IBM partners with Nvidia AMD and Microsoft on AI initiatives

BMO’s Fall Business Outlooks, released Nov. 17, 2026, describe AI investment as a growing economic force in both Northern California and Texas. In Northern California, spending by large technology companies on data centers, semiconductors, cloud infrastructure and networking is creating work for energy, construction, manufacturing and professional-services firms. The Bay Area remains a major investment hub, but job growth is more modest; high housing costs and limited labor supply remain constraints. BMO says the gap between investment and hiring will be important to watch in 2027.

Texas is also seeing more investment in AI, cloud computing, semiconductors and energy infrastructure, alongside growth in population, manufacturing and energy. BMO leaders Tony Sciarrino and Scott Anderson highlighted electricity and water availability, permitting, labor supply and housing affordability as issues that could limit growth. Businesses are also monitoring possible 2027 policy changes involving taxes, regulation, trade, energy and infrastructure.

Markets fell on Monday, Nov. 16, as AI-infrastructure stocks led losses. The S&P 500 dropped 1.92% to 7,416.59 and the Nasdaq Composite fell 2.49% to 24,509.76, giving both indexes three consecutive losing sessions; the VIX rose 19%, according to CNN. A separate Manila Times post also attributed the U.S. market decline to AI stocks, led by Nvidia, but provided no further details.

IBM announced Nighthawk quantum processors, a Chicago Quantum System Two deployment and co-design work with Nvidia, AMD, Fermilab and Oak Ridge. It also partnered with Microsoft on an AI agent platform. The initiatives could increase investment needs and pressure margins as IBM links its Software, Consulting and Infrastructure businesses. Investors will look for commercial projects and stronger Software revenue while the company manages high debt; analysts differ on its prospects.

Several companies introduced tools aimed at practical AI use. First Kind AI launched GoMirra, an adults-only service with 18 AI companions whose GOTEL system remembers user preferences and goals across conversations; users control what is retained. It offers voice and text in dozens of languages, web and mobile apps, modes including Devil’s Advocate and Standup Comedy, and Daily Briefings. Launch materials list a monthly price of $18, and the company says conversations are encrypted and can be deleted. Billables AI launched a free iOS app for its customers, letting lawyers dictate draft time entries, review activity from email, calendars, Microsoft Teams and Zoom, or run a timer that creates a billing narrative. Entries can be started with Siri, a widget or the iPhone Action Button. The app syncs across devices and records audio only when the user taps the microphone.

Research and infrastructure tests also focused on improving AI performance. Agent-Guided Process Verification (AGPV) checks a model’s mathematical reasoning claim by claim, stops at the first error and guides it through re-deriving the remaining work. Across five challenging math datasets and four models from DeepSeek, Qwen and GPT, it averaged 63.8% accuracy—13.5 percentage points above the strongest tested baseline. Crusoe, meanwhile, reported that an AI agent serving GLM-5.3 on 64 Blackwell GPUs increased output from 2,188 to 5,114 tokens per second, a 134% gain, and reduced median time to first token from 383 to 293 milliseconds. Crusoe cautioned that different pacing and traffic in the test sweeps make comparisons require care; it attributed most gains to system-layout choices.

Bank-compliance startup Arva AI raised $20 million in a Series A led by Arthur Ventures, joined by existing investors Gradient and Y Combinator. It plans to expand its New York sales team and AI research. Its agents automate sanctions and suspicious-transaction reviews, while banks decide whether to apply changes proposed by its learning system. Arva says one deployment reduced alerts sent to human analysts by 78%, cleared a backlog within 45 days and maintained a false-negative rate below 1% during 90 days of quality assurance.

SpaceX-related chip financing may also reach debt markets this week: some of a planned $40 billion in AI-chip financing could begin selling, with Goldman Sachs and Apollo expected to underwrite it. An Apollo-backed Atlas SP would buy Nvidia chips and issue asset-backed securities, a structure intended to keep the debt off SpaceX’s balance sheet and help preserve its investment-grade credit ratings. Separately, IBM and the UK National Cyber Security Centre are expanding free AI, cybersecurity and digital education for secondary schools through IBM SkillsBuild. The programme includes CyberFirst, a planned Futures course covering machine learning, natural language processing and data science, and the TechFirst Girls Competition; the article says it was set to reach UK schools and students from September 2023.

Key Takeaways

  • BMO says AI and infrastructure investment is creating opportunities in Northern California and Texas, but housing, labor and infrastructure constraints could limit growth.
  • Northern California’s strong technology investment has outpaced more modest job growth, a gap BMO says will be worth watching in 2027.
  • On Nov. 16, 2026, the S&P 500 fell 1.92% and the Nasdaq dropped 2.49%; a Manila Times post separately cited AI-stock declines led by Nvidia.
  • IBM announced Nighthawk quantum processors, a Chicago Quantum System Two deployment, projects with Nvidia and AMD, and an AI agent collaboration with Microsoft.
  • GoMirra launched 18 AI companions with user-controlled memory, voice and text in dozens of languages, and an $18 monthly subscription listed in launch materials.
  • Billables AI’s free iOS app lets lawyers create and review time entries, including through Siri, and syncs entries across mobile, desktop and web.
  • AGPV scored 63.8% across five math datasets and four models from DeepSeek, Qwen and GPT, beating the strongest tested baseline by 13.5 percentage points.
  • Crusoe reported a 134% increase in GLM-5.3 serving output on 64 Blackwell GPUs, while cautioning that test conditions varied.
  • Arva AI raised $20 million; it reports one deployment cut human-reviewed alerts by 78%, cleared a backlog in 45 days and kept its false-negative rate below 1% over 90 days.
  • SpaceX chip financing may involve up to $40 billion, with Nvidia chips purchased through an Apollo-backed structure intended to keep debt off SpaceX’s balance sheet; IBM and the UK NCSC are also expanding free school courses in AI and cybersecurity.

BMO says AI investment is reshaping Northern California’s economy

BMO released its Fall Business Outlook for Northern California on Nov. 17, 2026. It says strong company profits and investment in data centers, semiconductors, cloud infrastructure and networking are creating opportunities in energy, construction, manufacturing and professional services. The report says Bay Area investment by large technology companies remains a major economic driver, while job growth has been more limited. Housing costs and a constrained labor force remain challenges, and BMO says the gap between strong investment and modest hiring will be important to watch in 2027.

Texas businesses weigh growth as AI, energy and infrastructure spending rise

BMO released its Fall Business Outlook for Texas on Nov. 17, 2026. It says population growth, manufacturing, energy and business investment support the state’s economy, while spending on AI and infrastructure creates more opportunities. Texas is also investing in AI, cloud computing, semiconductors and energy infrastructure, though labor and housing limits could slow growth. Businesses are watching policy changes ahead of 2027, especially those affecting taxes, regulation, trade, energy and infrastructure. BMO leaders Tony Sciarrino and Scott Anderson also pointed to electricity, water, permitting, labor availability and housing affordability as important issues.

BMO sees AI investment spreading across Northern California’s economy

BMO released its Fall Business Outlook for Northern California on Nov. 17, 2026, describing AI as a broader economic force beyond technology companies. Investment in data centers, semiconductors, cloud infrastructure and networking is creating work for businesses in energy, construction, manufacturing and professional services. BMO says Northern California remains an innovation hub, but high housing costs and a limited labor force pose challenges. Large technology companies are investing heavily in the Bay Area, while job growth has been more modest. BMO’s Scott Anderson said this gap between investment and hiring will be important to watch in 2027.

Nasdaq falls 2.5% as AI infrastructure stocks lead market losses

Stocks tied to artificial intelligence infrastructure led Wall Street lower on Monday, Nov. 16, 2026. The S&P 500 fell 1.92% to 7,416.59, and the Nasdaq Composite dropped 2.49% to 24,509.76. Both indexes recorded a third straight losing day. CNN reported that the VIX, a measure of expected market swings, rose 19%.

AI stock declines, led by Nvidia, drag US market lower

A post from The Manila Times says declines in artificial intelligence stocks dragged the U.S. market lower on Monday, led by Nvidia. The post does not provide further details and directs readers to find the full story in the comments.

IBM’s Quantum Push Could Reshape Its Investment Outlook

International Business Machines announced new Nighthawk quantum processors, a Chicago Quantum System Two deployment, and co-design projects with NVIDIA, AMD, Fermilab, and Oak Ridge. It also announced an AI agent platform collaboration with Microsoft. These efforts could raise investment needs and affect margins as IBM connects its Software, Consulting, and Infrastructure businesses. Investors will watch whether IBM can turn its quantum and AI work into commercial projects and stronger Software revenue while managing its high debt. Analysts disagree about the company’s future, with some expecting slower growth and others more optimistic.

GoMirra Offers 18 AI Companions With Memory Across Conversations

First Kind AI, founded by Tony Nelson and Oliver Rozynski, launched GoMirra for adults aged 18 and older. Its GOTEL system uses an Identity Map and Behavioural Map to remember users’ interests, goals, communication preferences, and changing patterns, and users can control what it remembers or forgets. Subscribers can consult 18 AI companions with different personalities, use modes such as Devil’s Advocate and Standup Comedy, and receive Daily Briefings connected to their goals. The service supports voice and text in dozens of languages and is available on the web, iOS, and Android. Launch materials list an $18 monthly subscription; GoMirra says conversations are encrypted and users can delete them.

Billables AI Launches iOS App for Lawyers’ Mobile Timekeeping

Billables AI announced its iOS mobile app on November 16, 2026, making it available to all customers as a free download. Attorneys can dictate work to create draft time entries, review and approve entries captured from email, calendars, Microsoft Teams, and Zoom, or use a timer that creates a billing narrative. They can start voice entries through Siri, a Home Screen widget, or the iPhone Action Button. The app syncs entries across mobile, desktop, and web, and requires no additional setup for existing customers. It records audio only when users tap the microphone and does not monitor phone activity.

AGPV Helps AI Models Find and Fix Errors in Math Solutions

Researchers proposed Agent-Guided Process Verification, or AGPV, to help large language models correct errors in complex math reasoning. It breaks a solution into key claims, checks them one by one, and stops at the first error; a diagnosis then guides the model to re-derive the remaining work. The method also checks that a solution is complete and reuses successful tool results. Tests on five challenging math datasets with four models from DeepSeek, Qwen, and GPT produced an average accuracy of 63.8%. That was 13.5 percentage points higher than the strongest evaluated baseline, which scored 50.3%.

Arva AI Raises $20 Million to Grow Bank Compliance Tools

Arva AI raised $20 million in a Series A funding round led by Arthur Ventures, with existing investors Gradient and Y Combinator also taking part. The bank-compliance startup plans to expand its New York sales team and increase AI research. Its agents automate sanctions and suspicious transaction reviews for large banks, while customers decide whether to deploy changes suggested by its continuous-learning system. Arva reported that one deployment cut alerts sent to human analysts by 78%, cleared the backlog within 45 days, and kept its false-negative rate below 1% during a 90-day quality-assurance period. The company aims to win approvals at large banks and add more workflows to each deployment.

SpaceX chip financing may reach debt markets this week

Some of the $40 billion in financing for artificial intelligence chips for SpaceX may begin selling as soon as this week. Goldman Sachs Group Inc. and Apollo Global Management Inc. are expected to underwrite the deal. An Apollo-backed Atlas SP will buy Nvidia Corp. chips and issue asset-backed securities, keeping the debt off SpaceX’s balance sheet. This approach is intended to help SpaceX maintain its investment-grade credit ratings.

Crusoe uses AI to boost GLM-5.3 serving speed by 134%

Crusoe tested an AI agent on 64 Blackwell GPUs and raised output from 2,188 to 5,114 tokens per second, a 134% increase. The test also cut median time to first token from 383 milliseconds to 293 milliseconds. Crusoe’s Martin Cala ran the trials with Andrey Cheptsov of dstack, who provided the test setup. Cala said most of the improvement came from system layout choices, such as how to split prefill and decode and where to place the KV cache. The company cautioned that the test sweeps used different pacing and traffic, so their results need careful comparison.

IBM and UK cyber agency expand free school courses in AI and cybersecurity

IBM and the National Cyber Security Centre are expanding free cybersecurity, AI and digital education for UK secondary schools through IBM SkillsBuild. The programme, announced in 2022, includes CyberFirst, which teaches cybersecurity basics such as threat analysis, incident response and security protocols. A planned Futures course will cover machine learning, natural language processing and data science. The programme also includes the TechFirst Girls Competition for girls in UK secondary schools. It was set to reach schools across the UK and be available to students from September 2023.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI investment AI infrastructure data centers semiconductors cloud computing Northern California Texas economy energy infrastructure labor and housing constraints financial markets IBM quantum computing AI agents AI companions legal technology AI reasoning verification GPU performance AI compliance automation AI chip financing AI education

Comments

Loading...