrid quantum-classical workflows
VQE, QAOA, quantum chemistry, optimization — every near-term algorithm pairs IonQ hardware with substantial classical compute. Alpine is the obvious classical partner for CU researchers running on IonQ. New Blackwell and H200 capacity matters here.
-
GPU-accelerated quantum simulation
Researchers prototype, debug, and benchmark on simulators before paying for real hardware time. cuQuantum, Qiskit Aer GPU, Pennylane-Lightning-GPU all benefit from the GPU capacity we're standing up. IonQ's own science org may want capacity here too.
-
Software environment & access infrastructure
Open OnDemand, container support (Apptainer/Singularity), module systems, JupyterHub-style notebook access. We can host curated quantum SDK environments so researchers don't fight installs. This is the operational layer that makes the partnership feel frictionless to a faculty member.
-
Data hosting, benchmarks, reproducibility
PetaLibrary for joint research data and benchmark archives. For comparative work — IonQ hardware vs. classical simulation — data needs a long-term home with reproducibility guarantees.
-
User support & workforce-development infrastructure
The seven-person team you lead, plus the broader CURC user support function, builds the training, documentation, and onboarding pathways researchers actually use. Workforce development is a stated IonQ priority. The operational pipeline lives here.
08 · Your owned ideas
Concrete ideas to put on the table
These are the candidates for your 2–3 "owned" ideas. Pick what resonates; don't try to land all five.
-
Joint benchmarking testbed
CU runs algorithms on IonQ hardware and on CURC GPU simulators side by side. Outputs are jointly published benchmarks, accuracy/scaling comparisons, and noise-model validation. Maps directly to the "owned idea" success criterion and is publishable for both sides.
-
Hosted quantum SDK environment for CU researchers
CURC stands up a curated, supported software stack — Qiskit, Cirq, Pennylane, IonQ's SDK — accessible through Open OnDemand. New researchers get to first-experiment fast. Lowers the activation energy for any CU faculty member to start collaborating with IonQ.
-
Hybrid quantum-classical reference workflow
Pick one application area where CU and IonQ both have strong interest (chemistry, optimization, ML), build a reference VQE-style or QAOA-style workflow using Alpine for classical and IonQ Cloud for quantum, document it as a template other CU researchers can adopt. Quick win that compounds.
-
Workforce-development infrastructure
If IonQ wants to expand its university-trained pipeline, CURC's user-support function can run hands-on workshops using our HPC plus their cloud, with persistent training environments. Differentiator: we've done this for general HPC for years; quantum is an extension, not a new build.
-
Secure compute pathway for federal/SLED work
Your 100+ hours of CMMC/CUI compliance work at CSU and ongoing NIST 800-171 implementation at CURC. If IonQ's federal capture involves CU researchers on CUI work, RIT could be the secure-compute partner. Frame this carefully — future-state capability, not something we operate today — but flag it if the federal angle comes up.
09 · Your asks
Questions to ask
Asking sharp questions signals you're a sophisticated partner and surfaces what IonQ actually wants from us. For each question below: what a useful answer sounds like, and how to keep the conversation moving.
What does your existing university partnership operating model look like — Maryland, Cambridge, Tennessee, Duke? What did the infrastructure side of those agreements actually cover?
Listen for
Whether the partner universities provide the classical compute and SDK hosting, or whether it's purely a cloud-credit relationship. The former is your seat at the table. The latter means workforce is your stronger angle.
How to engage
"How much of the infrastructure burden sat with the university versus IonQ? Does that pattern hold here?" — gets them describing the actual division of labor, which is the shape of your eventual role.
For your reference applications — chemistry, optimization, ML — what's the typical classical-compute footprint per run? What hardware do partners usually use?
Listen for
Specific numbers — GPU type, count, classical hours per quantum hour. Anything that lands inside Alpine's envelope (21K cores, growing GPU footprint, Blackwell + H200 incoming) is in your wheelhouse.
How to engage
"Alpine has 21,000 cores plus Blackwell and H200 capacity coming online by mid-June. What's the gap between that and what you typically expect from a partner?" — turns their answer into a sizing conversation, not an abstract one.
Would joint benchmarking — your hardware versus classical simulation on our GPUs — be useful to your science org, or is that a solved problem internally?
Listen for
A yes — especially "third-party validation" or "co-publications." That's publishable, fundable, and puts RIT in a co-author role. A no usually means they already do this internally; pivot to one of your other ideas without burning energy.
How to engage
If yes: "What metrics matter most — time-to-solution? Accuracy at fixed cost? Noise-model validation?" If no: drop it cleanly and move to the hosted SDK environment idea (Idea B).
What software stack do you expect on the classical side? Are there containerized environments you'd want us to host so CU researchers don't have to fight installs?
Listen for
Specific package names, container preferences (Apptainer is what you already do), version-pinning needs, hardware-specific builds. The more specific they get, the easier it is for you to commit to something concrete.
How to engage
"We already run Open OnDemand and Apptainer broadly. Would you want a curated bundle we maintain, or do researchers pull what they need themselves?" — surfaces whether they want a managed environment or a self-serve one, which shapes the support model.
On workforce development: what's the ideal shape of a training partnership — credit-bearing courses, certificates, hands-on workshops, hackathons?
Listen for
Scale (ten students or a thousand?) and format. Workshops and persistent training environments are squarely yours. Credit-bearing courses are faculty territory — even then, you can still offer the infrastructure layer underneath.
How to engage
"We run HPC training and onboarding regularly. Persistent training environments — sandboxes researchers can return to — are something we can stand up quickly. Would that fit the shape you're imagining?"
For the federal/SLED side: is secure compute infrastructure a partnership area you've explored, or is that handled elsewhere?
Listen for
Whether they need cleared university partners on CUI/CMMC work, or whether all federal work runs through their own facilities. The first lights up your CSU background. The second closes the door — don't push.
How to engage
If open: "I led the secure-enclave build at CSU — 100+ hours on CMMC frameworks. We're early on the secure-compute side at CURC, but it's a credible direction if the demand is real." Pre-cleared with Shelley first (see Boundaries section).
10 · Boundaries
Flag with Shelley before May 13
Two reasons: ED-level commitments shouldn't surface from you first, and you want her air cover on the boundary of what you can speak to.
- Compute allocation language. Don't promise specific Alpine Peak SUs, dedicated reservations, or Blanca-style contributor terms with an industry partner without her review. Default answer in the room: "We have allocation pathways; let's scope what makes sense after the visit."
- Industry contributor model. Whether the Blanca contributor model extends to industry partners is an open policy question. Worth confirming her current read — she may already have a stance from ESIIL conversations.
- Secure compute / CUI. If you float secure compute as a future capability, it should land as exploratory, not committed. Five minutes with Shelley on how far forward to lean.
- RIO coordination. Diana Sankovic's office is running this. Quick check whether RIT commitments need to flow through RIO formally, or whether the breakout discussion is OK as-is.
11 · Read the room
What to listen for
Signals from IonQ that change what we offer:
- Heavy emphasis on application-layer partnerships → lean into the hybrid workflow and benchmarking ideas (A & C).
- Discussion of cloud-credit programs → that's their existing channel; ask what classical-side support they expect from partners.
- Federal/SLED priorities surface → flag the CMMC background, but cautiously (E).
- Workforce / pipeline language → workforce-development infrastructure is the strongest match (D).
- Specific Boulder-based interest (Allcock, QEC group) → physical proximity argues for closer integration than a typical remote partnership.
12 · Close the loop
After the meeting
Whatever lands as the 2–3 collaboration next steps from your breakout, capture them in writing within 24 hours. Send a short follow-up to Shelley with what was discussed, what RIT was associated with, and any ED-level decisions queued up. Loop Diana for the RIO record.
Sources & methods. Brief drafted with AI assistance (Thor) using two source documents: the IonQ Partnership Visit agenda dated May 13, 2026 (CU Industry Research Partnerships, PDF), and the strawman prep brief drafted May 7, 2026 for Candace Ramsey. IonQ company context (financials, acquisitions, university partnerships, federal contracts, Boulder office) verified via web searches on May 12, 2026 using IonQ investor relations, BusinessWire, Quantum Computing Report, JILA news, and SEC filings. Stage: STRAWMAN — partnership ideas are starting points, not finalized commitments. None have been pre-cleared with Shelley or RIO. Refine based on any pre-meeting context received between now and the morning of May 13. Notes: Sections marked with ✎ sync to Netlify Blobs via /api/notes using a passphrase stored in this browser. Anything you type stays on your own infra.