AI Tech Safar — Latest AI News, Tools & Trends

By Imran Khan (AI Tech Safar) The Pentagon calls Alibaba and Baidu "Chinese military-aligned companies." The Commerce Department has placed Z.ai on its Entity List, severely restricting its access to US technology. Both are official positions of the Trump administration. And both companies' AI models are available right now on an AI marketplace that a Trump-backed crypto firm is actively promoting — with the Trump family earning revenue from every transaction. A Reuters review found that 43 of the 90 AI models available on the platform, WorldClaw, were built by Alibaba, Baidu, Z.ai, DeepSeek, and other Chinese firms the same administration has flagged as national security risks. None of this appears to be illegal. It is, however, a genuinely awkward look for an administration that has staked much of its tech policy on being tough on Chinese AI. How the Pieces Connect Entity Role The Conflict World Liberty Financial Trump-backed crypto firm, issues USD1 a...

By Imran Khan (AI Tech Safar) Nvidia never set out to become one of SpaceX's biggest shareholders. It happened almost by accident — the company wrote a $10 billion check to Elon Musk's AI lab, xAI, back in January, and then SpaceX turned around and acquired xAI for $1.25 trillion in February. Overnight, a chip-relationship investment quietly turned into 122.8 million shares of one of the most closely watched IPOs of the year. When SpaceX went public in June, that stake had a real price tag attached: roughly $21 billion. It's the same story we've been tracking since Jensen Huang stood in front of six Wall Street firms and pitched GPUs as an "investable asset class" — except this time, the position wasn't even planned as an investment. Nvidia's Biggest Disclosed Equity Positions Holding Value at Q2 Close (Jun 30) Value as of Aug 14 Change Intel $30 billion ~$22 billion -27% SpaceX $21 billion ~$17.2 billion -18% CoreWeave —...

By Imran Khan (AI Tech Safar) Google's AI models were downloaded 418 million times in 2026. Meta's hit 227 million. Alibaba says its Qwen family was downloaded more than 3 billion times — in six months. That's not a close race; that's Alibaba claiming a number that dwarfs its nearest rival by roughly seven times, and it just quietly became the most-downloaded AI model family on Earth without most of the world noticing. There's a catch worth flagging before you accept that number at face value: it didn't come from an independent audit. It came from Alibaba itself, citing a Hugging Face report — the same report that verified Google's and Meta's much smaller numbers directly. The Numbers, Side by Side Company Reported Downloads Source of the Number Alibaba (Qwen) 3 billion+ (past 6 months) Alibaba's own statement, citing Hugging Face data Google 418 million (2026 total) Hugging Face's own state-of-open-models report Meta ...

By Imran Khan (AI Tech Safar) Three days before this announcement, Anthropic made it harder to hide that you used AI. Now Google is making it easier. On August 14, 2026, Google confirmed that users can now switch off the visible watermark on images, videos, and songs generated with Gemini — the same week Claude users were publicly furious that Anthropic's new watermarks would expose them for using AI at work and in school. Same industry, same eight-day window, two AI giants moving in exactly opposite directions. Google's version comes with a catch that matters more than the headline: turning off the visible mark doesn't touch the invisible one. What's Changing: Visible vs. Invisible Watermarks Watermark Type Can You Turn It Off? What It Does Visible Watermark Yes, as of this update The small on-image logo/mark you can see, removable via Settings > Media Watermark SynthID (Invisible) No, stays on regardless Embedded signal detectable via Gemini...

By Imran Khan (AI Tech Safar) Google just rewrote its own AI org chart in a single memo — and buried inside it is one of the biggest exits in Silicon Valley history. Demis Hassabis is stepping back from running Google DeepMind day-to-day. Jeff Dean, employee number 30 and the architect behind half the infrastructure the modern internet runs on, is leaving Google entirely after 27 years. And he's not leaving quietly — he's taking three of Google's most important AI researchers with him to build a rival AI research company, backed by Google's own money. This isn't a routine reshuffle. It's Google simultaneously trying to win the AGI race and manage the exit of the person many consider its most technically important employee ever. Table of Contents What Just Changed at Google DeepMind? Why Did Demis Hassabis Step Back From Running DeepMind? Who Is Koray Kavukcuoglu, DeepMind's New Chief? Why Is Jeff Dean Leaving Google After 27 Years? What Is Discov...

By Imran Khan (AI Tech Safar) Four days. That's all it took for Wall Street to completely flip its mood on AI stocks. Just weeks after we covered Nvidia trading at its cheapest valuation since 2019 and AI-linked stocks shedding over $1 trillion in value, the same stocks have staged one of the sharpest reversals in recent memory — the Nasdaq 100 surged 9.3% in four trading days, adding roughly $3.5 trillion back in investor value, and a Bank of America sentiment gauge just hit its most bullish reading since 2021. If you're feeling whiplash reading that after our last AI stocks piece, you're not alone. This is exactly the kind of violent mood swing that's come to define AI investing in 2026 — and understanding why it happened matters more than just watching the number bounce around. Table of Contents What Just Happened to AI Stocks? Why Did AI Stocks Suddenly Rebound? Selloff vs. Rebound: The Numbers Side by Side Are AI Stocks Going to Keep Rebounding? What Ri...

By Imran Khan (AI Tech Safar) Six years ago, Firmus Technologies was a Bitcoin mining operation in Tasmania. Today, it just closed a $2 billion funding round with Nvidia as its lead backer, at a valuation of roughly $15.5 billion — nearly double what the company was worth just three months earlier. If you've never heard of Firmus, that's the point of this story: while everyone watches OpenAI, Anthropic, and the usual Big Tech names, Nvidia has quietly been building an entirely separate empire of AI infrastructure bets across Asia-Pacific, and this deal is its biggest one yet in the region. Table of Contents What Just Happened With Nvidia and Firmus? Who Is Firmus, and Why Does Nvidia Care? Firmus's Funding Timeline: How It Got Here What Company Did Nvidia Just Invest $2 Billion In? Why Does This Deal Matter Beyond Australia? What Happens Next: The ASX IPO Frequently Asked Questions Quick Summary & Key Takeaways The Deal: Firmus Technologies has drawn $...

By Imran Khan (AI Tech Safar) Twice in three weeks, a Chinese AI lab has walked up to Anthropic's front door and knocked. First it was Moonshot's Kimi K3. Now Alibaba has released Qwen3.8-Max — its largest model ever, built on 2.4 trillion parameters — and the company isn't being shy about the comparison it wants people to make: benchmark scores it claims land right alongside, and sometimes ahead of, Anthropic's Claude Fable 5, the model widely considered the current global frontier. Investors reacted before most people even finished reading the announcement. Alibaba's stock jumped 4.5% in premarket US trading and 7% in Hong Kong within hours of the release. Quick Summary & Key Takeaways What Launched: Alibaba released Qwen3.8-Max on August 3, 2026 — its largest and most capable AI model to date, built on 2.4 trillion parameters. The Big Claim: Alibaba says it performs on par with, and sometimes beats, Anthropic's Claude Fable 5 on several coding,...

Here's a number that doesn't get repeated enough: at the exact same time Big Tech was posting some of its largest quarterly layoff totals in 2026, there were roughly 275,000 AI-related job postings sitting open in the United States. Companies were cutting headcount and hiring aggressively — often in the same building, in the same quarter. That contradiction is the real story behind every "AI is taking our jobs" headline this year, and it's a lot messier than either side of the debate wants to admit. 2026 has already produced more than 170,000 tech layoffs, spread across 322 announced events at 173 companies, and 54% of those announcements explicitly named AI, automation, or machine learning as a driving factor. But "AI caused it" and "AI is the excuse" are two very different stories wearing the same headline — and this year, both are true, just for different companies. Quick Summary & Key Takeaways 170,945+ tech workers have been lai...

For a decade, Silicon Valley's biggest companies were reliable cash-printing machines, turning products like Google Search and Facebook into steady profit engines that fattened retirement portfolios across America. That era is quietly ending. According to a new Washington Post report, America's tech giants are now feeding every available dollar into what the paper calls the "cash-incinerating maw" of AI infrastructure—and the stakes now extend well beyond Silicon Valley, reaching directly into the U.S. economy and millions of ordinary retirement accounts. The numbers behind this shift are staggering. Just five companies—Alphabet, Microsoft, Meta, Amazon, and Oracle—burned a combined $563 billion in free cash flow across just five quarters. And according to Goldman Sachs, that spending is only accelerating, with total AI capital expenditure among the megacaps projected to hit $765 billion this year before climbing toward nearly $1.2 trillion in 2027. Quick Summary ...