Author
@harukaze5719 harukaze5719
grade Bx.com/harukaze5719 ↗
Korean-language semiconductor and AI-compute researcher relaying deep chip supply-chain analysis
Summarizes and synthesizes semiconductor research, benchmark
trader score
+0.29
hit rate
50%
mean α
+0.23%
signals 14d
39
Grade = how their written analysis reads (A best). Trader score = how their last-20 timestamped calls performed vs SPY.
Their picks, scored
Across their last 20 scored bets: 50% hit rate, +0.23% mean alpha, trader score +0.29. Their last-14d mentions, direction-adjusted, have moved +0.5% since posting (mean over 17 mentions with price data).
Recent signals15receipts included
| date (PT) | ticker | author | sent | what they said | since then | receipt |
|---|---|---|---|---|---|---|
| 2026-08-27 | · | @harukaze5719 | · | Guosen Securities expects HBM and enterprise-SSD supply tightness through 2027 before easing in 2028. | · | tweet ↗ |
| 2026-08-26 | NVDA | @harukaze5719 | -0.10 | The Information reports Nvidia agreed to buy Hugging Face for $12.9 billion. | +3.8% | tweet ↗ |
| 2026-08-26 | NVDA | @harukaze5719 | +0.20 | Urges Nvidia's CEO to release more NVFP4 GGUF models if the Hugging Face acquisition occurs. | +3.8% | tweet ↗ |
| 2026-08-26 | · | @harukaze5719 | · | Lists explicit Samsung P9 and P7 prices across four storage capacities. | · | tweet ↗ |
| 2026-08-25 | · | @harukaze5719 | · | Questions the validity of stated FLOPS calculations based on precision scaling ratios. | · | tweet ↗ |
| 2026-08-25 | · | @harukaze5719 | · | Cites Gartner predicting AI inference cost per agentic workflow will rise more than fivefold by 2028. | · | tweet ↗ |
| 2026-08-23 | SSNLF | @harukaze5719 | -0.35 | Mocks the apparent omission of Samsung packaging technology from referenced material. | · | tweet ↗ |
| 2026-08-23 | · | @harukaze5719 | · | Warns that corporate PR slides should only be half trusted, without naming an issuer. | · | tweet ↗ |
| 2026-08-23 | SSNLF | @harukaze5719 | -0.25 | Questions whether Samsung's HPB cooling reflects inefficient DRAM and whether rivals have alternatives. | · | tweet ↗ |
| 2026-08-23 | SKHY | @harukaze5719 | +0.05 | Questions whether Samsung's HPB cooling reflects inefficient DRAM and whether rivals have alternatives. | · | tweet ↗ |
| 2026-08-23 | MU | @harukaze5719 | +0.05 | Questions whether Samsung's HPB cooling reflects inefficient DRAM and whether rivals have alternatives. | -3.5% | tweet ↗ |
| 2026-08-23 | AMD | @harukaze5719 | +0.10 | Uses AMD 3D V-Cache as evidence that stacked SRAM can help, then questions Cerebras feasibility. | -1.6% | tweet ↗ |
| 2026-08-23 | · | @harukaze5719 | · | Explains how offloading the memory controller to a base die could free area for compute and SRAM. | · | tweet ↗ |
| 2026-08-23 | · | @harukaze5719 | · | Questions whether a DRAM-covered GPU can cool a stated 1,200-1,600W power draw. | · | tweet ↗ |
| 2026-08-23 | · | @harukaze5719 | · | Suggests high-capacity 128GB LPDDR may be preferable to faster 48GB GDDR for some workloads. | · | tweet ↗ |
Grade is our human read-worthiness rating; trader score is a rolling 20-bet hit-rate/alpha composite — different things, often disagreeing. “Since then” is direction-unaware in the table; the summary line above adjusts for which way they leaned.