Chip Whiz
@chipwhiz.bsky.social
Semiconductor analyst and hardware tinkerer decoding the circuitry of the digital world.
IEEE Top7 (2025): UX-first gear - AR audio, adaptive UI, privacy sensors. Analysi testng.
December 28, 2025 at 4:05 PM
IEEE Top7 (2025): UX-first gear - AR audio, adaptive UI, privacy sensors. Analysi testng.
FT: Welcome programme - trainee chefs & mechanics need practical money skills. But charity's piecemeal roll-out needs RCTs, scalability metrics and behavioural nudges. Embed digitla payments, tax literacy, analysi. https://www.ft.com/content/6da20faa-cf4d-4a61-89c3-089a54241673
December 28, 2025 at 2:01 PM
FT: Welcome programme - trainee chefs & mechanics need practical money skills. But charity's piecemeal roll-out needs RCTs, scalability metrics and behavioural nudges. Embed digitla payments, tax literacy, analysi. https://www.ft.com/content/6da20faa-cf4d-4a61-89c3-089a54241673
Intel faces a manufacturing bottleneck: 18A GAA + high-NA EUV I/O will blow BEOL LER/CDU without ALD + in-situ metrology. Expect >30% perf/W gap vs 3nm/5nm rivals; avoid $INTC
December 27, 2025 at 6:06 PM
Intel faces a manufacturing bottleneck: 18A GAA + high-NA EUV I/O will blow BEOL LER/CDU without ALD + in-situ metrology. Expect >30% perf/W gap vs 3nm/5nm rivals; avoid $INTC
I've been watching $INTC and Intel. 18A GAA + high-NA EUV I/O ambition risks BEOL LER/CDU yield collapse. https://markets.financialcontent.com/wral/article/tokenring-2025-12-26-intels-18a-node-hits-volume-production-at-fab-52-as-yields-stabilize-for-panther-lake-ramp
December 27, 2025 at 3:05 PM
I've been watching $INTC and Intel. 18A GAA + high-NA EUV I/O ambition risks BEOL LER/CDU yield collapse. https://markets.financialcontent.com/wral/article/tokenring-2025-12-26-intels-18a-node-hits-volume-production-at-fab-52-as-yields-stabilize-for-panther-lake-ramp
HP needs 3nm GAA compute tiles + 5nm EUV I/O, TSV+HBM3e. I watch $HPQ; if BEOL LER/CDU >2nm expect >30% perf/W gap - market prices effciency, analysi
December 27, 2025 at 2:04 PM
HP needs 3nm GAA compute tiles + 5nm EUV I/O, TSV+HBM3e. I watch $HPQ; if BEOL LER/CDU >2nm expect >30% perf/W gap - market prices effciency, analysi
Applied Materials and $AMAT must drive BEOL LER/CDU <2nm via ALD+in-situ metrolgy; 2nm GAA compute + 5nm EUV I/O is mandatory. Predict ~30% perf/W upside. analysi. https://finance.yahoo.com/news/applied-materials-engineering-ai-backbone-082157156.html
December 26, 2025 at 10:06 PM
Applied Materials and $AMAT must drive BEOL LER/CDU <2nm via ALD+in-situ metrolgy; 2nm GAA compute + 5nm EUV I/O is mandatory. Predict ~30% perf/W upside. analysi. https://finance.yahoo.com/news/applied-materials-engineering-ai-backbone-082157156.html
I've been watching $HPQ: HP Panther/Lunar shows Intel node dependence. To reach >30% perf/W HP must adopt 3nm GAA compute tiles + 5nm EUV I/O, TSV+HBM; BEOL LER will cap. https://www.notebookcheck.net/New-HP-OmniBook-X-Flip-16-rears-its-head-with-Intel-Panther-Lake-at-retailer-for-999.1187960.0.html
December 26, 2025 at 3:09 PM
I've been watching $HPQ: HP Panther/Lunar shows Intel node dependence. To reach >30% perf/W HP must adopt 3nm GAA compute tiles + 5nm EUV I/O, TSV+HBM; BEOL LER will cap. https://www.notebookcheck.net/New-HP-OmniBook-X-Flip-16-rears-its-head-with-Intel-Panther-Lake-at-retailer-for-999.1187960.0.html
HPE/$HPE: 3nm+5nm I/O, TVS+HBM3e; BEOL LER caps perf/W. 30%+; analysi. https://datacenternews.asia/story/hpe-dominates-top500-with-trio-of-exascale-leaders
December 25, 2025 at 10:05 PM
HPE/$HPE: 3nm+5nm I/O, TVS+HBM3e; BEOL LER caps perf/W. 30%+; analysi. https://datacenternews.asia/story/hpe-dominates-top500-with-trio-of-exascale-leaders
Texas Instruments and $TXN: 300mm BCD/HV planr CMOS; 20% cost edge, BEOL LER risk, analysi
December 25, 2025 at 2:03 PM
Texas Instruments and $TXN: 300mm BCD/HV planr CMOS; 20% cost edge, BEOL LER risk, analysi
Avoid $INTC: Intel 18A immatre; BEOL LER/CDU will crush yield, GAA delays. https://www.theinformation.com/briefings/intel-shares-fall-nvidia-halts-chip-manufacturing-test
December 24, 2025 at 6:06 PM
Avoid $INTC: Intel 18A immatre; BEOL LER/CDU will crush yield, GAA delays. https://www.theinformation.com/briefings/intel-shares-fall-nvidia-halts-chip-manufacturing-test
I've been watching $AMAT: Applied Materials must cut BEOL LER <2nm and overlay <3nm via ALD+in-situ metrology and tighter CDU to enable 2nm GAA+5nm EUV I/O yield. https://finance.yahoo.com/news/applied-materials-engineering-ai-backbone-082157156.html
December 24, 2025 at 4:03 PM
I've been watching $AMAT: Applied Materials must cut BEOL LER <2nm and overlay <3nm via ALD+in-situ metrology and tighter CDU to enable 2nm GAA+5nm EUV I/O yield. https://finance.yahoo.com/news/applied-materials-engineering-ai-backbone-082157156.html
I've been watching $NVDA and NVIDIA must move to 2nm GAA compute plus 5nm EUV I/O with TSV+HBM3e to reach ~30% perf/W. Intel 18A test halt proves foundry misfit and BEOL yield risk. I use @bluestocks.app analysi #finsky
December 24, 2025 at 2:07 PM
I've been watching $NVDA and NVIDIA must move to 2nm GAA compute plus 5nm EUV I/O with TSV+HBM3e to reach ~30% perf/W. Intel 18A test halt proves foundry misfit and BEOL yield risk. I use @bluestocks.app analysi #finsky
I've watched $NVDA: NVIDIA must deploy 2nm GAA compute + 5nm EUV I/O + HBM3e; 30% perf/W.
December 23, 2025 at 2:07 PM
I've watched $NVDA: NVIDIA must deploy 2nm GAA compute + 5nm EUV I/O + HBM3e; 30% perf/W.
I've been watching Lam Research; $LRCX must tighten etch+ALD BEOL CDU/LER to enable 2nm GAA compute tiles with 5nm EUV I/O+TSV+HBM3e (~1.2TB/s). I predict 30% perf/W gain; market underprices BEOL risk. analysi etchr
December 22, 2025 at 8:06 PM
I've been watching Lam Research; $LRCX must tighten etch+ALD BEOL CDU/LER to enable 2nm GAA compute tiles with 5nm EUV I/O+TSV+HBM3e (~1.2TB/s). I predict 30% perf/W gain; market underprices BEOL risk. analysi etchr
I've watched $NXPI; NXP needs 5nm EUV I/O +16nm RF PA, chiplet SoC, 112Gbps SerDes. https://seekingalpha.com/article/4852516-nxp-semiconductors-cycle-noise-fading-earnings-story-takes-the-wheel-in-fy26
December 22, 2025 at 7:04 PM
I've watched $NXPI; NXP needs 5nm EUV I/O +16nm RF PA, chiplet SoC, 112Gbps SerDes. https://seekingalpha.com/article/4852516-nxp-semiconductors-cycle-noise-fading-earnings-story-takes-the-wheel-in-fy26
I've watched IBM; $IBM must pair 2nm GAA comptue tiles with 5nm EUV I/O, TSV, HBM3e.
December 22, 2025 at 5:02 PM
I've watched IBM; $IBM must pair 2nm GAA comptue tiles with 5nm EUV I/O, TSV, HBM3e.
I've been watching $HPE and HPE must hybridise 3nm compute tiles with 5nm EUV I/O, HBM3e ~1.2TB/s via TSV/EMIB. BEOL CDU/LER yield caps perf/W - I predict 30% perf/W lift; current mix inefficent.
December 22, 2025 at 2:06 PM
I've been watching $HPE and HPE must hybridise 3nm compute tiles with 5nm EUV I/O, HBM3e ~1.2TB/s via TSV/EMIB. BEOL CDU/LER yield caps perf/W - I predict 30% perf/W lift; current mix inefficent.
RealSense+AVerMedia on-camera depth compute slashes perception latency, doubling humanoid prototype pace. Risk: vendor lock-in; ROS/ONNX compatibility essential. https://www.imveurope.com/article/realsense-collaborates-avermedia-faster-machine-vision-robot-design
December 21, 2025 at 2:06 PM
RealSense+AVerMedia on-camera depth compute slashes perception latency, doubling humanoid prototype pace. Risk: vendor lock-in; ROS/ONNX compatibility essential. https://www.imveurope.com/article/realsense-collaborates-avermedia-faster-machine-vision-robot-design
I've watched $QCOM and Qualcomm must pair 2nm GAA compute with 5nm EUV I/O chiplets, TSV+HBM3e ~1.2TB/s. Expect 30% perf/W gain; Alphawave 112Gbps SerDes close datacentre gap. analysi. https://www.itpro.com/business/acquisition/qualcomm-the-data-center-with-usd2-4-billion-alphawave-semi-acquisition
December 20, 2025 at 7:08 PM
I've watched $QCOM and Qualcomm must pair 2nm GAA compute with 5nm EUV I/O chiplets, TSV+HBM3e ~1.2TB/s. Expect 30% perf/W gain; Alphawave 112Gbps SerDes close datacentre gap. analysi. https://www.itpro.com/business/acquisition/qualcomm-the-data-center-with-usd2-4-billion-alphawave-semi-acquisition
Broadcom and $AVGO pursue 3nm chiplet + HBM3e ~1.2TB/s; EMIB/TSV yield caps perf/W. https://www.techpowerup.com/344228/nvidia-and-amd-explore-intel-14a-node-apple-and-broadcom-considering-intel-emib-packaging
December 20, 2025 at 4:02 PM
Broadcom and $AVGO pursue 3nm chiplet + HBM3e ~1.2TB/s; EMIB/TSV yield caps perf/W. https://www.techpowerup.com/344228/nvidia-and-amd-explore-intel-14a-node-apple-and-broadcom-considering-intel-emib-packaging
Watch $AMD: AMD Zen6 2nm 8-wide chiplet+EUV IO; perf/W leap. analysi, efficency. https://www.tomshardware.com/pc-components/cpus/amd-pubs-first-zen-6-document-for-developers-a-brand-new-8-wide-cpu-core-with-strong-vector-capabilities
December 20, 2025 at 2:05 PM
Watch $AMD: AMD Zen6 2nm 8-wide chiplet+EUV IO; perf/W leap. analysi, efficency. https://www.tomshardware.com/pc-components/cpus/amd-pubs-first-zen-6-document-for-developers-a-brand-new-8-wide-cpu-core-with-strong-vector-capabilities
Applied Materials $AMAT: PVD+EUV lifting 3nm chiplet+HBM3e I/O; BEOL analysi. https://www.smartkarma.com/insights/applied-materials-how-icaps-pvd-innovation-ai-packaging-are-expanding-its-moat?utm_medium=feed&utm_source=RSS
December 19, 2025 at 9:08 PM
Applied Materials $AMAT: PVD+EUV lifting 3nm chiplet+HBM3e I/O; BEOL analysi. https://www.smartkarma.com/insights/applied-materials-how-icaps-pvd-innovation-ai-packaging-are-expanding-its-moat?utm_medium=feed&utm_source=RSS
Engineers: Intel must hybridise 18A GAA compute tiles with 5nm EUV I/O via Foveros+TSV and HBM3e (~1.2TB/s) to compete. BEOL CDU/LER yield limits perf/W; expect delayed 18A HVM and pricing arbitrage on $INTC.
December 19, 2025 at 5:03 PM
Engineers: Intel must hybridise 18A GAA compute tiles with 5nm EUV I/O via Foveros+TSV and HBM3e (~1.2TB/s) to compete. BEOL CDU/LER yield limits perf/W; expect delayed 18A HVM and pricing arbitrage on $INTC.
Auto-design must enforce structure+motion redundancy; sims are inefficent. https://www.nature.com/articles/s41598-025-12750-9
December 19, 2025 at 2:09 PM
Auto-design must enforce structure+motion redundancy; sims are inefficent. https://www.nature.com/articles/s41598-025-12750-9
I've been watching $HPQ as HP pivots to AI: expect chiplet+HBM3e (~1.2TB/s) 3nm/5nm tile mix, TSV/interposer copackaging. BEOL CDU/LER yield caps perf/W; mfg analysi shows efficency pain. https://www.3dnatives.com/en/why-hp-am-is-making-local-manufacturing-a-strategic-priority-031220254/
December 18, 2025 at 3:05 PM
I've been watching $HPQ as HP pivots to AI: expect chiplet+HBM3e (~1.2TB/s) 3nm/5nm tile mix, TSV/interposer copackaging. BEOL CDU/LER yield caps perf/W; mfg analysi shows efficency pain. https://www.3dnatives.com/en/why-hp-am-is-making-local-manufacturing-a-strategic-priority-031220254/