Hot Chips 2026 (HC38) — A Symposium on High Performance Chips
by IEEE / Hot Chips Symposium Committee
Where the chips powering AI are unveiled first
About This Event
Hot Chips 2026 (HC38) is the industry's leading symposium on high-performance processors and AI accelerators, held August 23-25, 2026 at Memorial Auditorium on the Stanford campus in Palo Alto, California. Chip architects from NVIDIA, AMD, Intel, IBM, Arm and Fujitsu present the real silicon behind AI training and inference — GPU platforms, AI accelerators, memory, networking and interconnect — usually months before the products ship.
At a Glance
- Date
- August 23–25, 2026
- Location
- Memorial Auditorium, Stanford University, Palo Alto, California, United States
- Format
- In-Person
- Status
- Upcoming
- Pricing
- Paid
- Organizer
- IEEE / Hot Chips Symposium Committee
- Source
- daily-ingest
Topics & Focus Areas
Who Should Attend
- ✓CTO
- ✓VP of Engineering
- ✓Chief Architect
- ✓AI Infrastructure Leader
- ✓Chip Architect
- ✓Hardware Engineer
- ✓Data Center Leader
Why Attend
For a CTO or head of AI infrastructure, Hot Chips is the earliest reliable read on the compute you will be buying in the next 18-24 months: the accelerator, memory, power and interconnect trade-offs are presented by the architects who made them, which is a far better basis for capacity and vendor planning than vendor roadmaps. It is the cheapest way to pressure-test assumptions about AI training and inference cost per token before committing budget.
Event Features
- ✓Sunday tutorial day: memory technology and AI data requirements; RISC-V standards and adoption
- ✓Two days of keynotes and technical sessions on GPUs, CPUs, AI accelerators, FPGAs and networking
- ✓Student poster session and system/prototype demonstrations
- ✓Evening networking receptions with chip architects from the major silicon vendors
What Makes This Unique
- ★Vendors disclose real, shipping silicon at architectural depth — not marketing roadmaps
- ★38th edition; the long-standing venue where major AI accelerator and GPU architectures are first detailed publicly
- ★Sunday tutorials on memory technology for AI and on RISC-V precede the two conference days
Industries Represented
Organizer
IEEE / Hot Chips Symposium Committee
hotchips.org/Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
Subscribe