Transcript: Anthropic Makes Its Move

Source: Tech Brew Ride Home | Duration: 19 min

Summary

Here is a comprehensive summary of the "Anthropic Makes Its Move" podcast episode:

Opening context:

• This episode covers the latest news and developments in the AI industry, with a focus on Anthropic's record-breaking $30 billion funding round and the rapid growth of other AI companies like Cohere.

Key discussion points and insights:

• Anthropic has raised a $30 billion Series G round, giving it a $380 billion post-money valuation - the second largest private tech fundraising round on record after OpenAI.
• Anthropic's annualized revenue has hit $14 billion, growing over 10x annually in the past 3 years. Its AI coding tool, CloudCode, has seen its revenue grow to over $2.5 billion.
• Cohere, another AI company, hit $240 million in annual recurring revenue last year, with over 50% quarter-over-quarter growth. Its capital-efficient model and high gross margins (around 70%) have allowed it to invest aggressively in customer acquisition and R&D.
• Meta (Facebook) plans to add facial recognition capabilities to its smart glasses, despite privacy concerns. The company sees the current "political tumult" as an opportunity to launch the controversial feature.
• Amazon's Ring has cancelled its partnership with Flock Safety, which would have allowed law enforcement agencies to request footage from Ring doorbell users, following significant public backlash.

Notable technologies, tools, or concepts mentioned:

• Anthropic's CloudCode AI coding tool
• Cohere's AI models and software tools for businesses
• Meta's facial recognition plans for its smart glasses
• Ring's integration with Flock Safety for law enforcement access to user footage

Practical implications or recommendations discussed:

• The AI industry is seeing exponential growth, with some companies like Anthropic and Cohere achieving hockey-stick revenue trajectories.
• The ethical concerns around facial recognition technology and law enforcement access to private surveillance footage remain a highly contentious issue.
• The use of AI in creative industries like romance novel writing is disrupting traditional publishing models, with some authors embracing the technology while others oppose it.
• The rapid advancement of AI capabilities, such as the potential for ChatGPT to generate on-demand creative content, raises fundamental questions about the future of art and human creativity.

Full Transcript

[00:00:00] Welcome to the TechBrew right home for Friday the 13th of February 2026. I'm Brian McCullough.

[00:00:09] Today, Anthropic raises the second largest financing round of all time. Other AI players

[00:00:14] are beginning to show hockey stick revenue growth. Meta wants to add facial recognition

[00:00:18] to its glasses. Ring pulls back from some recognition partnerships for its cameras.

[00:00:23] And of course, your weekend long read suggestions. Here's what you missed today in the world of tech.

[00:00:30] anthropic has raised a 30 billion dollar series g round led by gic and co2 and co-led by d.e shaw

[00:00:40] dragon near founders fund iconic and mgx at a 380 billion dollar post money valuation quoting cnbc

[00:00:48] open ai has the largest private tech fundraising round on record rival anthropic now has the second

[00:00:54] largest anthropic announced on thursday the close of a 30 billion dollar funding round at a 380

[00:00:59] billion post-money valuation, more than double what the artificial intelligence company was worth

[00:01:04] in September when it last raised money. Anthropix said the round includes a portion of the previously

[00:01:09] announced investments from Microsoft and NVIDIA, which said in November that they plan to commit

[00:01:13] up to $5 and $10 billion, respectively. Anthropix's annualized revenue has climbed to $14 billion,

[00:01:19] the company said, after revenue last year reached roughly $10 billion. Whether it is entrepreneurs,

[00:01:25] startups, or the world's largest enterprises, the message from our customers is the same.

[00:01:28] Claude is increasingly becoming more critical to how businesses work, Anthropics CFO Krishna Rao

[00:01:34] said in a statement. This fundraising reflects the incredible demand we are seeing from these

[00:01:38] customers. The company said the Fresh Capital will support Anthropics infrastructure, expansion,

[00:01:43] research, and its continued investment in enterprise-grade products. OpenAI is also

[00:01:48] engaged in fundraising talks with investors for a round that could close at around $100 billion,

[00:01:52] dollars, as CNBC previously reported. The company has to pad its cash position after inking $1.4

[00:01:59] trillion worth of infrastructure deals last year. Anthropic gets about 80% of its business from

[00:02:05] enterprises, CEO Dario Amadai told CNBC last month. That's partly thanks to the company's viral AI

[00:02:11] coding tool, CloudCode, which can automate parts of the software development process. CloudCode's

[00:02:17] annualized revenue has increased to $2.5 billion, and business subscriptions have quadrupled since

[00:02:22] the start of the year, Anthropic said Thursday. Enterprise users represent more than half of

[00:02:27] Claude Code's revenue, end quote. Just to underline that a bit, Anthropic says its run rate revenue

[00:02:34] has hit $14 billion, growing over 10x annually in each of the past three years, and Claude Code's

[00:02:41] revenue has grown to more than $2.5 billion. We are, folks, seeing revenue in some of these places.

[00:02:48] Go hockey stick.

[00:02:52] Here's another data point to that end, also from CNBC.

[00:03:01] Cohere hit roughly $240 million in annual recurring revenue last year, surpassing its $200 million target, according to a February investor memo viewed by CNBC.

[00:03:12] It saw quarter-over-quarter growth of more than 50% throughout 2025, the memo said.

[00:03:17] Our thesis is clearly resonating in the market, the company wrote.

[00:03:20] our sales pipeline continues to grow as global organizations across regulated sectors choose

[00:03:25] Cohear as their trusted partner for secure AI adoption at scale. Founded in Toronto in 2019,

[00:03:32] Cohear develops models and builds software tools for businesses. The company is backed by investors

[00:03:37] including NVIDIA and Salesforce Ventures and its valuation has swelled to roughly $7 billion.

[00:03:42] Cohear's investor memo comes after CEO Aiden Gomez said in October that the startup hopes to

[00:03:48] its public market debut, quote, soon. He told Bloomberg that he thinks investors would welcome

[00:03:53] a, quote, pure play AI investment opportunity. Cohere told investors that its capital-efficient

[00:03:58] model sets it apart from its rivals in the industry. The company primarily generates

[00:04:02] revenue from software and it said it can avoid hefty infrastructure costs because customers can run its models through their managed cloud services or on their own hardware directly This approach allows Cohere to invest more aggressively in customer acquisition and research and development

[00:04:17] according to its investor memo. Cohere's gross margins averaged around 70% in 2025,

[00:04:23] expanding by 25 basis points year over year, the memo said. By scaling compute resources

[00:04:29] proportionally to customer demand, we remain insulated from the speculative excesses

[00:04:33] surrounding the broader AI market, positioning Cohere for more sustainable growth, Cohere wrote.

[00:04:45] The New York Times has seen a memo suggesting that Meta plans to add facial recognition to

[00:04:50] its smart glasses later this year, saying the political tumult in the US could distract critics

[00:04:57] from the feature's release. Quote, five years ago, Facebook shut down the facial recognition

[00:05:02] system for tagging people and photos on its social network, saying it wanted to find the

[00:05:06] right balance for a technology that raises privacy and legal concerns. Now it wants to bring facial

[00:05:11] recognition back. Meta, Facebook's parent company, plans to add the feature to its smart glasses,

[00:05:16] which it makes with the owner of Ray-Ban and Oakley, as soon as this year, according to four

[00:05:21] people involved with the plans who were not authorized to speak publicly about confidential

[00:05:25] discussions. The feature, internally called NameTag, would let wearers of smart glasses

[00:05:30] identify people and get information about them via Meta's artificial intelligence assistant.

[00:05:34] Meta's plans could change. The Silicon Valley company has been conferring since early last

[00:05:39] year about how to release a feature that carries, quote, safety and privacy risks,

[00:05:43] according to an internal document viewed by the New York Times. The document from May described

[00:05:48] plans to first release NameTag to attendees of a conference for the blind, which the company did

[00:05:53] not do last year before making it available to the general public. Meta's internal memo said,

[00:05:58] the political tumult in the United States was good timing for the feature's release. We will launch

[00:06:03] during a dynamic political environment where many civil society groups that we would expect to attack

[00:06:08] us would have their resources focused on other concerns, according to the document from Meta's

[00:06:13] Reality Labs, which works on hardware including smart glasses. Facial recognition technology has

[00:06:18] long raised civil liberty and privacy concerns for its potential use by governments to monitor

[00:06:22] citizens and suppressed dissent, by corporations to track unwitting customers, or by creeps at bars.

[00:06:30] Some cities and states have restricted or banned use of the technology by the police over concerns

[00:06:35] about its accuracy. Meta is exploring who should be recognizable through the technology, two of the

[00:06:40] people said. Possible options include recognizing people a user knows because they are connected on

[00:06:45] a meta platform, and identifying people whom the user may not know but who have a public account

[00:06:50] on a meta site like Instagram. The feature would not give people the ability to look up anyone they

[00:06:56] encountered as a universal facial recognition tool to people familiar with the plan said.

[00:07:00] Meta smart glasses require a wearer to activate them to ask the AI assistant a question or to take

[00:07:06] a photo or video. The company is also working on glasses internally called a super sensing

[00:07:11] that would continually run cameras and sensors to keep a record of someone's day similar to how AI

[00:07:16] notetakers summarize video call meetings, three people involved with the plan said. Facial

[00:07:21] recognition would be a key feature for super-sensing glasses so they could, for example,

[00:07:26] remind wearers of tasks when they saw a colleague. Mr. Zuckerberg has questioned if the glasses

[00:07:31] should keep their LED light on to show people they are using the super-sensing feature or if

[00:07:35] they should use another signal, one person involved in the plan said. As part of an FTC settlement,

[00:07:42] Meta in the past has agreed to review every new or modified product for potential risks to the

[00:07:48] privacy of the company's users. In January 2025, Meta relaxed that process for reviewing privacy

[00:07:53] risks according to an internal post viewed by the Times. The company's privacy teams have less

[00:07:58] influence over product releases, and there are new limits on how long the risk review process takes.

[00:08:04] Around that time, employees who worked on risk review questioned whether Meta

[00:08:07] would still be in compliance with its FTC settlement under the changes.

[00:08:12] Andy Millen a director of risk review in Reality Labs told them that she believed the changes could quote push the bounds of Meta agreement with the FTC according to a recording of an internal meeting obtained by the Times

[00:08:24] Mark wants to push on it a little bit, Ms. Millen said, referring to Mr. Zuckerberg, end quote.

[00:08:30] picture this you've been chosen to lead ai onboarding for your team that includes everyone

[00:08:44] from seasoned technical pros to ai newcomers no pressure right that's where area comes in

[00:08:51] areas no code low code and pro code platform makes it easy for your organization to embrace

[00:08:56] AI regardless of technical experience. With ARIA's unified security layer, advanced threat

[00:09:02] detection, and robust compliance measures, your AI ecosystem can stay safe while your team

[00:09:08] innovates. Worried about staying on budget? ARIA's cost optimization tools can help you

[00:09:13] manage and forecast your AI spend so there are no surprises. Get started at ARIA.com slash

[00:09:20] morningbrew. That's ARIA.com slash morningbrew.

[00:09:27] Amazon's Ring has cancelled its partnership with Flock Safety, which would have let law

[00:09:33] enforcement agencies request footage from Ring doorbell users after a backlash, quoting The Verge.

[00:09:39] Following intense backlash to its partnership with Flock Safety, a surveillance technology

[00:09:43] company that works with law enforcement agencies, Ring has announced it is cancelling the integration.

[00:09:48] In a statement published on Ring's blog and provided to The Verge ahead of publication,

[00:09:51] the company said, quote, following a comprehensive review, we determined the

[00:09:55] planned Flock Safety integration would require significantly more time and resources than

[00:09:59] anticipated. We therefore made the joint decision to cancel the integration and continue with our

[00:10:03] current partners. The integration never launched, so no Ring customer videos were ever sent to Flock

[00:10:08] Safety, end quote. The statement goes on to say that Ring's mission to make neighborhoods safer,

[00:10:13] quote, comes with significant responsibility to our customers, to the communities we serve,

[00:10:17] and to the trust you place in our products and features. Trust is the big one there. Over the

[00:10:22] last few weeks, the company has faced significant public anger over its connection to Flock,

[00:10:26] with Ring users being encouraged to smash their cameras and some announcing on social media that

[00:10:31] they are throwing away their Ring devices. The Flock partnership was announced last October,

[00:10:35] but following recent unrest across the country related to ICE activities, public pressure against

[00:10:40] the Amazon-owned Ring's involvement with the company started to mount. Flock has reportedly

[00:10:44] allowed ICE and other federal agencies to access its network of surveillance cameras,

[00:10:48] and influencers across social media have been claiming that Ring is providing a direct link

[00:10:53] to ICE. While that claim probably isn't accurate as the flock integration has never gone live,

[00:10:58] Ring has a history of partnering with police and the new partnership quickly came under intense

[00:11:03] criticism. Adding fuel to the fire, this weekend Ring aired a Super Bowl ad for its new AI-powered

[00:11:08] search party feature. While the company says the feature is designed to find lost dogs and

[00:11:12] maintains it is not capable of finding people, the ad raised fears that Ring cameras were being used

[00:11:18] for mass surveillance. The ad shows dozens of Ring cameras in a neighborhood scanning the streets.

[00:11:23] On top of this, the company recently launched a new facial recognition feature called Familiar

[00:11:27] Faces. Combined with Search Party, the technological leap to using neighborhood

[00:11:31] cameras to search for people through a mass surveillance network suddenly seems very small.

[00:11:36] Flock was the second partner Ring announced for community requests, the first being Axon,

[00:11:41] a law enforcement technology company known for making tasers. With the new service,

[00:11:45] only law enforcement agencies that use the company's software can submit requests,

[00:11:49] but the end result is the same. Law enforcement gets video from users if they choose to share it.

[00:11:55] Ring spokesperson Yassi Yarger says the Axon partnership is unaffected by the end of the

[00:12:01] Flock integration. Additionally, she says no other integrations are currently being explored.

[00:12:05] End quote.

[00:12:06] time for the weekend long read suggestions first up from the journal the ai race has broken a

[00:12:18] long silicon valley no quote notion is among a host of startups and tech companies including Stripe OpenAI Anthropic Databricks and SpaceX that are giving employees eager to access wealth tied

[00:12:31] up in private shares a way to cash out some of it. It marks a shift in Silicon Valley culture

[00:12:36] as companies stay private longer. Selling startup shares early was long viewed as a taboo and a sign

[00:12:41] of lacking long-term commitment. Companies have been able to successfully remove the stigma of

[00:12:47] doing secondary offers for employees, said Charlie Franklin, chief executive of Compa,

[00:12:52] which tracks compensation data. The dam is broken. The number of tender offers completed on Carta,

[00:12:58] which offers financial services to startups, climbed 60% in 2025 from the prior year.

[00:13:03] Even young startups are pursuing tender offers as demand surges for private artificial intelligence

[00:13:08] stocks, and founders recognize that such deals are crucial to retaining talent. If on paper you're

[00:13:13] worth 10 million bucks, you're going to want liquidity for that because you're living in a

[00:13:17] studio apartment in San Francisco when you could instead go and buy a 6,000 square foot house

[00:13:21] in Atherton, an exclusive suburb, said Noel Moldvay, co-founder and chief executive of private

[00:13:27] stock marketplace Augment. Stripe, founded in 2010, is in the early stages of a tender offer

[00:13:33] that will value the payments company at over $140 billion, up from $91.5 billion a year earlier,

[00:13:39] according to people familiar with the matter. Anthropic has also allowed employees to sell

[00:13:43] shares in the past and is planning another tender offer at a $350 billion valuation,

[00:13:47] according to people familiar with the matter, end quote. And from the New York Times,

[00:13:52] a look at how AI is disrupting the romance novel industry. Quote, Ms. Rompodi began writing romance

[00:14:01] novels with the help of artificial intelligence in 2024 using the program PseudoWrite. As a plus

[00:14:06] size woman, Ms. Rompodi wanted to see heavier set heroines she could relate to in romance fiction.

[00:14:12] AI supercharged her writing process, enabling her to produce 10 novels in a little over a year,

[00:14:18] including The Billionaire's Curvy Match and Curves to Own, Pregnant by the Billionaire.

[00:14:24] But while the program made writing faster and easier, it was terrible at describing a plus-sized

[00:14:28] heroine, perhaps because there are so few in mainstream fiction, which many AI programs were

[00:14:32] trained on. Whenever Ms. Rompodi's curvy protagonist, an event planner named Sienna,

[00:14:37] was in a scene. The AI constantly referenced her weight, for example, noting that a chair groaned

[00:14:43] when she sat down. When you make a point that someone is plus-sized, it will exaggerate.

[00:14:48] Somebody is suddenly humongous, she said. It's impossible to gauge how many romance novels are

[00:14:53] produced with artificial intelligence. Many authors don't reveal they use chatbots for fear

[00:14:58] of alienating readers. A survey of more than 1,200 authors across genres showed that about

[00:15:03] a third were using generative AI for plotting, outlining, or writing, and the majority said they

[00:15:08] did not disclose their AI use to readers, according to BookBub, a book discovery site that released

[00:15:14] the poll last May. Some authors who publicly oppose the technology are secretly signing up

[00:15:19] for Ms. Hart's classes, she said. The rapid incursion of AI-generated stories is rattling

[00:15:25] some in the romance business. Publishers and authors worry that books by real authors are

[00:15:29] getting lost in a sea of digital slop as AI-enabled novels flood the market, end quote.

[00:15:35] But here's my bigger question, the question I've asked before. When Sam Altman turns on the ability

[00:15:42] to have ChatGPT spin up your romance story for you, on demand, exactly how you want, do you need

[00:15:50] to turn to someone else to write the story you want to read right now? That's basically the

[00:15:56] question for all of art. Full stop, right? And again, I think this is coming sooner than everybody

[00:16:03] thinks. If I get time to edit it tonight or tomorrow, you will have a bonus episode this

[00:16:16] weekend wherein our good friend Chris Mims of the Wall Street Journal and I talk in a pretty

[00:16:21] philosophical way about using AI in our daily lives. An interesting conversation given the

[00:16:26] headlines from this week. Like I said, I hope to get that to you tomorrow. We shall see.

[00:16:31] Talk to you on Monday.