Anthropic wants to build its own AI chip

Sunday, July 5, 2026

Anthropic is reaching for the one layer of the stack it doesn't yet control: silicon. Meanwhile the geopolitical lines around AI hardened this weekend, from Alibaba's ban on Claude Code to Modi's new fab.

Anthropic Joins the AI Chip Race, Courting Samsung to Build Its Own Silicon

Anthropic has begun early-stage work on a custom AI chip and held preliminary talks with Samsung about manufacturing it, per The Information. It's a bid to cut dependence on Nvidia and control costs as inference volumes explode — following the vertical-integration playbook Google and Amazon already run.

Read the brief →

Alibaba Bans Claude Code as Anthropic Cracks Down on China Access

Alibaba is barring employees from using Anthropic's Claude Code starting July 10, as Anthropic tightens restrictions on China access. The move underscores how AI tooling is becoming a front line in US-China tech decoupling, pushing Chinese firms toward domestic alternatives.

Read the brief →

Palantir and Nvidia Team Up to Put Sovereign AI Behind the Firewall

Palantir and Nvidia are deepening their partnership to deliver 'sovereign AI' that runs behind government firewalls for national-security work. It signals a fast-growing market for on-premises, air-gapped AI that governments can deploy without ceding data to the cloud.

Read the brief →

Nvidia's Stock Stalls in 2026, Leaving Investors Weighing Their Next Move

After years of blistering gains, Nvidia's stock is nearly flat for 2026, per The Motley Fool. The pause hints that investors are recalibrating expectations for the AI trade's biggest single beneficiary.

Read the brief →

Modi Opens Sanand Chip Plant, Bets India's Youth Will Power AI With Homegrown Silicon

Modi inaugurated a semiconductor plant at Sanand in Gujarat, framing it as a bet on homegrown silicon and Indian talent. It's another data point in the global scramble to localize chip supply chains beyond Taiwan.

Read the brief →

GLM-5.2 Arrives as AI Rivals Scramble to Catch Up

A new large language model, GLM-5.2, has launched amid an intensifying race among AI developers to close the capability gap. Details are thin, but the steady cadence of frontier-adjacent releases keeps pressure on incumbent labs.

Read the brief →

The Biggest AI Winners May Not Be AI Stocks, Research Suggests

New research argues the biggest financial winners from AI may not be the AI stocks themselves but the companies that deploy it to widen margins. For investors, it reframes where the durable value in the AI boom actually accrues.

Read the brief →

NHS App to Use AI to Point Patients to the Right Service

The NHS app will use AI to route patients to the most appropriate service for their needs, per the BBC. It's a concrete example of AI moving into high-stakes public-sector triage at national scale.

Read the brief →

The AI Briefing The day's most important AI, chip & tech stories — one email each morning. Free.