BIG TECH

The chip wars get another sequel with Nvidia's new product

The US-China semiconductor standoff continues its game of technological whack-a-mole, with Nvidia's latest B30 chip representing the newest attempt to thread the needle between export restrictions and market access.

After the Biden administration's export controls blocked Nvidia's H20 chips in April — costing the company $5.5 billion in charges — Nvidia is reportedly developing the B30 specifically for the Chinese market. The chip uses Blackwell architecture but swaps restricted high-bandwidth memory for conventional GDDR7, staying below performance thresholds that trigger export licenses.

• Nvidia plans to produce over one million B30 units in 2025 for the Chinese market

• China represents 14% of Nvidia's revenue ($17.1 billion in fiscal 2024)

• Chinese competitor Huawei shipped 200,000 Ascend AI chips in 2024, compared to Nvidia's million H20s sold to China

• Trump administration reportedly struck a deal allowing chip sales to China for a 15% revenue cut

As we've covered before, the Trump administration's approach to semiconductor exports has been inconsistent, announcing sweeping tariffs before carving out exemptions. The administration has also been grappling with how to support domestic chipmakers, including Intel, which remains the only major American company manufacturing advanced chips on U.S. soil.

The B30 development shows how policy zigzags create costly engineering puzzles for American companies. Nvidia's repeated chip redesigns — from A800 to H800 to H20 to B30 — represent millions in R&D spending to create deliberately inferior products while Chinese companies gain time to develop domestic alternatives.

The broader context reveals an accelerating technological decoupling. Huawei's Ascend processors are gaining ground, though their software still lags behind Nvidia's CUDA ecosystem. Recent policy shifts, including treating export controls as a taxation mechanism, reflect Washington's uncomfortable reality: maintaining AI leadership while constraining China creates contradictory incentives.

Each restriction spawns engineering solutions that ultimately benefit competitors while imposing costs on American companies. Export restrictions designed to handicap Chinese AI development force American companies into expensive workarounds while accelerating Chinese efforts toward semiconductor independence.

The administration's revenue-sharing approach acknowledges this reality, effectively admitting that complete technological decoupling isn't viable.

The semiconductor industry's global nature makes isolation strategies self-defeating. Instead of endless cat-and-mouse games with chip specifications, a more sustainable approach would focus on maintaining American innovation advantages through targeted partnerships rather than blanket restrictions.

TOGETHER WITH GLEAN

Is your organization paying the “AI tax”?

In 2025, AI is rapidly becoming an integral part of business operations—but fragmented implementation is creating an “AI tax” for organizations, leading to inefficiency and increased maintenance burdens.

In The State of AI at Work in 2025, Glean explores how leading orgs are avoiding this tax by centralizing their AI capabilities on unified platforms. The whitepaper provides reference architectures and evaluation checklists to help leaders standardize tooling, control spending, and accelerate delivery without sacrificing security. 

Download the report for free to access strategic guidance on consolidating AI services across data access, security, governance, and observability—while increasing your team’s productivity. 

Download now to future-proof your org.

SOCIETY

Americans deeply worried about AI permanently displacing workers

Seven in ten Americans fear artificial intelligence will permanently put people out of work, according to a new Reuters/Ipsos poll that reveals widespread anxiety about the technology's impact on society.

The six-day survey of 4,446 U.S. adults found 71% are concerned AI will be "putting too many people out of work permanently." The poll, conducted with a 2% margin of error, shows Americans' fears extend well beyond employment concerns.

Seventy-seven percent worry AI could be weaponized to create political chaos, reflecting growing unease about deepfakes and disinformation. The concern appears grounded in recent examples like Trump's social media post featuring an AI-generated video of Obama being arrested — an event that never occurred.

Military applications drew significant opposition, with 48% saying the government should never use AI to determine military strike targets, compared to just 24% who support such use. Another 28% remain unsure about AI's role in warfare.

• 61% worry about AI's massive electricity consumption as data centers proliferate

• 67% fear people will abandon human relationships for AI companions

• Americans split on education: 36% think AI will help, 40% disagree

The survey comes as unemployment remains low at 4.2%, suggesting Americans are anticipating future disruption rather than reacting to current job losses. Major companies continue investing heavily in AI infrastructure, with recent announcements including SoftBank's $375 million purchase of a Foxconn facility in Ohio to manufacture AI servers for the Stargate project.

The poll reflects broader questions about how quickly AI adoption will reshape American work and society.

TOGETHER WITH MOMENTUM

Try This AI Prompt Cheat Sheet

A solid prompt can be the difference between confused agents and seamless workflow – but what constitutes a good one when it comes to AI? Great question.

These best-in-class examples from Momentum’s AI Prompt Library are the perfect start. 

With more than 200+ go-to-market-ready prompts specific to sales, customer success, marketing, or beyond, this library is your AI prompt cheat sheet. Use it to improve your existing prompts, uncover powerful new ones to immediately uplevel your agent output, or simply get a feel for what makes a great prompt successful. 

It’s all free, and it’s all courtesy of Momentum — plus, you’ll get access to a free Live Prompt Training.

Access their AI Prompt Library right here.

VENTURE CAPITAL

Databricks joins $100 billion club as AI drives data analytics boom

Data analytics company Databricks is raising funds at a valuation exceeding $100 billion, joining an exclusive club of private companies valued above that threshold.

The San Francisco-based company announced Monday it signed a term sheet for its Series K round, expecting to raise over $1 billion from existing investors, including Thrive Capital and Andreessen Horowitz. The valuation represents a 60% jump from its $62 billion price tag just eight months ago, when it raised $10 billion.

Databricks operates what's called a "lakehouse" platform — technology that combines data storage flexibility with structured management capabilities of traditional databases. This allows companies to store vast amounts of raw data while running sophisticated analytics and AI models on top of it.

The platform has become increasingly valuable as companies rush to build AI applications. Databricks' recent products include Agent Bricks, which builds AI agents optimized on enterprise data, and Lakebase, an operational database designed for AI workloads.

• Over 15,000 customers, including 60% of the Fortune 500, use the platform

• Annual recurring revenue hit $3.7 billion as of July 2025, up 50% year-over-year

• Recent partnerships with Microsoft, Google Cloud, Anthropic and Palantir

CEO Ali Ghodsi told CNBC that investor interest reflects "the momentum behind our AI products, which power the world's largest businesses and AI services." The funding will accelerate AI product development and fuel global expansion as companies seek to transform their data into AI-powered applications.

The round also helps Databricks delay IPO plans as it continues rapid growth in the competitive data analytics market alongside rivals like Snowflake.

LINKS

• Eight Sleep raises $100m to expand its AI-powered sleep tech

• China’s DeepSeek releases v3.1, boosting AI model’s capabilities

• Bill Gates funds $1m AI Alzheimer’s prize

• Functionize nabs $41m to speed up software testing with AI

• Mark Zuckerberg shakes up Meta’s A.I. efforts, again

• Meta rolls out AI-powered translations to creators globally, starting with English and Spanish

• Why Andreessen and Bain backed a self-described ‘Zendesk killer’

• Qwen-Image Edit gives Photoshop a run for its money

• Tracking AI: Tracks and ranks the IQ test of the top AI models

• ElevenLabs Music API: First music API trained on licensed data, commercial-ready

• Qwen Image Edit: Make precise modifications to any image ($0.03 per edit)

• Notion AI: Can use Notion across multi-step tasks like editing multiple pages or updating entire database—like a power user, built in

• Camunda: Senior Field Marketing Manager

• Grammarly: Staff Growth Designer

• Dataiku: Senior Data Scientist

• Lambda: Senior Site Reliability Engineer

GAMES

POLL RESULTS

Should AI platforms disclose system prompts?

• Yes, full transparency (42%)

• Partial/behind a gate (21%)

• No, security risk (17%)

• Not sure (20%)