CLAUDE DETHRONES CHATGPT: ANTHROPIC BEATS OPENAI IN BUSINESS AI — AND RAISES AT $900 BILLION

📡 SOURCE: VentureBeat / CNBC

Anthropic’s Claude has, for the first time, overtaken ChatGPT in U.S. enterprise AI adoption. Now 34.4% of businesses pay for Claude vs. 32.3% for OpenAI.

The company hit a $30 billion annualized revenue run rate and is reportedly raising fresh capital at a staggering $900 billion valuation, eclipsing OpenAI itself.

Hold the phone and your wallet.

The AI industry just witnessed what pundits will be calling “The Flippening” for years to come: Claude, Anthropic’s quiet, safety-obsessed AI assistant, has officially pulled ahead of ChatGPT in the corporate world.

The Ramp AI Index for May 2026 dropped the bombshell this week: 34.4% of U.S. businesses are now paying for Anthropic’s Claude, while OpenAI’s ChatGPT trails at 32.3%. For a company that spent its first two years being described as “the responsible AI lab that nerds love but nobody actually uses,” this is a massive, category-defining win.

But, it gets wilder. Anthropic just hit $30 billion in annualized revenue. $30 billion. As in, the company founded by a bunch of ex-OpenAI safety researchers who left in a huff over ethical concerns is now a bonafide financial juggernaut. Claude Opus 4.7, their latest model, dropped earlier this month and apparently it rips.

And if you thought the valuation news couldn’t get more jaw-dropping, think again my friend.

Anthropic is reportedly in talks to raise fresh capital at a $900 billion valuation, which is a couple billions higher than OpenAI.

Okay, a little more than couple.

The number is somewhere around fifty billion if my math is right.

That’s not a typo. Nine hundred billion dollars for a company that launched four years ago and brands itself on being the “responsible” option.

TRUMP FLIES TO BEIJING WITH AI ON THE TABLE WHILE THE WORLD IS WATCHING

📡 SOURCE: Axios / CNBC

President Trump is in Beijing this week for a landmark summit with President Xi Jinping, with AI safety now officially on the agenda. The U.S.-China AI model gap has effectively closed according to Stanford researchers, raising urgent questions about export controls, AI weapons, and whether the world’s two superpowers can actually agree on any guardrails before it’s too late.

When the two most powerful men on the planet sit down in Beijing to talk about artificial intelligence, you know the tech race has officially moved from Silicon Valley fever dream to geopolitical reality.

President Trump landed in Beijing this week for a high-stakes summit with Xi Jinping and for the first time, AI safety isn’t just a side note on the agenda, it’s front and center.

The urgency is real. Stanford’s annual AI Index report dropped a quiet bombshell this year: the gap between U.S. and Chinese AI models has “effectively closed.”

That’s Stanford saying it, not some Chinese state propaganda outlet.

And all of us who use DeepSeek.

American firms no longer have a commanding technical lead. And now Washington is very, very interested in what happens next.

On one side, the Trump administration wants to “open up a channel of communication on AI matters” — diplomatic speak for “we’re scared too, let’s talk.” On the other, Washington has been loudly accusing Beijing of running “industrial-scale” campaigns to steal American AI technology. So it’s a nice conversation, held with all the trust of two poker players comparing hands while hiding cards under the table.

Meanwhile, back in D.C., the White House was supposed to drop a landmark AI safety executive order this week.

But, surprisingly reports emerged that internal infighting has stalled it. And honestly, I’m shocked.

Not ideal optics when you’re flying to Beijing to lecture Xi about AI governance. The irony is delicious: the country trying to set global AI safety norms can’t even get its own paperwork together in time for the meeting.

What comes out of Beijing could shape the next decade of AI development. Or it could produce a polite communiqué that nobody reads.

GOOGLE GOES ALL-IN ON ANDROID AI — VIBE CODING YOUR PHONE WIDGETS IS NOW A THING

📡 SOURCE: TechCrunch / CNBC / 9to5Google

SUMMARY

Google unveiled a sweeping Android AI overhaul this week, with Gemini Intelligence now embedded as an operating layer across phones, cars, watches, and laptops. New features include agentic app automation, generative UI widgets you can create by describing them in plain language, and a revamped voice input called Gboard “Rambler” — all racing to market before Apple’s anticipated AI reboot at WWDC.

Google just looked Apple in the eye, cracked its knuckles, and said: “We’re not waiting for WWDC.” The Android Show this week was essentially Google firing a $40-billion AI flare into the sky, screaming “Gemini is everywhere now, deal with it.” And honestly? The features are genuinely impressive.

Let’s start with the wildest one: “vibe-coded widgets.”

Yes, believe it or not, that is the official terminology. You can now describe a widget to Gemini in plain English, eg “show me a live summary of my unread emails and today’s traffic” and the AI will generate a custom widget on the fly. This is what the tech industry means when it says AI is becoming a “generative UI” platform. Your phone interface is now, in a very real sense, programmable by vibes.

And I can still remember that we used to pay for ringtones. Like, it was…ah nevermind, let’s not digress.

Then there’s the agentic stuff — Gemini can now move across apps autonomously. It can build a shopping cart in one app, pull info from your Gmail, and book a restaurant reservation, all while you do something more interesting.

Like mindlessly scroll tiktok.

Google is careful to note that “humans stay in the loop” before any transactions. Most likely because the alternative is the AI accidentally ordering 14 pounds of quinoa in your name.

Gboard’s new “Rambler” feature uses Gemini to clean up your verbal word soup in real time — handling filler words, self-corrections, and pauses. Dictation that actually works like a human assistant, not a court stenographer with hearing loss.

The whole push is timed to arrive before Apple’s expected AI reboot at WWDC. Whether Apple has something equally dramatic up its sleeve remains to be seen. But for now, Google is swinging first. Android users, prepare for your phone to get a lot smarter — and a lot more opinionated.

And I still miss my Nokia 3310.

GOOGLE STOPPED HACKERS FROM USING AI TO LAUNCH A MASS CYBERATTACK.
THIS TIME.

📡 SOURCE: CNBC

Google’s Threat Intelligence Group says it thwarted an effort by a criminal hacker group to weaponize AI for a “mass vulnerability exploitation event.” Google says it identified hackers using an AI model to discover and exploit a zero-day vulnerability — a first of its kind detection. Security researchers warn this is a preview of the AI-powered cyberattacks to come.

Somewhere in the world this week, a hacker sat down at a keyboard, pointed an AI model at the global internet, and said “find me every exploitable system.” Google noticed. And stopped them. But make no mistake — this is a preview, not a conclusion.

Google’s Threat Intelligence Group published findings this week revealing that it likely prevented a criminal threat actor from pulling off what they described as a “mass exploitation event” — using AI to systematically hunt for zero-day vulnerabilities at a speed and scale no human team could match. Google detected the AI-powered attack chain before it reached its targets. High confidence, they said. Mass exploitation, they said. This time.

The cybersecurity community has been warning about this for years: the same large language models that help developers write better code and help users draft emails can be weaponized to find weaknesses in software at machine speed. A human hacker manually probing systems is slow and exhausting. An AI doing it is fast, relentless, and doesn’t need coffee.

The fact that Google caught this one is reassuring, right up until you remember that Google is one company watching one slice of the internet. Every security researcher worth their salt will tell you that for every attack you detect, there are others you don’t.

The arms race is officially on. AI security versus AI offense. The good guys have bigger data centers and better funding. For now at least.

The bad guys have nothing to lose and everything to gain. This week’s news is a reminder that the most dangerous applications of AI won’t look like science fiction robots — they’ll look like invisible code silently mapping every crack in the digital world.

CEREBRAS PRICES IPO ABOVE RANGE, NETS $5.55 BILLION AS AI CHIP WAR GOES PUBLIC

📡 SOURCE: CNBC TECHNOLOGY

AI chip maker Cerebras priced its IPO at $185 per share — above the expected range — raising $5.55 billion in what Wall Street is calling a landmark bet on the AI infrastructure boom. The company recently signed a $20 billion deal with OpenAI for 750 megawatts of computing capacity.

There are IPOs. Then there are moments. And then there’s Cerebras walking into a room, spiking the football before the game even starts, and watching Wall Street throw money at it like it’s raining free drinks at an open bar.

Let’s set the scene: Cerebras, the plucky AI chip upstart that decided NVIDIA needed a nemesis, priced its IPO at $185 per share which was above the already generous expected range.

And they walked away with $5.55 billion.

But wait, it gets better. Before the IPO even landed, Cerebras had already locked in a jaw-dropping $20 billion deal with OpenAI for 750 megawatts of computing capacity. For reference, 750 megawatts is enough electricity to power a small city.

OpenAI apparently needs its own small city’s worth of electricity just to keep the lights on. Relatable, honestly.

What does this tell us? The AI infrastructure arms race is no longer a metaphor — it’s an economic reality that is now publicly traded, securities-regulated, and available for retail investors to panic-buy. The real story here isn’t Cerebras specifically; it’s that the market has decided AI chips are the new oil wells, and everyone from hedge funds to your brother-in-law’s Robinhood account wants a piece of the drilling rights.

Critics will note and correctly — that Cerebras still operates in NVIDIA’s enormous shadow. NVIDIA has the software ecosystem, the developer loyalty, and the kind of market dominance that makes competitors’ business plans look like strongly-worded suggestions.

But the Cerebras IPO signals something important: investors believe there’s room for more than one winner in the AI chip gold rush.

The question, as always, is whether the gold is real or if everyone’s just carrying shiny rocks. For now, Wall Street has spoken, and it says: more chips, please.

FULL STORY: https://www.cnbc.com/2026/05/13/cerebras-prices-ipo-above-expected-range-wall-street-expects-ai-flood.html

GOOGLE RACES TO MAKE GEMINI YOUR PHONE’S BRAIN BEFORE APPLE WAKES UP

📡 SOURCE: CNBC TECHNOLOGY

Google has launched sweeping Android AI updates, positioning Gemini as an operating layer that can move across apps, read your screen, manage your Gmail, build shopping carts, and book reservations. And they did all that before Apple is expected to unveil its own AI-powered iOS overhaul later this year.

In the great AI phone war of 2026, Google just showed up to the knife fight with a flamethrower — and Apple is still fumbling around in the drawer looking for scissors.

Google announced this week that Gemini is no longer just a chatbot you occasionally ask for a recipe. It is now, in Google’s words, “an operating layer” for your entire digital life. It can cross between apps. It reads your screen. It rifles through your Gmail (with permission, presumably — though let’s be honest, it was probably already in there). It builds shopping carts. It books reservations.

At this rate, by the end of 2026, Gemini will be attending your meetings for you while you nap.

This is a deeply strategic move, and the timing is not accidental. Apple is reportedly gearing up for a major Siri overhaul — a long-overdue upgrade that has been the butt of more tech jokes than any assistant deserves. Apple has even quietly partnered with Google to power some of its AI features using Gemini’s technology, which is the corporate equivalent of asking your rival to ghostwrite your homework.

Google knows this. And Google is not waiting politely. They are moving fast, pushing Gemini into every Android crevice possible, so that by the time Apple announces its shiny new AI features at WWDC, Android users will have been living in an AI-integrated world for months. First-mover advantage is everything when you’re trying to win the hearts and home screens of 3 billion smartphone users.

The consumer experience, meanwhile, sounds genuinely impressive. An AI that can actually complete multi-step tasks across apps without you babysitting it? That’s not a feature — that’s a personal assistant. The kind people used to pay salaries for. The kind that is now contributing to mass layoffs.

Progress is funny like that.

FULL STORY: https://www.cnbc.com/2026/05/12/google-races-put-gemini-at-center-of-android-before-apples-ai-reboot.html

PALO ALTO SOUNDS ALARM: AI CYBERATTACKS TO HIT ‘CRITICAL MASS’ WITHIN MONTHS

📡 SOURCE: CNBC TECHNOLOGY

Palo Alto Networks is warning that companies have a ‘narrow three-to-five-month window’ to fortify their defenses before AI-powered cyberattacks become the new normal. The cybersecurity giant says hackers are already using AI to rapidly identify and exploit software vulnerabilities at a scale previously impossible.

Three to five months. That’s it. That’s the window. That’s what the people who protect the internet’s vital organs are telling us before AI-powered hackers go from “emerging threat” to “Tuesday.”

Palo Alto Networks is a company that exists specifically because the internet is a lawless hellscape — issued what can only be described as a politely worded apocalypse warning this week. Their tech chief Lee Klarich told the world that organizations have roughly three to five months to get their cybersecurity act together before AI-driven exploits become the “new norm.” The phrase “new norm,” used in this context, should cause anyone responsible for enterprise IT infrastructure to immediately put down their coffee and have a very serious meeting.

Here’s what makes AI cyberattacks genuinely different from the old-fashioned kind: speed and scale. A human hacker takes time. They poke around, they probe, they occasionally stop to get a sandwich. An AI attacker doesn’t need sandwiches. It scans millions of endpoints simultaneously, finds the vulnerabilities, writes the exploit code, deploys it, and moves on.

All that while you’re still reading this sentence. Traditional cybersecurity was built for human-speed threats. AI threats operate at machine speed, and most defenses simply aren’t ready.

The implications here go beyond corporate IT departments sweating through their polo shirts. Critical infrastructure like power grids, water systems, hospitals, financial systems, they all run on software. Software has vulnerabilities. AI can find those vulnerabilities faster than humans can patch them. This is not a hypothetical. This is the trajectory.

Someone please get John McClaine.

But, to be completely fair, defenders also have AI now. The arms race is mutual. Of course, like always, the attackers have one significant advantage: they only need to get it right once. The defenders need to get it right every single time.

Tick tock.

FULL STORY: https://www.cnbc.com/2026/05/13/palo-alto-ai-cyberattacks-mythos-gpt.html

COMPANIES ARE FIRING WORKERS FOR AI THAT ISN’T EVEN WORKING, GARTNER STUDY REVEALS

📡 SOURCE: FORTUNE

A bombshell Gartner study reveals that companies cutting jobs in the name of AI automation are frequently not seeing the returns they promised — and in many cases, the AI isn’t even delivering results yet. The data shows workforce reduction rates were nearly equal for companies reporting high ROI and those reporting worsened outcomes, exposing AI-driven layoffs as potentially premature or driven by optics over economics.

Someone finally said the quiet part loud: companies are firing humans for AI robots that aren’t performing any better. And in some cases, are actively making things worse. This is the corporate equivalent of firing your entire kitchen staff because you bought a $40,000 coffee machine that keeps brewing lukewarm decaf.

A new Gartner study dropped this week like a brick through the floor of every board presentation that used the phrase “AI-driven efficiency gains” in the past 18 months. The findings: 80% of companies that piloted AI automation reported workforce reductions. The kicker? The companies reporting the biggest ROI from AI were not the same ones cutting the most jobs — and the layoff rate was nearly identical whether the AI was generating returns or wrecking them.

In other words, the layoffs are happening regardless of whether the AI works.

Companies are cutting people because it looks decisive, because the stock market rewards it, because the earnings call narrative sounds better when you say “we’re investing in automation” rather than “we bought expensive software that sort of works sometimes.” The humans get fired. The AI gets the headcount. The ROI spreadsheet stays aspirational.

This isn’t a small problem. Over 92,000 tech workers have been laid off in 2026 alone. The broader narrative being sold to investors and board members is that AI will pay for itself by replacing human labor. But if the Gartner data holds up, many companies are discovering that AI handles the easy stuff and struggles badly with the nuanced, judgment-heavy work that actually required humans in the first place.

The real danger isn’t that AI is too good. It’s that companies are treating it as though it is — before it actually is — and real people are paying the price for the hype.

The robots will get there eventually. Just maybe not by Q3.

FULL STORY: https://fortune.com/2026/05/11/ai-automation-layoffs-gartner-study-roi/

AI MEMORY CHIP STOCKS SURGE 60% IN SIX WEEKS; HARVARD EXPERT SCREAMS BUBBLE

📡 SOURCE: FORTUNE

The Philadelphia Stock Exchange Semiconductor Index has surged 60% in just six weeks, with memory chipmaker Micron posting its best week since 2008 — a 38% single-week gain. Harvard economists are sounding the alarm that the AI chip frenzy has the hallmarks of a speculative bubble, even as Apple warns that rising memory prices are squeezing its margins.

Sixty percent. Six weeks. Read that again and then go check what your index fund did in six weeks. Right. Now you understand why every hedge fund manager on Wall Street is currently developing a very spiritual relationship with the AI chip sector.

The Philadelphia Stock Exchange Semiconductor Index — a mouthful that traders have mercifully shortened to “the SOX” — has launched itself 60% higher in just six weeks. Micron, a memory chipmaker that most people couldn’t have named at a cocktail party in 2024, just posted a 38% gain in a single week. Its best week since 2008. For context: 2008 was not a great year for most things, which tells you how extraordinary a 38% single-week pop actually is.

The driver, as with approximately everything in tech right now, is AI. Training and running large AI models requires enormous amounts of high-bandwidth memory — the specialized DRAM chips that Micron and its competitors produce. AI infrastructure buildout is consuming these chips faster than manufacturers can make them, driving prices up sharply and making every memory chipmaker on earth feel briefly like a tech genius.

But here’s where the fun part ends: Harvard economists are waving red flags with the enthusiasm of people who’ve seen this movie before. The hallmarks are all there — parabolic price movements, narrative-driven valuations, retail FOMO piling in after the smart money, and underlying demand projections that assume AI adoption will continue on its current exponential trajectory forever. Spoiler: things that go up 60% in six weeks sometimes do not continue going up 60% every six weeks.

Apple CEO Tim Cook is already feeling it — warning analysts that rising memory prices are crushing his gross margins. When a company with Apple’s purchasing power is wincing at chip costs, the ripple effects downstream are going to be felt by everyone.

FULL STORY: https://fortune.com/2026/05/11/ai-memory-chips-semiconductor-stock-boom-price-hikes-dram-shortage-hbm/

Keywords: AI news May 2026, Anthropic Claude ChatGPT enterprise, Trump Beijing AI summit China, Google Android Gemini AI widgets, AI cyberattacks Google hackers, Cerebras IPO AI chips, Palo Alto AI security, AI layoffs Gartner study, semiconductor stocks bubble, AI chip war NVIDIA Micron