
Tech Express | July 5, 2026 AI Chip War Escalates: Anthropic Courts Samsung, Alibaba Bans Claude
This Week in Tech: The AI chip battleground expands—Anthropic follows OpenAI into custom silicon with Samsung, while Alibaba bans Claude Code. Kling AI raises .8B, Grok 4.5 enters private beta with 1.5T parameters, Hong Kong chip trade hits records, SpaceX clears Starship static fire, Google loses €4.1B EU antitrust appeal, and Midjourney pushes Hollywood for AI transparency.
Tech Express | July 5, 2026 AI Chip War Escalates: Anthropic Courts Samsung, Alibaba Bans Claude
📰 Anthropic in Talks with Samsung for Custom 2nm AI Chip
The AI chip race is heating up. Just a week after OpenAI unveiled "Jalapeño," its first custom inference chip built with Broadcom, Anthropic is reportedly in early discussions with Samsung Electronics to manufacture its own 2nm custom AI chip. No design specifications, timeline, or server integration details have been finalized.
Anthropic told TechCrunch that a diversified hardware stack including chips from Google, Amazon, and NVIDIA remains central to its compute strategy, but declined to comment on the Samsung talks. Notably, Anthropic hired Clive Chan—the second engineer to join OpenAI's custom chip team—in early June. Chan spent two and a half years building the Broadcom-designed Jalapeño accelerator.
The move signals an accelerating "de-NVIDIA-fication" trend among top AI labs. Amazon and Google already offer custom TPUs, while Meta and Microsoft are developing in-house chips. Samsung, already a key NVIDIA manufacturing partner, would further cement its role in AI chip fabrication if it lands Anthropic's business.

📰 Alibaba Bans Claude Code Effective July 10
Alibaba Group has issued an internal directive banning all employees from using Anthropic's AI coding tool Claude Code starting July 10, 2026. The Chinese tech giant has classified all Anthropic model products as high-risk software and instructed staff to migrate to the company's in-house Qoder tool.
The immediate trigger was the revelation that Anthropic had embedded experimental code in Claude Code capable of identifying Chinese users, intended to prevent account abuse and model distillation. Despite Anthropic engineer Thariq Shihipar stating stronger mitigations have since replaced the approach, Alibaba is taking no chances with data security.
The ban underscores escalating US-China AI tensions. Anthropic already prohibits Chinese companies and their affiliates from accessing its models. Alibaba's move signals that China's tech giants are accelerating their pivot toward domestic AI alternatives.

📰 Kuaishou's Kling AI Raises $2.8 Billion at $15 Billion Valuation
Kuaishou Technology's AI video generation unit Kling AI has secured 19 billion yuan (approximately $2.8 billion) from a consortium including Alibaba, Tencent, and Baidu, according to a Hong Kong Stock Exchange filing dated July 2. The round values the AI video platform at $15 billion pre-money, with a two-month window to admit one additional investor, potentially raising total financing to 20.45 billion yuan.
Kling AI has seen explosive growth since its June 2024 launch, generating over 650 million yuan in Q1 2026 revenue. CEO Cheng Yixiao expects revenue to more than double this year. The platform serves both individual creators and enterprise clients with video generation and API services.
The deal marks a record for China's AI video sector, reflecting continued investor appetite for applied AI companies in the world's second-largest economy.

— Pulse2
📰 Grok 4.5 Enters Private Beta at SpaceX and Tesla
Elon Musk's xAI has begun private beta testing of Grok 4.5 at SpaceX and Tesla as of June 28, 2026. The model is built on a 1.5 trillion-parameter V9 foundation architecture with training data from the AI-powered code editor Cursor. Musk has characterized internal evaluations as showing performance "close to, perhaps exceeding" GPT-5.
The upgrade represents a significant leap from Grok 4.4 in both reasoning capability and code generation. xAI has simultaneously released the Grok Voice Agent Builder beta, enabling developers to build voice AI products on top of the Grok Voice Agent API, which has been available since December 2025.
Grok's evolution from an X-platform chatbot toward enterprise-grade AI infrastructure is accelerating, positioning xAI as a serious contender in the foundation model race.

📰 Hong Kong Now Handles Over Half of China's Chip Imports
Hong Kong accounted for more than 50% of China's $239 billion in semiconductor imports during the first five months of 2026, a record share up from roughly one-third a decade ago, according to Bloomberg analysis of official data. The city re-exported $124 billion worth of chips to the mainland during the period.
The former British colony's free-port status—zero import tariffs and no capital controls—combined with its strong air cargo network have made it the critical middleman in Asia's $2 trillion AI trade network. AI-related electronics now account for 57% of Hong Kong's exports, with Barclays estimating the figure could be as high as 70%.
The intermediary role cuts both ways, however. Hong Kong lacks the chip fabrication capabilities of Taiwan and South Korea, leaving it exposed to US-China tensions. As Washington tightens advanced chip curbs on China, Hong Kong has been ramping up purchases of American-made semiconductors through third countries, drawing scrutiny from both US and Taiwanese authorities.
📰 Canadian Defense Tech Firm Dominion Dynamics Lands CA$139M Series A
Ottawa-based Dominion Dynamics has raised CA$139 million (approximately $98 million) in what appears to be Canada's largest-ever defense tech Series A, led by Georgian. The company builds software, sensors, and autonomous drones for Arctic surveillance, positioning itself as "Canada's Palantir."
Founder and CEO Eliot Pence said the round reflects not only what Dominion has accomplished but also the larger strategic shift under Prime Minister Mark Carney as Canada executes an about-face on defense. With US-Canada relations remaining strained, the country is accelerating defense sovereignty.
The deal aligns with a global surge in defense tech investment—European drone company Quantum Systems simultaneously closed a $1.2 billion Series D round at an $8 billion valuation.

— BetaKit
📰 SpaceX Completes Full-Duration Starship Static Fire Ahead of Flight 13
SpaceX achieved a critical milestone for the upcoming Starship Flight 13 mission on July 2, completing a 60-second full-duration static fire test of all six Raptor engines on Ship 40 at its Starbase facility in Texas. Three sea-level and three vacuum-optimized engines burned flawlessly, clearing a major hurdle for the launch targeted for late July.
Ship 40 had previously passed two cryogenic proof tests in early May. Booster 20, the Super Heavy first stage, is undergoing independent testing. If successful, Flight 13 will be the first Starship mission to use the Booster 20 configuration for an orbital attempt.
SpaceX continues to advance the Starship program at a remarkable pace, moving closer to Musk's vision of high-cadence, fully reusable spaceflight.

— ECIKS
📰 Google Loses Final Appeal Against €4.1 Billion EU Antitrust Fine
The Court of Justice of the European Union (CJEU) on July 2 dismissed Google's final appeal against the €4.1 billion ($4.7 billion) antitrust fine originally imposed in 2018. The European Commission had ruled that Google illegally abused Android's mobile dominance to promote its Chrome browser and search service.
The ruling ends an eight-year legal battle. While the fine is a fraction of Alphabet's annual revenue, the precedent is significant—it affirms the EU's authority to penalize tech platforms for leveraging operating system control to bundle their own services.
The verdict arrives as the EU's Digital Markets Act (DMA) enters full enforcement, placing Google under fresh compliance pressure across search, advertising, and app store operations.

📰 Midjourney Pushes Hollywood to Disclose AI Usage
AI image generation powerhouse Midjourney is pushing Hollywood studios to publicly disclose the extent of AI usage in their productions. The company, valued at over $10 billion, is reportedly engaging major studios to establish transparency standards that would require film and TV productions to detail AI-generated content in credits or promotional materials.
The timing is notable. Midjourney is itself expanding beyond image generation into hardware, having recently unveiled a controversial medical scanning device prototype. Meanwhile, Hollywood writers' and actors' unions have already incorporated AI-related protections into new collective bargaining agreements.
Industry observers see the move as a strategic play to shape industry standards while positioning Midjourney's tools for deeper integration into Hollywood production pipelines.

This Week in Tech: The AI chip battleground is expanding on all fronts—Anthropic follows OpenAI into custom silicon with Samsung, while Alibaba retaliates by banning Claude Code. China's AI video sector soars with Kling AI's $2.8B raise, and Grok 4.5 enters private beta with 1.5 trillion parameters eyeing GPT-5. Hong Kong's chip trade hub role hits record highs, SpaceX's Starship clears a key test toward a late July launch, and Google's eight-year EU antitrust fight ends in defeat.
评论 (0)
更多优惠
-26%Ninja Express Chop Food Chopper — Chop, Mince, Puree at the Touch of a Button, Now $29.99 CAD (Save 25%)
Amazon
-22%Souke Sports Men's Cycling Underwear — 4D Padded Bike Liner Shorts! Limited Time $29.77 CAD (Save 22%)
Amazon
-41%Roborock Qrevo S Pro Robot Vacuum — 18500Pa, 75°C Self-Cleaning, 2026 Model! Now CAD 599.99 (Save 40%)
Amazon
-26%SocTone 10.5FT Telescoping Ladder — Triangle Stabilizers, 330lbs Capacity! Now 119.98 CAD (25% Off)
Amazon