- AI KATANA
- Posts
- Nvidia Places Huge H20 Order
In partnership with
Create How-to Videos in Seconds with AI
Stop wasting time on repetitive explanations. Guidde’s AI creates stunning video guides in seconds—11x faster.
Turn boring docs into visual masterpieces
Save hours with AI-powered automation
Share or embed your guide anywhere
How it works: Click capture on the browser extension, and Guidde auto-generates step-by-step video guides with visuals, voiceover, and a call to action.
Hello!
Over the last 24 hours the AI landscape has ricocheted from Wall Street trading desks to community tech hubs in Accra. Meta’s audacious “super-intelligence” push is rattling investors even as Google funnels $37 million into African research and talent. Taiwan’s chip fabs are back at full throttle to satisfy a surprise 300-thousand-GPU order, while Tokyo’s test-gear champion is hiking guidance on the same demand boom. Citi is tearing up the sell-side playbook by rating still-private AI darlings, and Pittsburgh has rolled out an “AI Avenue” plan to muscle into the robotics big-leagues. Put together, the day’s news shows capital and policy makers scrambling to keep pace with hardware hunger, regional ambitions and a widening race for talent.
Sliced just for you:
💾 Nvidia Places Huge H20 Order
🇯🇵 Advantest Ups Forecast on AI Test Boom
🧠 Meta’s Super-Intelligence Spending Spree
🌍 Google Pours $37 M into Africa AI
🏙️ Pittsburgh Sets Sights on AI Capital
📊 Citi to Rate Private AI Stars
Nvidia has reversed course on chip rationing and asked TSMC to fabricate another 300,000 H20 accelerators, signalling that freshly-lifted U.S. export curbs have unleashed a torrent of pent-up mainland demand. The order enlarges an existing stockpile of roughly 650,000 units and underscores the H20’s role as a lower-power workaround that still meets China’s training appetite. Sources say the new wafers will take nine months to reach board-level assembly; in the interim, buyers must submit volume forecasts and compliance paperwork while Washington finalises export licences. Lawmakers critical of the policy shift warn the move could erode America’s lead just as Blackwell-class GPUs ramp up elsewhere.
Japan’s Advantest lifted its full-year operating-profit outlook by 24 percent to ¥300 billion after hyperscalers and chipmakers pulled forward orders for its V93000 tester line. Management told analysts that AI-centric HBM packages and 2.5-D substrates require 40 percent more test hours per wafer, driving both volume and mix. The company is adding capacity in Gunma Prefecture and expects margins to widen as software-driven “Edge AI” options roll out next quarter. Investors cheered the news, sending shares up 5 percent as traders bet on a protracted validation cycle for next-gen neural accelerators.
Meta’s quest to outpace rival labs has ballooned into a cash-burning sprint: the firm is now on track to post its slowest profit growth in two years as it pours billions into data-centres and a newly formed “Superintelligence Lab.” Insiders say Mark Zuckerberg has green-lit hundreds of elite hires, lured Scale AI’s CEO with a $14.3 billion stake and is considering additional capex hikes despite ad-market headwinds. Analysts remain split—some praise open-sourcing plans around Llama while others flag uncertain timelines and rising political risk from tariff-hit advertisers.
Google has unveiled a $37 million package to deepen its African AI footprint, highlighted by an Accra-based AI Community Center and a $25 million hunger-forecasting initiative that pairs local NGOs with Google.org researchers. Additional grants support Masakhane’s multilingual datasets and 100,000 Career-Certificate scholarships covering “AI Essentials.” Executives say the strategy blends philanthropy with long-term talent cultivation, positioning the continent as both market and lab for low-resource language models.
Pennsylvania’s new “AI Strike Force” is pitching Pittsburgh as the world hub for “physical AI,” leveraging Carnegie Mellon’s robotics pedigree and a $90 billion wave of data-centre pledges. Officials are marketing an “AI Avenue” corridor, a $50,000 startup contest and relaxed zoning to lure hardware, defence-tech and autonomous-mobility firms. Critics worry about grid strain and fossil-fuel reliance, but backers argue the region’s steel roots and nuclear-adjacent power deals give it a pragmatic edge in the resource-heavy AI era.
Breaking with tradition, Citi will start publishing event-driven research on roughly 100 high-valuation private companies—many in AI—mirroring similar moves by rival banks. Reports will track product launches, customer traction and competitive threats but omit price targets, allowing clients earlier insight into pre-IPO heavyweights like OpenAI and SpaceX. The shift reflects investors’ demand for guidance in a market where private valuations can eclipse S&P 500 constituents, blurring public-private boundaries.
🛠️ AI tools updates
Slack’s July change-log introduces an analytics dashboard that breaks down how often summarisation, translation and huddle-note features are used across paid workspaces. Admins can now toggle individual AI tools, set retention rules for AI-generated notes and add Confluence or Teams as enterprise-search sources—part of a broader push to court security-conscious customers ahead of Slack’s fall pricing revamp. Read more
Prop-tech firm Keyway has launched an AI platform that ingests messy rent-rolls and trailing-twelve-month statements, normalises them against chart-of-accounts templates and pipes audit-ready data into CRE dashboards. Early adopters report deal timelines cut by 90 percent and cost savings near 50 percent, positioning the product as a finance-first operating system for mid-market real-estate portfolios.
💵 Venture Capital updates
San Francisco-based Julius, which bills itself as a “data-analyst in your pocket,” secured $10 million led by Bessemer. The platform writes millions of lines of code daily to clean data, run predictive models and output visualisations from natural-language prompts. Funds will scale enterprise connectors and expand the freemium user base that already tops two million. Read more
Bengaluru startup Drizz raised a $2.7 million seed round led by Stellaris Venture Partners to apply vision-AI to mobile-app testing. Its system records session videos, detects UI regressions and autogenerates bug tickets, slashing manual QA cycles for fast-shipping teams. Capital will fund hiring and expansion into iOS and gaming-engine pipelines.
🫡 Meme of the day
⭐️ Generative AI image of the day
Before you go, check out ChatGPT Outsmarts “I’m Not a Robot”
