From the Research Universities edition of November 3, 2026
UW's Tillicum Supercomputer Marks First Anniversary
University of Washington Information Technology (UWIT) · University research computing (shared GPU supercomputer) · United States (Washington)
- Publisher
- UW Information Technology
- Original publication
- 2 November 2026 (UWIT news)
- Source retrieved
- 2026-11-04
What happened
UW reports that Tillicum, a $10 million UWIT GPU investment, supported 336,606 jobs and 1,018,768 GPU hours in its first year for 940 users from 159 research groups, 64 departments, 17 colleges and schools, and all three campuses. Students can access it through Student Technology Fee-funded credits. UW says shared GPUs reduce the need for groups to buy their own hardware.
Why it matters
Directly relevant to public research universities deciding whether to centralize GPU investment: a first-year usage baseline from a peer institution. Adoption breadth is shown; research outcomes and cost per GPU hour are not.
Evidence and measured results
Institutional self-report of usage counts and rankings (No. 7 U.S. higher ed on Green500, 10th on TOP500 per the post). No utilization rate, queue times, cost recovery, or research outcomes reported. Partner-sponsored events (Dell, NVIDIA) noted. Classified as a self-reported institutional claim.
Limitations and uncertainty
Self-reported; no denominator for capacity, so utilization cannot be computed; rankings not independently checked here; partner-sponsored events.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-11-03; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Departments buy and run their own GPUs, leaving capacity fragmented and access uneven for smaller groups and students.
- Stakeholders
- CIO
- Research computing director
- Vice President for Research
- Faculty governance
- Student technology fee committee
- Discovery questions
- How many departments buy GPUs independently today?
- What utilization and wait-time data exist?
- How would student access be funded?
- What cost-recovery model is acceptable?
- Value hypothesis
- A central shared cluster may broaden access across departments, as UW reports, though cost effects remain unmeasured.
- Potential engagement
- A demand and governance assessment for a shared GPU service, using UW's published counts as a reference point.
- Unsupported claims
- Cost savings from consolidation
- Research productivity gains
- Utilization levels
Pre-sales engineering
Role takeaway
- Fit
- Fits universities with scattered departmental GPU demand.
- Architecture and integration
- Central cluster, scheduler fair-share, campus identity, research storage, optional cloud burst.
- Prerequisites
- Demand survey
- Funding/cost-recovery model
- Data classification policy
- Constraints
- No utilization or queue data in source
- Rankings self-reported
- Hardware pricing volatility
- Security
- Tenant separation, controlled-data enclaves where needed, and access reviews for student accounts.
- Proof of value
- Measure own GPU demand and wait times
- Pilot shared allocation with several departments
- Track utilization monthly
Delivery
Role takeaway
- Work
- Design allocation policy and onboarding; Launch usage reporting and student credits
- Dependencies
- Central IT funding
- Data center power/cooling
- Scheduler and identity integration
- Ownership
- Central IT research computing.
- Skills and adoption
- User training and office hours for new GPU users across disciplines.
- Governance checkpoints
- Fair-share policy approval
- Data classification review
- Annual usage and outcome review
- Acceptance criteria
- Proposed: utilization and queue-wait published quarterly
- Proposed: number of departments served tracked
- Proposed: student credit uptake reported
- Risks
- Oversubscription
- Unfunded refresh cycle
- Equity gaps between groups
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Centrally funded shared GPU cluster serving multiple campuses, with student access through fee-funded credits and cloud providers participating in a symposium, suggesting hybrid on-prem/cloud discussion.
Governance
Who approves, reviews and stays accountable for outcomes?
Allocation across 159 groups and student access implies governance for fair share and credits; not described in the post.
Security and privacy
What data, permissions and controls need testing?
Shared clusters require tenant separation and data classification policies; not described.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Student access through fee-funded credits and cross-campus users suggests broader workforce development, though outcomes are not reported.
Procurement
What should contracts, pricing and exit terms secure?
Centralized purchasing can replace scattered departmental buys; post does not provide price or contract details.
Operating model
Which teams own the service once it runs?
Central IT-run shared service model with student credit program.
Publication history
- 2026-11-03Research Universities · Issue 565 resources
Stable resource ID: uw-tillicum-gpu-supercomputer-first-year-usage-20261102