Researchers ‘recognize that AI will never replace HPC,’ says AMD executive — high-precision workloads will ‘exist for a very long time’ even as AI usage expands
- ID
- 47247
- Status
- summarized
- Published
- 17 Nov 2026, 5:30 AM
- Fetched
- 17 Nov 2026, 6:59 AM
- Provider
- Tom's Hardware
- Category
- technology
- Original URL
- https://www.tomshardware.com/tech-industry/supercomputers/researchers-recognize-that-ai-will-never-replace-hpc-says-amd-executive-high-precision-workloads-will-exist-for-a-very-long-time-even-as-ai-usage-expands
- Source URL
- https://www.tomshardware.com/feeds/all
Summary
- Score
- 3.0
- Created
- 17 Nov 2026, 7:01 AM
- Tags
- Audience
- developersai_ml_learners
What happened
Tom's Hardware published a news-analysis piece by Jake Roach on 16 November 2026 under the headline that researchers "recognize that AI will never replace HPC," quoting an AMD executive saying high-precision workloads will "exist for a very long time" even as AI usage expands. The supplied excerpt ends mid-sentence in the opening paragraph, so no named executive, benchmark, chip SKU, price, or FP64-vs-AI performance figure is actually available in the text. What is verifiable is only the framing: a vendor executive arguing that FP64-style supercomputing persists alongside AI, not that any specific workload or product changed.
Why it matters
There is no concrete takeaway to act on here: the excerpt contains zero numbers, no product names, and no measured comparison, so it gives you nothing to change about a GPU/CPU spec decision or a precision choice in your own pipeline. If you were hoping for data on whether reduced-precision AI accelerators can cover FP64 simulation work, this text does not supply it — the claim is an executive assertion, not a measurement.
Discussion angle
Ask the room where precision actually bites them: has anyone shipped something where dropping from FP64 to FP32/BF16 silently changed results, and how did they catch it? Use that to test the headline's claim with first-hand experience, since the article excerpt itself offers no supporting data.