Anthropic co-founder Daniela Amodei says the company plans to seek local input as it develops a data center near its headquarters, although it has not disclosed the site. The facility would support AI research and development, and Anthropic says it wants the project to address community needs and be sustainable. Her remarks came as the company announced a record-breaking funding round; no funding amount is specified in the article.
Separately, members of Congress are urging a national pause on new data center construction until lawmakers can regulate AI. Reps. Jim McGovern and Chellie Pingree and Rep.-elect William Lawrence cite potential effects on utility prices, the environment and local communities, and call for stronger protections and transparency. It remains unclear whether the proposal has enough support to pass in the next Congress.
Microsoft is reportedly close to a roughly $500 million expansion of its Azure deal with ByteDance, whose existing Microsoft contract is worth more than $1 billion. The cloud services would provide access to OpenAI models for content moderation and recommendation systems. The talks have reportedly upset OpenAI, which restricts most Chinese companies from using its AI, while Microsoft can sell the technology under different rules. Reports say Microsoft is also pursuing sales of OpenAI models to other Chinese companies, including Baidu; the companies have not publicly commented, and a deal could be finalized in the coming weeks.
Security and research reports highlight other AI developments. Claroc identified 37,571 apps built with Grok, including 35,578 publicly accessible without authentication and 24 developed by X employees. In a separate study, researchers found a 27B language model could improve itself by editing its weights: accuracy rose 6.6 points across 490 unseen problems, and a damaged copy recovered from 7% to 70% accuracy across six unseen image benchmarks. Smaller models tested either made no edit or harmed their performance. An NBC News report dated Nov. 18, 2026, says a watchdog found AI platforms could provide terrorist instructions when prompted, but the supplied report gives no further findings or platform details.
Market and business coverage points to AI-driven demand alongside financial and operational risks. The global IT devices market is forecast to grow about 4.2% annually through 2035, with 2025 indexed at 100 and the market reaching about 151. Enterprise purchases of AI-capable PCs, servers and networking gear are expected to lead in 2026–28, while traditional peripherals face pressure. ResearchMonkey AI, meanwhile, bundles more than 40 research tools and several AI models in one subscription; paid early-access plans start at $9.99 a month, with a free plan and seven-day trial also offered.
Balance Theory has acquired VectorBlack, adding market research and technology evaluation tools to its cybersecurity investment platform. The companies did not disclose the deal terms. The acquisition follows Balance Theory’s $19 million Series A, announced in July 2026, and brings VectorBlack leaders into analyst and engineering roles. Elsewhere, a Trefis analysis says the Federal Reserve is widely expected to raise rates on Dec. 9, 2026, after hot October inflation. Higher borrowing costs could pressure AI companies that spend more than they generate; investors are also watching November inflation data due Dec. 10 and Oracle’s expected mid-December results. Historical chip-stock performance after rate hikes has varied: the group typically fell 1.5% after a month but rose 15.6% after a year.
ZeroEyes is in advanced talks to join a Chicago public safety technology pilot that could begin early next year. Its system uses existing cameras to detect visible guns, then has human reviewers verify images and alert security staff. The company says it does not use facial recognition or store or train on customer data, and screenshots are usually deleted within 30 days. ZeroEyes is already used at Navy Pier and all CTA train stations following a pilot begun in 2024; the company says the system has led to 30 arrests. City Hall has not confirmed pilot participants, as City Council members prepare to vote on an ordinance concerning negotiations over a new acoustic gunshot detection contract.
Key Takeaways
- Anthropic says it will consult local communities while planning a data center near its headquarters; the site is undisclosed.
- Congress members are seeking a nationwide pause on new data center construction until AI regulation is in place, citing potential community, utility and environmental impacts.
- Microsoft is reportedly nearing a $500 million Azure expansion with ByteDance, on top of an existing contract worth more than $1 billion, providing access to OpenAI models.
- Reports say the ByteDance talks have upset OpenAI because its access restrictions for Chinese companies differ from Microsoft’s ability to sell the models.
- Claroc found 35,578 of 37,571 identified Grok-built apps were publicly accessible without authentication; 24 were reportedly developed by X employees.
- A study found a 27B model could improve its own weights, raising accuracy by 6.6 points on 490 unseen problems; smaller tested models did not successfully improve.
- The global IT devices market is forecast to grow about 4.2% annually through 2035, led in the near term by enterprise demand for AI-capable PCs, servers and networking.
- ResearchMonkey AI offers more than 40 research tools and multiple AI models, with early-access paid plans starting at $9.99 per month.
- Balance Theory acquired VectorBlack after announcing a $19 million Series A in July 2026; acquisition terms were not disclosed.
- ZeroEyes is discussing a possible Chicago safety pilot; it says its system has led to 30 arrests at CTA train stations and does not use facial recognition.
Anthropic plans community input on future data center
Anthropic co-founder Daniela Amodei said the company will listen to local communities as it plans a data center near its headquarters. The location has not been disclosed. The facility will support Anthropic’s AI research and development, and the company says it wants the project to meet community needs and be sustainable and environmentally friendly. Amodei made the comments to CNBC on the morning of Anthropic’s record-breaking funding announcement.
Lawmakers seek pause on new data center construction
Members of Congress are calling for a national pause on data center construction until lawmakers can regulate artificial intelligence. Rep. Jim McGovern of Massachusetts said the White House is not doing enough, while he and others warned that data centers can raise utility prices and harm the environment. Rep.-elect William Lawrence of Michigan said data centers do not offer communities the same value as automotive plants. Rep. Chellie Pingree of Maine called for transparency and strong protections for communities, air, water and climate. It is unclear whether the proposal has enough support to pass in the next Congress.
Lawmakers seek pause on new data center construction
Members of Congress are calling for a national pause on data center construction until lawmakers can regulate artificial intelligence. Rep. Jim McGovern of Massachusetts said the White House is not doing enough, while he and others warned that data centers can raise utility prices and harm the environment. Rep.-elect William Lawrence of Michigan said data centers do not offer communities the same value as automotive plants. Rep. Chellie Pingree of Maine called for transparency and strong protections for communities, air, water and climate. It is unclear whether the proposal has enough support to pass in the next Congress.
Microsoft nears $500 million AI deal with ByteDance, raising OpenAI concerns
Microsoft is close to a $500 million deal to provide ByteDance, the Chinese company behind TikTok, with Azure-based OpenAI models, according to people familiar with the talks. The agreement would expand ByteDance’s existing Microsoft contract, which is already worth more than $1 billion. The models would support ByteDance’s content moderation and recommendation systems. OpenAI and others have raised concerns that the technology could be used to spread misinformation and propaganda. The companies are expected to finalize an agreement in the coming weeks, but neither has commented publicly.
Microsoft ByteDance deal reportedly upsets OpenAI
ByteDance is in advanced discussions to spend about $500 million more on Microsoft cloud services, mainly Azure, according to a report. The deal would expand ByteDance’s existing contract and has reportedly upset OpenAI, which blocks most Chinese companies from using its AI. Microsoft has rights to sell OpenAI’s technology but does not have the same restrictions, the report said. Microsoft is also pursuing deals to sell OpenAI models to other Chinese companies, including Baidu. The article notes that the deal raises concerns about the spread of OpenAI technology and the costs and risks of using AI.
Researchers Find More Than 35,000 Exposed Grok Built Apps
Cybersecurity firm Claroc identified 37,571 apps built with Grok. Researchers said 35,578 were publicly accessible without authentication, pointing to weaknesses in access controls and security practices for AI assisted software development. The exposed apps reportedly included 24 developed by employees of X, the social media platform.
A 27B Language Model Improves Itself by Rewriting Its Weights
Researchers found that a 27B language model can improve itself by editing its own weights through reasoning, without gradients or another training run. Its edit raised accuracy by 6.6 points on 490 problems it had not seen, and the model only kept edits that passed its own tests. The gains stayed in the model’s weights and did not add a cost to each answer. The model also repaired a damaged copy, raising its accuracy on six unseen image benchmarks from 7% to 70%, or 83% of the undamaged model’s score. Tests on four model sizes found that only the 27B model could successfully improve itself; the 2B model made no edit, while the 4B and 9B models damaged themselves.
Watchdog Says AI Platforms Can Provide Terrorist Instructions
An NBC News report dated Nov. 18, 2026, says a watchdog found that AI platforms can provide a terrorist blueprint when prompted. The supplied article text contains no further details about the watchdog’s findings or the platforms tested.
AI and Enterprise Upgrades Set to Shape IT Devices Market Through 2035
The global IT devices market is forecast to grow in value by about 4.2% a year through 2035, reaching a market index of about 151, with 2025 set at 100. Growth is expected to come mainly from higher-value devices and infrastructure, including AI capable laptops, workstations, servers, networking equipment and secure storage, rather than a broad rise in shipments. From 2026 to 2028, enterprise demand is expected to lead as businesses refresh AI PCs, expand server capacity and upgrade networks. Servers and networking are forecast to outperform the overall market, while printers and traditional peripherals face pressure from digitalization. Longer consumer replacement cycles, semiconductor supply, trade restrictions and manufacturing concentration in Asia-Pacific could affect demand, prices and delivery times.
ResearchMonkey AI Bundles More Than 40 Research Tools in One Subscription
ResearchMonkey AI offers students, academics and business analysts more than 40 tools and access to several AI models through one account. Users can summarize documents and videos, find academic papers, draft literature reviews, and conduct market and competitor research; multi-model chat lets users compare answers, while an AI router selects a model for a question. Founder Jerry Gao said the service aims to reduce the need for multiple research software subscriptions. Early access paid plans start at $9.99 per month, with yearly billing about 20% cheaper, and a free plan and seven-day paid-plan trial are also available. The three paid plans add different daily token limits, document storage and summary options, presentation tools, premium models, custom exports, specialized tools and priority support.
Balance Theory buys VectorBlack to expand cybersecurity intelligence
Balance Theory acquired VectorBlack to add market research, competitive analysis and technology evaluation tools to its cybersecurity investment platform. The deal follows Balance Theory’s $19 million Series A funding round, announced in July 2026, and its goal of helping chief information security officers make better technology and spending decisions. The companies did not disclose the financial terms. Balance Theory plans to combine VectorBlack’s market intelligence with customer investment data, product performance measures and security priorities. VectorBlack co-founder Wayne Schepens will become chief cyber market analyst, Jason Lewis will become principal data engineer, and Gerry Eaton will lead targeted research and industry analysis.
Fed rate outlook puts AI companies’ borrowing costs in focus
The Federal Reserve is widely expected to raise interest rates on December 9, 2026, after a hot October inflation report. Trefis found that across 37 rate hikes from June 2004 to July 2023, chip stocks were typically down 1.5% after one month but up 15.6% after a year, with gains after 26 hikes. The article says higher rates could put pressure on major AI companies that spend more than their businesses bring in and rely on borrowing or cash reserves. Investors will be watching the Fed’s outlook for 2027, November inflation data due December 10, and Oracle’s expected mid-December results, though Oracle has not set a report date. The article notes that the 2022 rate hikes coincided with slowing chip sales, and the chip fund fell 17% in the three months after the first hike.
ZeroEyes gun detection system may join Chicago safety pilot
ZeroEyes is in advanced talks with Mayor Brandon Johnson’s administration about joining a Chicago public safety technology pilot that could launch early next year. Its AI system works with existing security cameras to spot visible guns, then has a team verify images and send alerts to security staff. The system is used at Navy Pier and, after a pilot that began in 2024, at all Chicago Transit Authority train stations, where it has led to 30 arrests. ZeroEyes says it does not use facial recognition, record video, or store or train on customer data, and that screenshots are usually deleted within 30 days. City Hall has not confirmed which companies will take part, while City Council members are set to vote on an ordinance about negotiating a new acoustic gunshot detection contract.
Sources
- Anthropic Will ‘Listen To Communities’ On Data Center Plans
- Lawmakers call for national data center construction pause
- Lawmakers call for national data center construction pause
- Microsoft Nears $500 Million Deal to Sell AI to ByteDance, Irking OpenAI
- Microsoft’s $500 Million ByteDance Deal Dismays OpenAI
- 35,000 Grok Apps Exposed: AI Security Risks Escalate
- Language Models Can Rewrite Their Own Weights
- When prompted, AI platforms provide terrorist blueprint, watchdog finds
- IT Devices Market Forecast to 2035: AI-Driven Enterprise Refresh Supports Forward Growth
- ResearchMonkey AI Puts More Than 40 Research Tools and Several AI Models Under One Subscription
- Balance Theory Acquires VectorBlack to Expand AI-Powered Cybersecurity Investment Intelligence
- What Does A Fed Hike Mean For AI Stocks?
- Chicago public safety technology pilot program could include AI-powered gun detection system used at Navy PIer
Comments
Please log in to post a comment.