What Huawei Is Building Today Is Almost Impossible to Believe

What Huawei Is Building Today Is Almost Impossible to Believe

Title history
  • What Huawei Is Building Today Is Almost Impossible to Believe — observed on 20261031-2309 UTC (YouTube)
🎙 Tech Meets East 👥 17K 📅 October 14, 2026 ⏱ 29 min 👁 75K 📄 news review 🧭 2026-10-31
Available in: English (current) Français

Keywords

HuaweichipslithographyAscendHarmonyOS

Summary

The video examines Huawei’s massive effort to build a complete technology supply chain indigenously, covering everything from lithography machines to operating systems and cars. It opens with the September 2026 announcement of the Atlas 960E superpod containing 4,096 processors and a petabyte of memory. The narrative traces Huawei’s response to US sanctions since 2019, highlighting its 121.4 billion yuan R&D spending in H1 2026. It details progress in lithography via a coordinated domestic industry, logic folding technology for chips, and the Kirin 9050 Pro with claimed performance gains. Memory constraints and price increases are addressed. The video also covers the Ascend AI chip roadmap, the unified bus interconnect, CANN software ecosystem, HarmonyOS adoption, and the Hima car alliance. It critically notes limitations: China’s lithography is roughly two decades behind, chips trail Apple’s newest, supply shortages persist, and profit fell 36% in H1 2026. The conclusion suggests Huawei doesn’t need to win every floor; it needs every floor to exist, and the flywheel effect of domestic substitution could accelerate progress.

169 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides substantial original synthesis of scattered news reports into a coherent narrative about Huawei’s vertical integration strategy. It argues that Huawei’s approach—building every layer itself—is unprecedented in the tech industry. The argumentation is structured as a physical metaphor of a skyscraper, effectively illustrating the interdependencies. It uses concrete data (revenue, R&D spending, chip performance figures, market share) from cited sources. The reasoning is balanced: it acknowledges Huawei’s achievements while also presenting counterpoints (e.g., leading one chip to start with, but still behind Apple’s A20 Pro; supply constraints; financial losses). The video avoids hyperbole and instead presents a realistic assessment, making the argument credible.

Scientific Rigor, Source Quality, Title Accuracy

Rigueur scientifique is high for a YouTube analysis: the video cites major financial and trade press (Financial Times, Reuters, Bloomberg, Nikkei) and analyst firms (Bernstein, SemiAnalysis, UBS). It attributes specific claims to named individuals and publications. However, it relies on second-hand reports and Huawei’s own announcements, which are not independently verified. The title matches the content: it explores the audacity of Huawei’s ambition with a critical eye. No comment analysis was possible as no comments were provided.

197 words

Title / Content Match

The title is attention-grabbing and accurately reflects the video's core thesis that Huawei is building an unprecedented vertically integrated technology stack; the content supports this claim with evidence and nuance.

Quality & Reliability

8/10

The video presents a detailed, well-structured overview of Huawei's technological self-reliance efforts, citing multiple reputable sources (Financial Times, Reuters, Bloomberg, Bernstein, SemiAnalysis, UBS, Nikkei). It balances claims with critical caveats on performance gaps and financial strain, but relies heavily on secondary reporting and Huawei's own statements without independent verification.

Key Moments

Cited Sources

  • Financial Times — Articles on Huawei's lithography coordination, satellite images of chip plants, and analyst quotes.
  • Reuters — Reports on Ascend accelerator card pricing, memory costs, and SMIC and Huawei production lines.
  • Bloomberg — Interviews with Huawei executives on chip market share and supply constraints.
  • Bernstein Research — Analyst reports on Huawei chip performance and process node analysis.
  • SemiAnalysis — Comment by Dylan Patel on Huawei's unprecedented effort to develop an AI supply chain.
  • Nikkei — Report on Huawei's Liangzhu Lake R&D center and lithography research mission.
  • UBS — Analysis of Chinese lithography progress compared to ASML's 2004 level.
  • Geekernel — Independent testing of Kirin 9050 Pro chip performance and transistor density.
  • The Register — Assessment of Ascend 960 DT computing power and superpod scalability.

Concurring Sources

  • Financial Times — Aligned with the video's claims on Huawei's role in coordinating lithography efforts.
  • Reuters — Confirms pricing and supply constraints mentioned in the video.

Dissenting Sources

  • The Register — Suggests the superpod's performance is lower than Huawei claims, and the million-processor cluster is theoretical.
  • Bernstein Research — Notes Kirin 9050 Pro still trails Apple's A20 Pro by ~30%, which the video acknowledges but downplays.

Contribution & Novelties

The video offers a unique consolidated view of Huawei’s multi-layered strategy, synthesizing recent news into a coherent ‘skyscraper’ model. It provides specific performance metrics and financial data that are often scattered, and it frames Huawei’s actions as a deliberate response to sanctions with a flywheel effect. The video’s original contribution is highlighting how the pressure of export controls forces incremental improvement that could accelerate over time.

Pour aller plus loin :

  • Huawei — Background on the company and its evolution.
  • Deep ultraviolet lithography — Explanation of the technology Huawei is helping develop domestically.
  • High Bandwidth Memory (HBM) — Context for memory challenges in AI chips.
  • Chiplet architecture — Related approach to boosting performance without shrinking transistors.
  • Open source AI ecosystems — Relevance of PyTorch support for Ascend.

127 words

Radar Profile

The radar profile shows high scores in quantity and quality of information, reflecting extensive data and credible sources. Technical level is high, indicating good depth for an informed audience. Reliability is slightly lower due to reliance on unverified claims and secondary reporting, resulting in a balanced but not flawless profile.

Reliability 7/10