"Oracle’s largest data centre partner Blue Owl Capital will not back a $10bn deal for its next facility, as the software group faces increased concerns about its rising debt and artificial intelligence spending. Blue Owl had been in discussions with lenders and Oracle about investing in the planned 1 gigawatt data centre being built to serve OpenAI in Saline Township, Michigan. But the agreement will not go forward after negotiations stalled, according to three people familiar with the matter. The private capital group has been the primary backer for Oracle’s largest data centre projects in the US, investing its own money and raising billions more in debt to build the facilities. Blue Owl typically sets up a special purpose vehicle, which owns the data centre and leases it to Oracle. Larry Ellison’s computing giant has deals to supply computing power from these data centres to AI groups such as OpenAI. The breakdown of funding discussions with Blue Owl leaves the financing of the Michigan facility in doubt, as Oracle has not yet signed a deal with a new backer, according to the people close to the matter."
Client Challenge
 #AI #Oracle #DataCenters #Debt #AIBubble #USA #Michigan #BigTech
"EU governments are pushing to widen data retention obligations for apps that citizens use every day – and the best VPN apps are among those targeted. A new internal document dated November 27 (first published by Netzpolitik) provides important insights into the current thinking of the Danish Presidency of the EU Council. It shows that member states largely agree on the need for a new framework on data retention, presenting an important overview of lawmakers main position on the matter. The topic has been debated since April, when the EU Commission first unveiled "ProtectEU," a strategy aiming to create a roadmap for "lawful and effective access to data for law enforcement." The Commission then presented the Roadmap in June, which outlined an intent to decrypt citizens' private data by 2030. Crucially, the document reveals that EU governments see metadata – specifically traffic and location history – as the most vital tool for law enforcement. Most member states argue that simply knowing who owns an account isn't enough. Instead, they want a new legal baseline where companies are forced to log exactly when and where a user was online, as well as the IP addresses they used to connect. The document notes that member states are aware of the legal hurdles of gathering this data and emphasize that any new system must include robust safeguards and strict proportionality to satisfy the courts. However, privacy experts and technologists have long warned that such 'safeguards' are not enough, arguing that you cannot weaken encryption or retain this data without fundamentally compromising user security. Besides virtual private network (VPN) companies, other online services targeted include messaging apps, hosting providers, file sharing services, cloud storage apps, and other over-the-top (OTT) services."
TechRadar
The EU prepares ground for wider data retention – and VPN providers are among the targets
A legislative proposal is expected to be introduced in the first half of 2026
 #EU #DataRetention #Privacy #VPNs #Metadata #Surveillance #DataProtection #DigitalRights #ProtectEU
"Democrats on the House Judiciary Committee have published a 27-page staff report titled “Trump, Crypto, and a New Age of Corruption”. According to the report, Trump has amassed billions of dollars from crypto ventures “from the Oval Office by steering investment to his family firm, shielding his investors from federal fraud and securities investigations and prosecutions, bilking his political base, and degrading the federal agencies ordinarily responsible for investigating bribery and tracking known bad actors online.” Among other things, the report cites Trump’s pardon of Changpeng Zhao, which it links to Zhao’s and Binance’s help in promoting Trump’s World Liberty Financial project; political donations to Trump from companies who later saw cases and investigations from the SEC and DOJ dropped [QPQ]; and the tangled mess of apparent quid pro quo surrounding MGX, Binance, the Trump family’s USD1 stablecoin, and the Emirati AI chips deal. The report concludes, “[T]he information we do have clearly demonstrates that foreign actors and corporate interests are buying access to and favors from the President and members of his Administration by investing in his family’s cryptocurrency ventures and making large, and plainly politically motivated donations.” House Judiciary Committee Ranking Member Jamie Raskin (D-MD) issued a statement alongside the report: Donald Trump has turned the Oval Office into the world’s most corrupt crypto startup operation, minting staggering personal fortunes for him and his family in less than a year. ... America has never seen corruption on this scale take place inside the White House. Since Zhao’s pardon, Binance has ramped up its support for the Trump family’s USD1 stablecoin. The exchange announced a set of new no-fee trading pairs for USD1, meaning that people can now exchange their bitcoin, ether, or various other tokens for USD1..."
Citation Needed
Issue 98 – The world’s most corrupt crypto startup operation
Terra’s Do Kwon gets 15 years, crypto banks get the green light, and the Trump family’s crypto grift expands even as one of their treasury comp...
 #USA #Trump #Crypto #Cryptocurrencies #Kleptocracy #Plutocracy #Binance
"[A] small cohort of teenage computer enthusiasts from the Princeton, N.J., area flaunted a clever work-around: They borrowed an acoustic coupler—a forerunner of the computer modem—and connected it to a nearby pay phone. With this hardware in place, the youngsters dialed in to an off-site minicomputer. The teenagers called themselves the RESISTORS, a retronym (they picked the moniker first and then matched words to the letters) for “Radically Emphatic Students Interested in Science, Technology, Or Research Studies.” The trade publication Computerworld gave the RESISTORS front-page billing—“Students Steal Show as Conference Opens”—and noted how the group drew a “fascinated crowd” of computer professionals. A reporter even suggested that the RESISTORS represented the vanguard of a small-scale social movement as the teens sought to engage with their counterparts from “underprivileged areas of Trenton” and introduce them to personal computing. In the modern history of computing, a story about a small cohort of teens “playing” with computers might seem tangential. But the previously untold history of the RESISTORS highlights the fact that, years before there were machines called personal computers, some people regularly accessed computers for activities unrelated to their professional lives. Motives varied, but entertainment as well as the display of technical prowess mattered. Just as important, the story of the RESISTORS expands our sense of the hobbyist community beyond later and better-known groups like the Bay Area’s Homebrew Computer Club." https://spectrum.ieee.org/teenage-hackers #Computers #ComputerHistory #Hacking #NewJersey #Princeton #Resistors
"As the pandemic waned, interest rates spiked, geopolitical tensions rose, and company exits ground to a halt. Investors retrenched, and by 2023, global venture funding fell to an eight-year low. Tiger went from striking just over 300 venture deals in 2022 to about 40 the following year, according to PitchBook. Some of the very startups that had ridden wave after wave of investment to stay afloat now found themselves beached, leading to mass layoffs, down rounds, and in some cases outright closures. A spate of governance issues and fraud allegations also plagued pandemic-era startup darlings around the world. The Bahamas-based crypto platform FTX crumbled, and its wunderkind founder Sam Bankman-Fried was convicted of fraud. The $22 billion Indian edtech firm Byju’s went bankrupt amid a flurry of lawsuits accusing the company of governance lapses and unpaid debts. A co-founder of the Indian car-servicing startup GoMechanic openly admitted to financial misreporting, saying he’d done it in the name of pursuing “growth at all costs.” Tiger was among the international investors who had invested in all three of those companies. (The source familiar with Tiger Global said the firm was “disappointed” with these outcomes, but emphasized that Tiger was not the largest investor in any of them.) Rest of World spoke with a wide cross-section of founders, executives, investors, and Tiger insiders about the global rise of the growth-at-all-costs model and its consequences. The founders of these now-infamous companies have rightly borne the brunt of scrutiny for the scandals and failures that followed. Yet questions remain: How much of the blame should lie with Tiger and other hyperaggressive investors for fueling the global unicorn bubble, and the slaughter that followed? And, as another bubble swells in the artificial intelligence era, has anyone learned their lesson?"
Rest of World
Big bets and broken unicorns: Tiger Global’s rise and reckoning
The fast-money pandemic era helped build billion-dollar startups around the world — then brought many down just as quickly. Will the AI boom be a...
 #India #VC #VCs #Tiger #StartUps #Unicorns #VentureCapital
"Predictive AI systems have also been shown to be incredibly useful when they leverage certain generative techniques within a constrained set of options. Systems of this type are diverse, spanning everything from outfit visualization to cross-language translation. Soon, predictive-generative hybrid systems will make it possible to clone your own voice speaking another language in real time, an extraordinary aid for travel (with serious impersonation risks). There’s considerable room for growth here, but generative AI delivers real value when anchored by strong predictive methods. To understand the difference between these two broad classes of AI, imagine yourself as an AI system tasked with showing someone what a cat looks like. You could adopt a generative approach, cutting and pasting small fragments from various cat images (potentially from sources that object) to construct a seemingly perfect depiction. The ability of modern generative AI to produce such a flawless collage is what makes it so astonishing. Alternatively, you could take the predictive approach: Simply locate and point to an existing picture of a cat. That method is much less glamorous but more energy-efficient and more likely to be accurate, and it properly acknowledges the original source. Generative AI is designed to create things that look real; predictive AI identifies what is real. A misunderstanding that generative systems are retrieving things when they are actually creating them has led to grave consequences when text is involved, requiring the withdrawal of legal rulings and the retraction of scientific articles."
MIT Technology Review
Generative AI hype distracts us from AI’s more important breakthroughs
It's a seductive distraction from the advances in AI that are most likely to improve or even save your life
 #AI #PredictiveAI #GenerativeAI
"On a low-profile blog that tracks product changes, the company said that it rolled back ChatGPT’s model router—an automated system that sends complicated user questions to more advanced “reasoning” models—for users on its Free and $5-a-month Go tiers. Instead, those users will now default to GPT-5.2 Instant, the fastest and cheapest-to-serve version of OpenAI’s new model series. Free and Go users will still be able to access reasoning models, but they will have to select them manually. (...) In practice, the router seemed to send many more free users to OpenAI’s advanced reasoning models, which are more expensive for OpenAI to serve. Shortly after its launch, Altman said the router increased usage of reasoning models among free users from less than 1 percent to 7 percent. It was a costly bet aimed at improving ChatGPT’s answers, but the model router was not as widely embraced as OpenAI expected. One source familiar with the matter tells WIRED that the router negatively affected the company’s daily active users metric. While reasoning models are widely seen as the frontier of AI performance, they can spend minutes working through complex questions at significantly higher computational cost. Most consumers don’t want to wait, even if it means getting a better answer."
WIRED
OpenAI Rolls Back ChatGPT’s Model Router System for Most Users
As OpenAI scrambles to improve ChatGPT, it's ditching a feature in its free tier that contributed to last summer's user revolt.
 #AI #GenerativeAI #OpenAI #Chatbots #GPT
"In the report, Ren and his colleagues first set out to estimate how much power California data centers had consumed in recent years, then to forecast how much they would use in the coming ones. To determine that, they looked at figures and forecasts from reports issued last year by Lawrence Berkeley National Laboratory and the Electric Power Research Institute. Using Berkeley Labs’ national numbers and calculating California’s from them by tapping the ratio of the state’s usage to the national number in the EPRI report, the researchers determined that California data centers consumed 5.5 terawatt hours of electricity in 2019. They estimated that amount rose steadily in ensuing years, hitting 10.82 terawatt hours in 2023. Ren and his team split the difference between Berkeley Lab and ERPI’s wildly different forecasts, estimating that data center energy usage in California would rise between 8.4% and 18.5% a year from 2023 to 2028. Based on that, they calculated that such power consumption by 2028 would grow to between 16.2 and 25.3 terawatt hours. Those projections represent as much as one-fourth of all commercial power consumption in the state last year, according to the California Energy Commission, and up to 8.9% of electricity usage across California as a whole. (...) All told, the on- and off-site pollutants emitted in California as a result of data centers resulted in $44.7 million in health-related costs in 2019, according to the Next10 study. That amount rose to $155.4 million in 2023. That jump was largely due to the large growth in the number of data centers over that period, according to the report. In all, health-care expenditures cost Californians $409 billion in 2023, according to the state Office of Health Care Affordability. The study forecast that the growth in costs would slow in coming years, increasing to between $167.1 million and $266.6 million in 2028."
San Francisco Examiner
Study: Data centers’ environmental impacts rising
Related emissions, water use, health costs all increasing
 #USA #California #DataCenters #Environment #Energy #Pollution
RT @TheHackersNews Attackers are abusing React2Shell to plant Linux backdoors like KSwapDoor and ZnDoor. This hits orgs that left React and Next.js servers unpatched. Microsoft saw reverse shells, Cobalt Strike, and stolen cloud tokens tied to CVE-2025-55182, and Shadowserver tracks over 111,000 exposed IPs. 🔗 Details →
The Hacker News
React2Shell Vulnerability Actively Exploited to Deploy Linux Backdoors
React2Shell vulnerability CVE-2025-55182 is actively exploited to deploy Linux malware, run commands, and steal cloud credentials at scale.
RT @rohanpaul_ai ☁️ Oracle’s latest quarterly filing shows $248B of future lease payments, mostly for data centers and cloud capacity. The key surprise is that this is roughly $150B more than earlier footnotes suggested. These are rent commitments that start between now and Oracle’s 2028 financial year. They are off the balance sheet today because many leases have not started, but they still signal a large fixed bill that will hit cash flow. Leasing can speed up AI data-center growth without buying everything upfront, but it adds debt-like pressure if cloud demand or pricing softens. Investors now have to judge whether Oracle’s AI revenue growth can comfortably cover a very rigid rent base.
Bloomberg.com
Oracle’s $248 Billion Rent is Another AI ‘Bombshell’
Of all the eye-popping numbers that Oracle Corp. published last week on the costs of its artificial-intelligence data center buildout, the most str...