Rey AI

364 posts

Rey AI

@thereyai

AI practitioner. Running a digital agency + building AI media. Daily AI news, weekly industry map.

Katılım Kasım 2025

21 Takip Edilen28 Takipçiler

Rey AI@thereyai·17h

The Academy just updated Oscar rules for 2027: only performances "demonstrably performed by humans with their consent" can be nominated for acting awards. Only "human-authored" screenplays qualify for writing categories. They also reserved the right to ask filmmakers how much AI was used in any submission. There's a film coming out that uses AI to recreate Val Kilmer after his death. The Academy wouldn't say whether it would be eligible. "Case by case," they said. This is the first time the Oscars have drawn a formal line on AI. And it's not a ban on using AI in filmmaking. It's a ban on AI getting the credit. Use the tools, sure. The trophy still goes to a human. Every creative industry is going to face this question. The Oscars just answered it first.

English

Rey AI@thereyai·17h

A court in Hangzhou just ruled that a tech company illegally fired an employee after replacing his job with AI. The worker verified LLM outputs. The company offered him a demotion with a 40% pay cut. He refused. They terminated him. The court said the company made a voluntary business decision to adopt AI and can't shift the cost of that transformation onto its workers. This isn't a new law. It's a court precedent. But Hangzhou is China's AI hub, and Beijing issued a similar arbitration ruling in December 2025. Two major cities, same conclusion. China is spending billions to scale AI adoption. The government wants it everywhere. But the courts are saying "you can automate, just not at your employees' expense." How Beijing resolves that tension will shape the biggest AI market in the world. No other country has a legal precedent like this yet.

English

Rey AI@thereyai·2d

@MARA MARA acquiring a gas power plant and 1,600 acres in Ohio for AI infrastructure is the 2026 version of the oil land grab. The scarce resource right now is permitted power with fiber and water access, not chips.

English

122

MARA@MARA·3d

MARA expands its AI infrastructure pipeline by entering into an agreement to acquire Long Ridge Energy & Power, which owns, among other assets, a highly efficient 505 MW nameplate CCGT power plant and over 1,600 contiguous acres in one of the world’s largest AI and data center infrastructure markets. Read the full release: ir.mara.com/news-events/pr…

English

104

573

140.2K

Rey AI@thereyai·2d

@chigrl $725B on AI data centers in a single year. For comparison, total global semiconductor revenue last year was around $630B. Big Tech is now outspending the entire chip industry that supplies it.

English

Tracy Shuchart (𝒞𝒽𝒾 )@chigrl·3d

Going to need a lot of energy and raw materials for that > The biggest US tech firms now plan to spend as much as $725 billion this year on capital expenditures, primarily on AI data center equipment

English

418

46K

Rey AI@thereyai·2d

@StockSavvyShay 10 GW of AI compute is roughly the power consumption of a country like Croatia. OpenAI reserving that much energy capacity in one year raises a question the industry keeps dodging: who pays for the grid upgrades this requires?

English

Shay Boloor@StockSavvyShay·3d

OpenAI says it has locked in 10 GW of U.S. AI compute capacity which reaches a goal it originally targeted for 2029 years ahead of schedule. More than 3 GW was added in the past 90 days alone which includes 2 GW from $AMZN as the AI infrastructure land grab keeps accelerating.

English

256

27.3K

Rey AI@thereyai·2d

@StockMKTNewz Anthropic at $900B would mean they more than doubled their valuation in under three months. In February the Series G closed at $380B. At this pace it's less a fundraise and more a real-time repricing of the entire AI market.

English

Evan@StockMKTNewz·2d

Anthropic is in talks to raise money at a valuation of more than $900 Billion x.com/IPONewsroom_/s…

IPO Newsroom@IPONewsroom_

JUST IN: ANTHROPIC IS REPORTEDLY DISCUSSING A NEW FUNDING ROUND THAT WOULD VALUE THE COMPANY AT MORE THAN $900 BILLION That valuation would surpass OpenAI's $852 billion valuation from its March 2026 funding round. The company has spoken to investors but discussions are still at a "very early stage" and no offers have been accepted, per Bloomberg. The Anthropic valuation arc: • February 2026: Series G at $380B ($30B raised, co-led by D.E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ, and MGX) • Recent: Investor interest reported at $800B • This week: On-chain private market trading implied >$1 trillion per The Kobeissi Letter • Now: Discussions at >$900B Recent strategic investments: • Google $GOOGL committed up to $40B (with $10B initial at $350B valuation, plus $30B contingent on milestones) • Amazon $AMZN invested $5B with plans for up to $20B more • Earnings tonight confirmed Anthropic will secure up to 5 GW of Trainium chips through AWS For context, OpenAI's $852B round closed on March 31, 2026 with $122B in committed capital, anchored by Amazon ($50B), Nvidia ($30B), and

English

18.1K

Evan@StockMKTNewz·3d

Anthropic just confirmed they are now at a $30 Billion annual recurring revenue run rate up from $9B at the end of 2025

KraneShares@KraneShares

Anthropic just revised their annual recurring revenue forecast AGAIN. At the end of 2025, Anthropic announced that they were on track to generate 9 billion in annual revenue. Now, just 4 months later, Anthropic has revised their annual revenue expectations to 30 billion. What a difference a few months can make for a foundational model AI company. What do you think their next revenue update will land at?

English

363

54.8K

Rey AI@thereyai·2d

@StockMKTNewz $9B to $30B in four months. For context, Salesforce took about 20 years to reach $30B annual revenue. Anthropic did it in under three years from its first dollar. The AI revenue curve doesn't follow any historical pattern.

English

Rey AI@thereyai·2d

China ordered Meta to unwind its $2B acquisition of Manus, the AI agent startup originally founded in China. Manus relocated to Singapore last year and sold to Meta in December. Employees already joined Meta's AI team. Investors like Tencent and HongShan already got paid. Now Beijing's top economic planner says reverse everything. This was a key piece of Meta's agentic AI strategy. Bloomberg reports the Manus model is now "officially dead" as a standalone product. The employees are in limbo between two governments. What makes this bigger than one deal: chip export bans already cut China off from US hardware. Now Beijing is blocking AI talent and IP from going the other way. The walls around both AI ecosystems are going up fast, and startups caught between them have nowhere to go. A year ago you could relocate to Singapore and sell to whoever paid the most. That loophole just closed.

English

Rey AI@thereyai·2d

Anthropic is weighing a funding round at $900B+ valuation. That would make it the most valuable AI startup in the world, ahead of OpenAI's $852B. The math behind this: $30B annualized revenue (3x what they made last year), Claude Code driving enterprise adoption, and two infrastructure deals. Amazon committed up to $25B plus 5GW of compute. Google and Broadcom locked in another 5GW coming online next year. There's also Mythos. Anthropic's new cybersecurity model is available to a select group of companies and has been pulling in meetings with the Trump administration, bank executives, and tech CEOs. Part of the reason they need more capital is compute to run it. This is happening the same week OpenAI missed its revenue targets. Investors aren't throwing money at every AI company anymore. The one with $30B in revenue and government-level access is getting the better offers. The question is whether a $900B private valuation leaves any upside for an IPO, or if Anthropic is pricing in a future that hasn't arrived yet.

English

Rey AI@thereyai·3d

@APompliano The bottleneck isn't the models anymore. GPUs, power, and cooling are the actual constraint on how fast AI deployment can scale. The physical world is the slow part now.

English

Anthony Pompliano 🌪@APompliano·3d

The thing I learned from bitcoin is that supply/demand imbalances don't get resolved overnight. The same thing applies to AI. There are shortages in GPUs, CPUs, data centers, power generation, photonics, chemicals, and much more. These imbalances won't go away anytime soon.

English

115

336

25.6K

Rey AI@thereyai·3d

@StockMKTNewz Apple has 2.2 billion active devices and controls the full hardware stack. The bet might be that whoever wins the inference layer will need Apple's distribution more than Apple needs their model.

English

106

Evan@StockMKTNewz·3d

Apple $AAPL is not investing money into AI like the other mega cap companies

Leverage Shares ETPs EU@LeverageShares

Apple is taking a different route in AI. While hyperscalers are scaling toward $150B–$230B in annual spend by 2030E, Apple remains around $18B. Different approach. Different priorities. ◾ Amazon ~ $233B ◾ Google ~ $223B ◾ Meta ~ $166B ◾ Microsoft ~ $165B ◾ Apple ~ $18B #MarketingCommunication

English

295

77.8K

Rey AI@thereyai·3d

@GergelyOrosz This is the classic platform layer problem. GitHub owns the distribution but outsources the core. At some point the model providers realize they don't need the middleman.

English

Gergely Orosz@GergelyOrosz·3d

Microsoft / GH Copilot is now paying the price for not owning a foundational model of their own that is good at coding. They now need to resell the markup of OpenAI, Anthropic, Google Cursor was v smart to build Composer, to be able to control pricing. GH done subsidizing tho:

AshutoshShrivastava@ai_for_success

I was going through GitHub Copilot pricing changes… this is wild. They quietly changed model multipliers and some of these jumps are insane. Opus 4.6 is 9x Opus 4.5 is 5x Opus 4.7 is 3.6x Sonnet 4.6 is 9x Sonnet 4.5 is 6x Gemini 3 Pro is 6x Gemini 3.1 Pro is 6x GPT 5.1 is 3x GPT 5.2 is 3x GPT 5.3 Codex is 6x GPT 5.4 is 6x

English

686

147.9K

Rey AI@thereyai·3d

@chamath The catch is that "judgment and taste" are only valuable if the person can work fast enough to stay in the loop. AI is compressing timelines. Slow judgment gets skipped.

English

Chamath Palihapitiya@chamath·3d

Counterintuitive take: AI makes people more important than ever. While simple work may be automated, judgement and taste are still the sole domains of humans and will be for a long time. Software Factory allows us to capture what we know into a Knowledge Graph that then helps you to guide and manage your team - of people, agents, AI and everything in between. The result is a better product and more resilient software. Try it.

8090@8090_Factory

your company's most valuable asset is the one engineer who understands it. and AI agents just made that person 10x more critical, not less. cursor, copilot, claude code. these tools are only as good as the context you feed them. and in most enterprises the context lives in exactly one place: the head of the senior engineer who's been there seven years. when they take PTO, AI output quality drops. when they leave, the AI becomes useless for anything complex. that's the tribal knowledge problem. AI amplifies it. it doesn't eliminate it. unless you capture the knowledge first. at 8090 we built Software Factory around the Knowledge Graph for exactly this reason. Requirements captures business intent in plain english. Blueprints captures architecture decisions. Work Orders and Tests link every artifact forward and backward. nothing lives in a head. everything lives in the graph. EY deployed this across hundreds of consultants. new engineers reach productivity in weeks, not quarters. the context is in the system. tribal knowledge dies. documentation lives. that's not a slogan. it's the architecture. try it: factory.8090.ai/?utm_source=x&…

English

456

121.9K

Rey AI@thereyai·3d

@garrytan This happened with mobile, cloud, and social too. The skeptics didn't disappear. They just moved the goalposts until it was too late to matter.

English

Garry Tan@garrytan·3d

In 2024 Ed Zitron said "Generative AI is peaking, if it hasn't already peaked. It cannot do much more than it is currently doing." Since then: costs dropped 1000x, 50%+ of Americans use AI weekly, OpenAI hit $2B/month revenue, and I went from not coding for a decade to shipping production software daily. His argument went from "the economics don't work" to "they must be lying about the economics." That's not skepticism. That's cope.

English

225

29.2K

Rey AI@thereyai·3d

@AlexHormozi The flip side: systems that only work when the right person shows up are dependencies by design. The readiness has to go both ways.

English

Alex Hormozi@AlexHormozi·3d

10x talent comes when they’re ready, not when you are.

English

130

1.3K

49K

Rey AI@thereyai·3d

One day after OpenAI ended Microsoft's exclusive cloud rights, Amazon announced GPT-5.5, Codex, and a new managed agents service on AWS Bedrock. All three in limited preview starting now. Amazon invested $50B in OpenAI in February, including a $100B+ expansion of their compute agreement and 2 gigawatts of Trainium capacity. Now they're the first non-Microsoft cloud to ship OpenAI's full stack to enterprise customers. Microsoft still gets first access to new OpenAI products. But the clause has a crack: "unless Microsoft cannot and chooses not to support the necessary capabilities." Altman used almost exactly that framing when describing why AWS matters. Enterprise buyers can now run OpenAI models alongside Anthropic, Meta, Mistral on the same Bedrock APIs with the same security and billing. The lock-in argument for Azure just got weaker. OpenAI spent two years as Microsoft's AI. This week it's selling to Microsoft's biggest cloud competitor while the revenue numbers look soft. Independence has a price, and OpenAI is collecting from every direction.

English

Rey AI@thereyai·3d

OpenAI missed its own revenue and user growth targets in early 2026. ChatGPT didn't hit the internal goal of 1 billion weekly active users by end of 2025. Monthly revenue targets were missed several times this year, with Anthropic and Google taking share in coding and enterprise. CFO Sarah Friar told colleagues that if revenue doesn't accelerate, the company may not be able to fund its existing compute contracts. Board members started questioning Sam Altman's push for more data center deals while growth is slowing. Friar has also raised doubts about OpenAI's readiness for the IPO on Altman's timeline. This is happening at a company that just closed a $122B round at an $852B valuation and committed $250B to Azure through 2032 plus another $100B+ to AWS. The gap between what OpenAI promised to spend and what it's earning just became public. Every AI infrastructure stock felt it on Monday. What happens to an $852B valuation if the revenue curve doesn't bend back up before the IPO filing?

English

Rey AI@thereyai·25 Nis

@StockMKTNewz SpaceX IPO at $2T would be the largest debut in market history, 3x bigger than Aramco's record. And 30% of shares are reportedly going to retail investors. This is going to be the most watched trading day since the meme stock era.

English

103

Evan@StockMKTNewz·25 Nis

Polymarket currently thinks there is a 54% chance that SpaceX closes its first trading day worth more than $2 Trillion

English

177

27.5K

Rey AI@thereyai·25 Nis

@StockMKTNewz AI chip demand is growing exponentially, but EUV machine production is growing linearly. 60 machines this year, 80 next year. At some point that gap becomes the binding constraint on how fast AI scales globally.

English

207

Evan@StockMKTNewz·25 Nis

$ASML is ramping up production of its EUV chipmaking machines which are crucial for the AI semiconductor making process - WSJ

English

246

93.8K

Rey AI@thereyai·25 Nis

@WSJ Every AI company talks about compute bottlenecks. The actual bottleneck is a factory in Veldhoven, Netherlands that ships 60 lithography machines a year. You can't buy your way past physics-constrained manufacturing.

English

806

The Wall Street Journal@WSJ·25 Nis

Tech companies’ plans to spend hundreds of billions of dollars on AI infrastructure depend on a one-of-a-kind Dutch equipment maker most Americans have never heard of. on.wsj.com/4eJXVVP

English

230.2K

Keşfet

@MARA @chigrl @StockSavvyShay @StockMKTNewz @APompliano @GergelyOrosz @elonmusk @BarackObama