Sravan Kundojjala

AMAT raised CY26 revenue growth guidance to 40% from earlier >>30%. Started the year with 20%+ growth expectation. Even 40% growth doesn't guarantee WFE outperformance. But AMAT has broad presence across tool chain and a big beneficiary of AI/DC trend. Expects DC wafer volume to be 2x vs smartphone wafer volume in two years after reaching parity this year.

TSMC Aug 2026 - $16.079B in USD, +10.8% M/M and +44.3% Y/Y; First $16B month; Strongest August M/M since 2022 - Sixth consecutive monthly ATH, seventh in eight months in ‘26 - 32 consecutive months of Y/Y growth; August annualized run-rate is $193B - The first 8 months of 26 $106.7B, +37.1%; Sep-Dec needs $64.7B (+45% Y/Y) to hit at least 40% y/y growth in CY26 -+$1.566B, the second-largest ATH increment ever and 2x of June's +$769M - Jul+Aug $30.6B is already 92.5% of the entire 3Q25 ($33.1B) and > all of 2Q25 ($30.2B)

Broadcom 3QF26 - FY27 AI $115B, (raised from >$100B); supply secured - FY28 AI doubling to $230B; supply secured; demand > supply - 3Q XPU up >3.5x y/y = 73% of AI; networking 27% (2Q was 40%, Hock had called 30% the normal share); AI networking up >2.5x y/y - 3Q AI revenue up 55% q/q and 221% y/y to $16.7B and 4Q AI revenue will grow 30% q/q to $21.7B - 4Q both XPU and AI networking to triple y/y - networking to grow just as fast as XPUs over the next few years - Content per GW $20-30B, stable across generations since power per chip rises with ASP

- demand doubling based customer forecasts; supply bottleneck through end of FY28 - Memory costs higher than anticipated and weigh on margins; 3Q GM guided down to 74% (+/- 50bp), 4Q bottoms at 71-72%, FY28 settles at 72-73%; price increases from 1QF28 - DC revenue $89.0B, up 117% y/y and 18% q/q; Blackwell Ultra remains the vast majority of revenue - Hyperscale $48.7B, up 102% y/y and 13% q/q; ACIE (AI clouds, industrial, enterprise) $40.3B, up 138% y/y and 25% q/q; ACIE is now 45% of DC; expects non-hyperscale to be roughly half the DC business - Edge Computing $7.2B, up 27% y/y; Blackwell workstations strong while consumer PC affected by memory

- Revenue per GW has gone from $18B (Hopper) to $25B (Blackwell) to $40B (VR); $60B cost per GW for DC - Grace TTM revenue >$5B; server CPU demand reiterated at $20B; FY28 server CPU revenue to >2x y/y - Hopper (H20/H200) shipments to China were <1% of DC revenue; dilutive to GM - China revenue (customer HQ basis) was $7.9B, up 98% y/y and 73% q/q, but this is mostly gaming and workstation - About $50B invested in frontier AI labs to date; the 10-Q shows $99B of equity investments held plus $25B committed; Neoclouds exit 2026 with 8GW installed vs about 3GW at end-2025

Nvidia QJul26 - First-ever year-ahead guide; FY28 revenue up 70% y/y; supply-constrained - Supply and capacity commitments jumped from $119B to $279B in one quarter - Inventory $31.6B, up 22% q/q, with raw materials tripling in first 6 months to $11.3B - VR production shipments started in Aug; guided to 20% of 3Q DC revenue; expected to be the fastest ramp in company history - Revenue $96.2B, up 106% y/y and 18% q/q, 4th straight quarter of y/y growth; beat the $91B guide by 5.7%; GAAP gross margin 75.0%, flat q/q - 3Q revenue $108B (+/- 2%), up 12% q/q and about 89% y/y; driven by ACIE; hyperscale re-acceleration in Q4 and FY28 (VR) - Networking had another record quarter, up 18% q/q to $17.5B

Qualcomm sees HBC (High Bandwidth Compute) coming to smartphones, XR and cars too in the future. HBC Gen 1 commercialization in 2027 and Gen 2 in 2028. Sees lots of interest from hyperscalers. Adopts system integratrator model with compute die from Qualcomm, TSMC manufacturing, DRAM from stacking from someone else. With HBC, Qualcomm claims 6x the bandwidth per watt of HBM designs.

- CPU demand off the charts; units growing again; double-digit unit growth; core count growing and ASP per core stabilizing - Ireland fab output to double next year; Fab 62 will start to be equipped from next year; Oregon to transition from just pilot line pilot plus high volume manufacturing fab for 14A - Arizona will be the center for 18A; Ohio Mod 1 key focus right now; sees Intel 3 as a good node for external base die customers too - Sacrificing client share for server CPU capacity as server capacity has become key for market share; sees desktop catch up with Nova Lake

Intel at DB conference - Upbeat on foundry and process node developments; 18A yields tracking ahead of targets; calls14A defect density curve the best since 22nm - Strong internal demand for 14A; increased confidence on 14A external engagements; 0.9 PDK in Oct - EMIB-T ramp from back half of ’27; multiple billions of dollars of EMIB per customer; lower capex and higher ROIC vs wafer business - Sees 40% GM and 30% OPM for advanced packaging; wafer business will also have similar margins and depends on 14A execution as the best in this business is now close to 70%

Foundry industry was nearly $220B annual run-rate in 2Q26 and $245B+ in 3Q26. Foundry utilization inflected to 90% from 85% as non-China mature fabs also caught up. TSMC took 90% of the $30.1B gross profit pool at 73.8% revenue share, but ex-TSMC growth doubling to +14% in 2026 from +3% in 2025 on the pricing turn.