AI chips head into orbit & Compute race tightens worldwide - Tech News (Sep 25, 2026)
Google sends AI chips toward orbit, Anthropic locks in infrastructure, Meta debuts new devices, and Waymo scales robotaxis.
Our Sponsors
Today's Tech News Topics
-
AI chips head into orbit
— Google is preparing a Project Suncatcher space test to see whether AI chips can operate in orbit, while China, NASA, and Blue Origin highlight a broader space infrastructure race. Keywords: Google, AI chips, Project Suncatcher, China satellites, Blue Origin, Moon south pole. -
Compute race tightens worldwide
— Anthropic's long-term Akamai deal, BNP Paribas expanding with Google Cloud, and Microsoft's Middle East buildout show how fierce the global AI infrastructure scramble has become. Keywords: Anthropic, Akamai, Google Cloud, Microsoft, CPU shortage, Oracle Project Jupiter. -
DeepSeek cash, Anthropic control
— DeepSeek is reportedly raising enormous new capital, and Anthropic is reportedly moving to lock in founder voting power before any IPO. Keywords: DeepSeek valuation, AI funding, Anthropic governance, voting control, startup power. -
Meta bets on AI hardware
— Meta introduced its Muse handheld and refreshed Ray-Ban smart glasses, signaling a bigger contest over consumer AI interfaces beyond the phone. Keywords: Meta Muse, Ray-Ban smart glasses, AI device, consumer hardware, assistants. -
Cyber models and RSA warning
— OpenAI is reportedly preparing a cybersecurity-focused model just as researchers reveal a new attack against some RSA signatures, increasing pressure on older security systems. Keywords: GPT-6 Cyber, RSA attack, blind signatures, Privacy Pass, cybersecurity. -
Robotaxis push toward scale
— Waymo keeps expanding paid robotaxi service, especially in Texas, while Tesla claims a major Cybercab autonomous driving milestone that still needs fuller context. Keywords: Waymo, robotaxi, Texas, Tesla Cybercab, autonomy. -
Solar efficiency record climbs
— LONGi says it has set a new crystalline-silicon solar efficiency record, a reminder that mainstream solar technology is still improving in meaningful ways. Keywords: LONGi, solar efficiency, photovoltaic, crystalline silicon, clean energy. -
X fight becomes policy clash
— The US government has stepped into Elon Musk's fight with the EU over X, turning a platform penalty into a larger argument about regulation and digital speech. Keywords: X, Elon Musk, EU fine, Digital Services Act, US intervention.
Sources & Tech News References
- → Anthropic signs up to $11.6 billion in cloud infrastructure commitments with Akamai
- → Qualcomm and Apple renew global patent licensing agreement through 2027
- → China launches the Yaogan-40 04 remote-sensing satellite group
- → LONGi reports a new 28.29% crystalline-silicon solar-cell efficiency record
- → BNP Paribas signs a five-year agreement to expand use of Google Cloud AI
- → Google prepares an orbital test of AI chips for space-based computing
- → DeepSeek reportedly completes a $7.5 billion financing round at a $75 billion valuation
- → OpenAI prepares a preview of the GPT-6 Cyber model
- → Anthropic seeks voting-control provisions for seven co-founders ahead of a potential IPO
- → Meta unveils the Muse handheld AI device
- → Microsoft commits more than $10 billion to digital infrastructure in the Middle East
- → Tesla reaches one million unsupervised miles with its Cybercab
- → Fireworks Launches Specialized Intelligence Index
- → Google’s first orbital AI data center test launches October 1
- → Bezos Has Poured $30 Billion Into Blue Origin
- → New Classical Attack Can Forge Some RSA Signatures Faster Than Expected
- → Waymo’s Robotaxi Fleet Is Growing Fast, Especially in Texas
- → AI Demand Is Driving a New CPU Shortage
- → Thinking in Systems, Shipping in Loops
- → OpenAI’s Recurrent AI and the Risk of “Neuralese”
- → Meta’s Muse AI Could Be Its iPod Moment
- → Fireworks Announces Forge Event on Specialized AI
- → NASA Chief Warns China Could Block Access to Key Lunar South Pole Sites
- → U.S. Backs X in Fight Against E.U. Fine
- → Why TLA+ Cannot Verify Everything
- → Why Debugging Is Hard for Many People
- → Oracle sends force majeure notice on New Mexico data center project
- → AI Makes Soft Skills the New Hard Skills
Full Episode Transcript: AI chips head into orbit & Compute race tightens worldwide
An AI data center in space still sounds a little unreal, but Google is about to test the idea with AI chips heading into orbit next week. Welcome to The Automated Daily, tech news edition. The podcast created by generative AI. It is September 25th, 2026, and I'm TrendTeller. Today, the scramble for AI capacity gets even more intense, consumer hardware takes another turn, and cybersecurity gets a fresh warning.
AI chips head into orbit
Let's start in orbit. Google is preparing the first real test of Project Suncatcher, which means sending experimental AI hardware into space to find out whether orbital computing is practical or just an expensive thought experiment. The first mission is expected to launch on October 1, and one of the biggest challenges is simple heat. The chips can only run briefly before the satellite has to cool down. Even so, this is a serious step toward the idea of space-based data centers.
Compute race tightens worldwide
That story fits into a much wider space race. China has launched another Yaogan remote-sensing satellite group, adding to its growing orbital network. NASA's administrator is also warning that key territory near the Moon's south pole could become strategically difficult to access if China gets there first. And on the commercial side, Jeff Bezos has now reportedly put around 30 billion dollars of his own money into Blue Origin. Space is looking less like a side project and more like core infrastructure.
DeepSeek cash, Anthropic control
Back on Earth, the AI capacity race is getting more expensive and more crowded. Anthropic has struck a massive long-term deal with Akamai for distributed AI infrastructure, a sign that major model makers are locking in compute years in advance. BNP Paribas also signed a five-year agreement with Google Cloud for broader access to AI systems, and Microsoft committed more than 10 billion dollars to digital infrastructure in the Middle East.
Meta bets on AI hardware
At the same time, supply is tightening. Reports say cloud customers are now running into CPU shortages as AI demand spreads beyond GPUs into general-purpose compute. And Oracle's big New Mexico data center project hit investor nerves after reports of a force majeure notice tied to the buildout, even though Oracle says the project remains on schedule. The message is pretty clear: demand for AI infrastructure is booming, but building enough of it is getting harder.
Cyber models and RSA warning
The money around AI companies is still arriving at remarkable speed. DeepSeek has reportedly completed a 7.5 billion dollar funding round at a valuation of about 75 billion dollars, although that report has not been independently confirmed in the material available here. If accurate, it would put DeepSeek among the most valuable AI companies in the world and give it a lot more room to buy talent, chips, and time.
Robotaxis push toward scale
Anthropic is also reportedly asking shareholders to approve a structure that would give seven co-founders majority voting control ahead of a possible public offering. That matters because the next phase of the AI industry is not just about who has the best model. It is also about who controls the company, who can overrule investors, and who gets to steer very powerful systems over the long term.
Solar efficiency record climbs
On the consumer side, Meta is pushing further into dedicated AI hardware. The company unveiled Muse, described as a small handheld AI device, alongside updates to its Ray-Ban smart glasses. The bigger point is that tech companies are still searching for the right way to make AI feel native in everyday life, whether that ends up being glasses, a handheld companion, or something else entirely.
X fight becomes policy clash
Staying with personal tech, Apple and Qualcomm renewed their global patent licensing agreement, with the new arrangement taking effect in 2027. It is not the flashiest headline of the day, but it matters. This kind of deal reduces legal uncertainty for two major players and helps stabilize a very important part of the smartphone industry while the next generation of AI-enabled devices takes shape.
In security news, OpenAI is reportedly preparing to preview GPT-6 Cyber, a model aimed specifically at cybersecurity work. A specialized model like that could be useful for defensive analysis, vulnerability hunting, and automating routine security tasks, though the report did not come with an independent OpenAI announcement in the material available here.
That arrives just as researchers have identified a new classical attack against certain RSA signatures that does not require factoring the private key. The good news is that common padded RSA schemes are not the main target. The bad news is that some blind-signature setups may be more exposed than people expected, including systems related to Privacy Pass. So this is not the end of RSA overnight, but it is another strong push away from older cryptographic assumptions.
Autonomous driving is also starting to look more commercial and less experimental. Waymo now operates in 15 US cities and says it is delivering about 500,000 paid rides each week. Most of its fleet is still concentrated in California and Texas, and Texas is growing especially fast as Waymo adds more of its newer minivans built for higher-volume service. The scale-up suggests Waymo is now focused as much on economics as on engineering.
Tesla, meanwhile, says its Cybercab program has reached one million miles driven without human supervision. That is the kind of milestone the company wants to use to show progress toward autonomous ride-hailing, although its significance depends on safety data and regulatory validation that were not detailed in the report. So for now, Waymo looks like the clearer story in real-world service, while Tesla is still making the case with milestone claims.
A quieter but still important development came from solar. LONGi says it has set a new record for crystalline-silicon solar-cell efficiency. That may sound technical, but the practical takeaway is straightforward: better efficiency can squeeze more power out of the same panel area and improve the economics of mainstream solar. It is a reminder that even mature clean-energy technology still has meaningful room to improve.
And finally, in tech policy, the US government has moved to intervene in Elon Musk's legal fight against the European Union over a 140 million dollar fine imposed on X. The EU says X misled users around verification, had unclear ad practices, and failed to give researchers proper data access under the Digital Services Act. Washington called the EU action regulatory overreach, which means this is no longer just a platform penalty. It is becoming a wider transatlantic argument about speech, platform rules, and who gets to set the boundaries for large online services.
That is your tech news snapshot for September 25th, 2026. From orbital AI experiments to tightening compute markets and expanding robotaxis, the common theme today is scale. Thanks for listening to The Automated Daily, tech news edition. I'll be back tomorrow with more.
More from Tech News
- September 23, 2026 U.S.-China AI Safety Talks & Alibaba Expands Sovereign AI Stack
- September 22, 2026 Meta Muse Sparks AI Commerce & China Builds Full AI Stack
- September 21, 2026 Military AI False Alarm & AI Rules Courts Diplomacy
- September 20, 2026 Chips become strategic battleground & China's rapid launch cadence
- September 19, 2026 Gemini breaches real company systems & Anthropic juggles IPO and oversight