What should enterprise leaders know about AI today?
On November 5, 2026, US prosecutors charged three men with using AI-cloned executive voices to steal about $129 million from companies. On November 6, OpenAI was reported to have raised $11 billion at a $1.25 trillion valuation, Microsoft was reported to be winding down rented Oracle AI capacity, and Google DeepMind said its AI agents found two weaknesses in legacy 1024-bit RSA. Leaders should harden payment approvals now and treat AI capacity and old cryptography as assets to re-check.
Executive takeaways
- Act nowAI Security: AI voice and video cloning of executives is now a charged, nine-figure fraud pattern; CFOs should require out-of-band verification for every urgent, confidential wire request.
- EvaluateAI Investments & Initiatives: OpenAI’s $11 billion first close came below the valuation it reportedly sought, a sign that buyers can negotiate harder on multi-year AI commitments without fearing for the vendor’s funding.
- MonitorAI Infrastructure: Microsoft retiring older rented GPU capacity shows AI compute ages fast; CIOs should check that contracted capacity runs on current chips.
US prosecutors charge three over a $129 million AI deepfake CEO fraud, putting payment controls to the test
AI Security · Associated Press, Gizmodo, November 5–7, 2026 Priority: Act now
For: Financial Services, Healthcare & Life Sciences
What changed?
Federal prosecutors in Manhattan charged three men on November 5, 2026, with an AI impersonation scheme that allegedly took about $129 million. From July 2024 to February 2026 the group targeted thousands of companies, posing as CEOs by email and then using AI to mimic the executive’s voice and sometimes image on calls (AP via SecurityWeek).
A fake “lawyer” from a large law firm then gave wire instructions for a supposedly confidential acquisition. Named losses include about $4.83 million at a Missouri hospital system, $2.68 million at a Rhode Island bank and $11.58 million at a UK fintech; a New York investment bank’s back office stopped a $2.85 million wire an employee had already approved (Gizmodo). The charges are allegations, not convictions.
Why does this matter for enterprises?
The attack used no technical exploit. It relied on secrecy, urgency and a convincing executive voice, which AI now makes cheap, so the defence is process, not detection software.
Hospitals, banks and fund managers appear among the victims because their finance teams move large wires routinely. This does not mean video calls are unusable; it means a call can no longer serve as proof of identity for a payment.
| Enterprise function | Business implication |
|---|---|
| Finance | Wire approvals need a second channel and a second approver for urgent, confidential requests |
| Legal | Deal confidentiality claims should trigger verification, not bypass it |
| Internal audit | Test whether a back office can stop an already-approved payment |
What should leaders do next?
Decision for CFOs, CISOs and general counsel: Can any employee release a large payment on the strength of a call from an executive or outside lawyer?
Within 30 days, require call-back verification through a known number and dual approval for wires above a set threshold, and run one simulated deepfake request against the finance team.
Sources: AP via SecurityWeek — 3 charged in AI CEO impersonation scheme · Gizmodo — Deepfaked CEOs and lawyers helped steal $129 million
OpenAI raises $11 billion at a $1.25 trillion valuation, below the price it reportedly sought
AI Investments & Initiatives · Reuters, Analytics Insight, November 6, 2026 Priority: Evaluate
What changed?
OpenAI has closed about $11 billion as the first part of a round valuing it at $1.25 trillion, Bloomberg News reported on November 6, 2026. Reuters relayed the report, which cites people familiar with the matter; OpenAI has not confirmed the terms and investors were not named (Reuters via GV Wire).
The valuation is below the $1.5 trillion OpenAI reportedly sought, and the round could add about $9 billion more. OpenAI also said it will keep expanding data centers it owns while paying their full power costs (Analytics Insight).
Why does this matter for enterprises?
Fresh capital lowers the near-term risk that a major supplier cuts service or capacity, which matters for anyone signing multi-year ChatGPT or API commitments.
A lower-than-sought price signals investors want proof of returns, the same pressure buyers feel. This does not mean OpenAI is short of money; it means pricing power is shifting toward customers who can show alternatives.
What should leaders do next?
Decision for CIOs and CFOs: Are our OpenAI commitments priced and structured as if the vendor holds all the leverage?
Within 60 days, benchmark renewal terms against at least one alternative model provider and seek price protection and exit clauses before signing multi-year deals.
Sources: Reuters via GV Wire — OpenAI secures $11 billion · Analytics Insight — OpenAI secures $11 billion as it expands data centers
Microsoft winds down rented Oracle AI capacity, showing how fast AI compute ages
AI Infrastructure · Reuters, November 6, 2026 Priority: Monitor
What changed?
Microsoft is winding down a 2023 deal to rent AI computing from Oracle, Bloomberg News reported on November 6, 2026. The capacity runs on older Nvidia chips; Oracle said demand for its AI infrastructure “remains strong, including for earlier-generation GPUs” (Reuters via Investing.com).
The same report says IREN delayed some deliveries after Microsoft requested network design changes. IREN moved its target for bringing over $4 billion in annualized revenue capacity online from year-end to January, possibly late February.
Why does this matter for enterprises?
Older GPUs lose value fast. Enterprises reserving dedicated capacity face the same risk of paying for hardware that no longer fits their workloads.
This does not mean AI capacity is in surplus: Microsoft is still tightening its neocloud projects, not exiting them. It means contract terms should track chip generations.
What should leaders do next?
Decision for CIOs and CFOs: Do our AI capacity contracts let us move to newer hardware without paying twice?
Within 90 days, review reserved GPU and AI capacity commitments for chip generation, refresh rights and term length, and avoid locking new deals beyond two years without upgrade clauses.
Sources: Reuters via Investing.com — Microsoft winding down Oracle AI capacity deal
Google DeepMind’s AI agents find two weaknesses in legacy RSA-1024, speeding up crypto retirement
AI Security · OfficeChai, Crypto Briefing, November 6, 2026 Priority: Evaluate
What changed?
Google DeepMind said on November 6, 2026, that its AI agents, working under human direction, found two weaknesses in legacy 1024-bit RSA. One makes signature forgery easier when an attacker controls much of the signed data; the other cuts the average cost of factoring each additional 1024-bit key by about 30 times (OfficeChai).
Coverage stresses that 2048-bit and larger RSA keys remain secure against these methods. DeepMind’s headline 1,600-times improvement is a theoretical estimate, and the measured result was tested on an already-solved benchmark key (Crypto Briefing).
Why does this matter for enterprises?
DeepMind says its agents made these advances in weeks, so weak keys that seemed safe enough in legacy systems carry rising risk.
This does not mean modern encryption is broken. The exposure sits in old devices, signing systems and partner integrations still using 1024-bit keys.
What should leaders do next?
Decision for CISOs and CTOs: Do we know where 1024-bit RSA keys still protect signatures or data in our estate?
Within 90 days, inventory certificates and signing keys below 2048 bits, set a retirement date for each, and fold the work into the post-quantum migration plan.
Sources: OfficeChai — DeepMind agents and legacy RSA-1024 · Crypto Briefing — DeepMind AI finds RSA-1024 weaknesses
What should enterprise leaders do next?
| Action | Owner | Timeframe | Success measure |
|---|---|---|---|
| Harden payment verification against AI impersonation | CFO with CISO | 30 days | 100% of wires above threshold verified by call-back and dual approval, and one deepfake simulation run |
| Renegotiate AI vendor terms with alternatives in hand | CIO and CFO | 60 days | every multi-year AI contract has price protection and an exit clause |
| Retire weak keys and stale capacity | CISO and CTO | 90 days | complete inventory of sub-2048-bit keys with retirement dates, and capacity contracts reviewed for upgrade rights |
The bigger picture
- AI lowers the cost of attack in very different places, from cloned voices in the deepfake fraud to cheaper factoring in DeepMind’s research; the answer in both is process and hygiene, not new tools.
- Capital and compute are both being repriced: OpenAI’s valuation came in below its ask, and Microsoft is shedding older GPUs, so buyers gain leverage over price and hardware terms.
Executive FAQ
How can companies stop AI deepfake CEO fraud?
Verify any urgent, confidential payment request through a separate known channel and require two approvers. In the $129 million case charged on November 5, 2026, a bank’s back office stopped one wire an employee had already approved.
Does OpenAI’s $11 billion raise change enterprise AI contract negotiations?
It lowers the risk of supplier disruption, while the $1.25 trillion valuation, below the reported $1.5 trillion target, suggests buyers can push for price protection and exit terms.
Is RSA encryption still safe after DeepMind’s AI findings?
RSA keys of 2048 bits or more remain secure against the methods DeepMind described on November 6, 2026; legacy 1024-bit keys should be retired.
About the curator & AI disclosure
Curated by Luciano Ayres
Luciano Ayres is Director of Artificial Intelligence at ArcTouch, with more than 20 years of experience in software and technology leadership and specialized studies in deep learning. He focuses on helping organizations connect AI capabilities with practical business outcomes, responsible adoption, and scalable implementation. Connect with him on LinkedIn to discuss enterprise AI opportunities, strategy, investment decisions and risks.
AI-generated briefing. This issue was produced through an automated AI research and writing workflow curated by Luciano Ayres. Sources are linked for independent verification; summaries and analysis may contain errors and do not replace professional due diligence.