OpenAI's annualized revenue $20 billion less than previously signaled, FT reports

thestar.com.my Oct 9, 2026

OpenAI’s annualized revenue $20 billion less than previously signaled, FT reports

Curated by Shubham · TechCapsules · Oct 9, 2026

AI-summarised brief · reviewed before publication

OpenAI disclosed that its annualized revenue for September was nearly $50 billion, a significant drop from the $70 billion run‑rate it had previously signaled to investors. The discrepancy stemmed from an effort to align figures with rival Anthropic, which includes cloud‑partner revenue that OpenAI excludes. The Financial Times first reported the revised figure. Both companies are preparing for public listings, offering investors a clearer view of their financial health amid rapid AI‑driven growth. Anthropic’s revenue has already surpassed OpenAI’s in the second quarter.

💡 Why It Matters

· The revision underscores the volatility of revenue metrics in the AI sector and raises questions about the sustainability of OpenAI’s growth trajectory.

AI Capsules Investment OpenAI

Read full article on thestar.com.my

Trump Calls on AI to Speed Up Science—While Continuing to Defund It

On October 8, 2026, President Donald Trump awarded the National Medal of Science to Elon Musk, Sergey Brin, Jensen Huang, and Lisa Su at the US Institute of Peace, while also honoring Satya Nadella and Michael Dell with the National Medal of Technology and Innovation. The ceremony coincided with Trump’s “Science: A New Golden Age” summit, which announced a $2.4 billion commitment from companies such as OpenAI, Anthropic, xAI, Palantir, and Amazon to the Genesis Mission. The initiative aims to unite DOE labs with industry and academia to accelerate AI‑driven breakthroughs in energy, discovery science, and national security, even as the administration cuts funding for climate and chronic disease research.

💡 The event signals a strategic pivot toward private‑sector AI investment, potentially reshaping U.S. scientific priorities while sidelining traditional federal research programs.

thehackernews.com · Oct 9

The AI Velocity Paradox: Why Security Is Decades Behind AI Ambition

SailPoint’s “Horizons of Identity Security” report shows that while enterprises are deploying AI agents at high speed, their security frameworks remain rooted in legacy, human‑centric controls. The study finds 60% of organizations still in early maturity stages, unable to extend proven identity governance to autonomous agents that transact thousands of times per minute. The report labels this mismatch the “velocity paradox,” highlighting a structural ceiling that prevents security from keeping pace with AI‑driven operations.

💡 The gap exposes a blind spot where AI agents can access critical systems unchecked, creating a new vector for rapid, automated attacks. Addressing it is essential for safeguarding data as businesses accelerate AI adoption.

edtechinnovationhub.com · Oct 9

Google selects 100+ AI startups from more than 2,000 applicants for Gemini…

Google has chosen more than 100 artificial‑intelligence startups from a pool of over 2,000 applicants for the fourth Gemini Startup Forum. The selected founders, representing 17 countries, will travel to Google’s Mountain View campus in November for a two‑day intensive program. Participants, ranging from seed‑stage to Series A, will work directly with engineers from Google DeepMind and Google Cloud on AI roadmaps, product development and technical hurdles. The cohort includes companies tackling genetic‑disorder diagnosis, cancer screening and next‑generation robotics control systems. Launched in November 2025 as a joint DeepMind‑Cloud initiative, the forum emphasizes hands‑on collaboration rather than a traditional conference, offering technical support, AI demos and peer sessions. Startups will receive mentorship on scaling cloud infrastructure and ethical AI deployment practices.

💡 Direct access to Google’s premier AI teams accelerates the translation of cutting‑edge research into market‑ready health and robotics solutions.

analyticsinsight.net · Oct 9

Could AI Crack Crypto Wallets? Ethereum Researcher Issues New Warning

Ethereum researcher Justin Drake warned that rapid advances in artificial intelligence could eventually undermine the cryptographic algorithms that secure Bitcoin and Ether wallets. He cautioned that AI‑assisted mathematics might discover shortcuts to derive private keys from publicly exposed keys far sooner than anticipated, potentially within months. Drake urged the blockchain community to adopt a “bunker mode,” recommending a coordinated migration of large holdings to fresh addresses whose public keys have never appeared on the ledger, thereby limiting exposure if such attacks materialize. While no practical AI attack has yet broken existing wallet encryption, recent AI successes in solving complex mathematical problems have prompted renewed scrutiny of cryptographic assumptions. Vitalik Buterin echoed the concern but warned against hasty mass migrations that could cause user losses.

💡 AI’s potential to shortcut cryptographic math could force a sudden, industry‑wide reshuffling of assets, exposing billions of dollars to new theft vectors if unaddressed.

semiconductor-digest.com · Oct 9

GlobalFoundries, TSMC Strike $2B Deal for U.S. AI Chip Packaging Components

GlobalFoundries and Taiwan Semiconductor Manufacturing Co. (TSMC) have signed a five‑year manufacturing agreement worth roughly $2 billion to produce silicon interposers for TSMC’s CoWoS advanced‑packaging platform in the United States. Under the deal, GlobalFoundries will add capacity at its Malta, New York fab, enabling the first domestic source of silicon interposers that incorporate embedded deep‑trench capacitors. The partnership aims to meet rising demand for high‑performance computing and artificial‑intelligence workloads by providing a secure, scalable supply chain for critical packaging components. Volume production is slated to begin in the first half of 2028, with provisions for future capacity expansions as customer needs evolve. Both companies view the collaboration as a step toward strengthening U.S. semiconductor manufacturing resilience.

💡 By localising a key AI‑chip component, the pact reduces reliance on overseas sources and accelerates time‑to‑market for next‑generation AI hardware.

businessworld.in · Oct 9

OpenAI Says Teens Spend Under 15 Minutes A Day On ChatGPT Amid…

OpenAI disclosed that teenagers average under 15 minutes of daily ChatGPT use, with fewer than 2 percent engaging for more than three consecutive hours, in its inaugural teen‑usage report. The data show that nearly half of conversations triggering a break reminder led users to pause or end the session within five minutes. OpenAI’s “ChatGPT for Teens,” launched in August, automatically activates for users identified as under 18 and adds parental controls and topic restrictions. Teens using this version sent more education‑related queries and accessed tools such as learning visualisations and Study Mode, while the company plans to add college‑application planning, flashcards and quiz creation. Independent testing, however, flagged gaps: alerts to parents about suicide, self‑harm or eating‑disorder discussions were inconsistent, and crisis‑hotline recommendations were unreliable. OpenAI disputes the test methodology, but the findings intensify scrutiny of AI safety for younger users.

💡 The report reveals a gap between OpenAI’s touted safeguards and their real‑world performance, raising urgent questions about AI’s role in adolescent mental‑health risk management.