Enterprise AI News This Week: November 13, 2026

The biggest change for businesses this week is that the AI vendors are moving from selling models to selling places where work happens. Anthropic shipped a new flagship model plus spreadsheets, whiteboards and a meeting agent, OpenAI opened a new purchasing channel through AWS, and DeepMind showed how AI agents are making old cryptography cheaper to attack. This week in AI is a procurement and governance story as much as a technology one.

Here is the short version, followed by what each item means for a business in plain words.

Item	What launched	Who it matters for	Promact take
Claude Fable 5.2	New top Anthropic model, $8/$40 per million tokens	Teams building on Claude	Watch
Claude Whiteboards, Sheets and Data	Shared AI workspaces in Claude, in beta	Team and Enterprise admins	Act now (set policy)
Claude Meetings	Claude joins Google Meet calls (beta)	Teams on Google Workspace	Watch
ChatGPT Enterprise on AWS Marketplace	Buy ChatGPT Enterprise against AWS commitments	Procurement and finance	Act now
DeepMind and legacy RSA	AI agents cut attack cost on 1024-bit RSA	Security and compliance	Act now (audit keys)
Anthropic IPO roadshow	Price range of $70 to $78 per share	Anyone betting on Claude	Watch

Claude Fable 5.2: a stronger model at the same top-tier price

Anthropic launched Claude Fable 5.2 and the restricted Mythos 5.2 on November 9. According to Anthropic’s launch post, standard pricing is $8 per million input tokens and $40 per million output tokens, with an optional fast mode at $16 and $80 that speeds up token generation by up to 2.5 times. Fable 5.2 is available through Anthropic’s own platform and on AWS, Google Cloud and Microsoft Azure, while Mythos 5.2 is limited to vetted cybersecurity and life-science organisations.

Anthropic says the models are “a step above Opus 5.5 in almost all the capabilities we evaluated,” and reports 69.9% on SWE v2 against 62.3% for Opus 5.5. These are vendor-reported results, so treat them as a reason to run your own test, not as a purchasing decision.

What it means for you: if you already use Claude for coding or analysis agents, put Fable 5.2 on a small evaluation set of your real tasks before switching. Measure cost per completed task, not price per token.

Claude Whiteboards, Sheets and Data: AI workspaces arrive by default

On November 10, Anthropic launched Claude Whiteboards and Claude Sheets in beta and made Claude Dashboards generally available as Claude Data. They work on all paid plans. On Team plans all three are on by default. On Enterprise plans Whiteboards and Sheets are off by default, and Claude Data turns on by default on November 19, 2026.

What it means for you: the risk is not the features, it is the defaults. A tool that switches itself on in a Team plan can start holding company data before anyone has written a rule for it. Admins should review the settings under Organization settings before the November 19 date. This is the same pattern we described in vendor-embedded AI becoming a governance blind spot.

Claude Meetings: an AI participant in your calls

Axios reported that Anthropic is launching Claude Meetings in beta for Google Meet on Team and Enterprise plans. Claude joins as a visible participant, announces itself in the call chat, and drafts plans and action items while people talk. Reports say it disconnects automatically if an external guest joins. Anthropic’s own pages describe the feature only briefly, so check the current plan and region availability before promising it to your teams.

What it means for you: meeting notes are some of the most sensitive text a company produces. Decide who can invite the agent, where the notes are stored and how long they are kept before your teams start using it.

ChatGPT Enterprise on AWS Marketplace: a procurement shortcut

On November 9, OpenAI announced that ChatGPT Enterprise can be bought through AWS Marketplace. Customers buy through their existing AWS billing, and eligible spend can count toward AWS commitments. OpenAI says AWS customers get the same models, features and functionality as buying directly. Pricing was not disclosed, and eligibility is confirmed through OpenAI or AWS account teams.

What it means for you: if you hold an AWS spending commitment, this can make an AI purchase easier to approve, because the money is already budgeted. It is a billing channel, not a data-residency guarantee, so ask your account team where data is processed. It also fits the multi-vendor direction we covered in IBM’s OpenAI partnership. Keep your architecture able to switch models.

DeepMind and legacy RSA: older keys just got cheaper to attack

On November 6, Google DeepMind said human-supervised AI agents found two weaknesses in legacy 1024-bit RSA. As SecurityWeek reported, the results apply only to 1024-bit keys, not to 2048-bit or larger keys. The measured result was a roughly 3.5 times CPU saving when reproducing a known signature forgery. The larger estimates, such as about 30 times cheaper batch factoring, are theoretical.

What it means for you: this is not a break of modern encryption. It is a reminder to find any 1024-bit RSA keys still lingering in old systems, certificates and embedded devices, and retire them. Legacy estates are where this kind of finding bites.

Anthropic’s IPO roadshow: your vendor is becoming a public company

RTTNews reported that Anthropic launched its IPO roadshow on November 10, with an expected range of $70 to $78 per share for 825 million shares and a planned Nasdaq listing under the ticker ANT. Pricing and trading dates reported by the press are targets, not confirmed dates.

What it means for you: a listing adds public financial disclosure, which helps vendor due diligence. It does not change your need for exit options. Our earlier piece on what going public means for enterprises betting on Claude covers the questions to ask.

What to do this week

1. Audit defaults. List every AI feature your Claude, Microsoft and Google plans switch on automatically, and set a written policy before November 19. Our AI agent governance framework is a starting point.
2. Check your AWS commitment. Ask procurement whether AI spend can draw down existing cloud commitments, and what the data-handling terms are.
3. Scan for 1024-bit RSA. Ask your security team for an inventory of old keys and certificates.

If you want help turning these into a working agent strategy, see our AI agent solutions service.

Frequently Asked Questions

What is the biggest enterprise AI news this week?

The biggest shift is vendors bundling AI into workspaces and purchasing channels: Anthropic’s new model and workspace tools, and ChatGPT Enterprise on AWS Marketplace.

Is Claude Meetings generally available?

Reports describe it as a beta for Google Meet on Team and Enterprise plans. Confirm current availability with Anthropic before planning a rollout.

Does the DeepMind result break RSA encryption?

No. It concerns legacy 1024-bit RSA only. Keys of 2048 bits or larger were not affected, and several of the larger figures are theoretical estimates.

Does buying ChatGPT Enterprise through AWS change the product?

OpenAI says customers get the same models, features and functionality. The change is in how you buy and bill, not in what the product does.

Should we switch to Fable 5.2 now?

Test it first. The benchmark figures are vendor-reported, so compare cost and quality on your own tasks.

Sources

• Anthropic: Claude Fable 5.2 and Mythos 5.2 - launch post with pricing, availability and benchmarks
• Anthropic: Claude Whiteboards, Sheets and Data - launch details and plan defaults
• Axios: Anthropic wants Claude to join your meetings - coverage of Claude Meetings
• OpenAI: ChatGPT Enterprise on AWS Marketplace - announcement of the new purchasing channel
• SecurityWeek: DeepMind AI agents make 1024-bit RSA keys cheaper to attack - report on the RSA findings
• RTTNews: Anthropic sets IPO price range and launches roadshow - IPO terms

Tracked with help from World of AI.