Section Insights
Anthropic's Product Diversification Strategy
What is Anthropic doing to diversify its product offerings?
Anthropic is focusing on expanding its product lineup beyond Claude Code, with notable successes in Co-Work and plans for other applications like Cyber and Managed Agents. Co-Work has seen significant adoption, indicating a positive reception from customers.
- Anthropic is diversifying its offerings to reduce reliance on Claude Code.
- Co-Work has gained traction, outperforming Claude Code in early adoption metrics.
- The company is exploring various applications to cater to different market needs.
Shifts in Customer Spending Towards Cyber Applications
How are customers allocating their spending on Anthropic's products?
Customers are increasingly shifting their spending towards Cyber applications, with some reports indicating up to 20% of their budget being redirected. However, many are using their own software rather than Anthropic's tools, which presents a challenge for the company.
- There is a notable shift in customer spending towards Cyber applications.
- Customers prefer using their own software for Cyber use cases over Anthropic's offerings.
- Anthropic's ability to attract customers to its own applications will be crucial for future success.
Growth of Cloud Managed Agents
What role do Cloud Managed Agents play in Anthropic's strategy?
Cloud Managed Agents are seen as a significant growth area for Anthropic, providing personalized AI solutions for customers. This category is expected to expand alongside Co-Work, indicating a strategic focus on personalized AI services.
- Cloud Managed Agents represent a key growth opportunity for Anthropic.
- These agents are tailored to provide personalized solutions, differentiating them from more generalized offerings.
- The success of these products will be vital for Anthropic's expansion beyond coding applications.
Anthropic's Discount Strategy and Customer Compliance
How is Anthropic incentivizing customers to use its products?
Anthropic is offering significant discounts and usage credits to encourage customers to adopt its products, particularly for enterprise plans. However, this comes with expectations for customers to allocate a certain percentage of their AI budget to Anthropic, raising concerns about transparency and compliance.
- Anthropic is using discounts to drive enterprise adoption of its products.
- Customers may be required to commit a portion of their budget to qualify for discounts.
- This practice raises questions about transparency and could face resistance from clients.
Future Pricing Concerns and Market Dynamics
What are the implications of Anthropic's current pricing strategy?
The current discounts are seen as a way to transition customers to usage-based enterprise contracts. While this strategy may benefit customers in the short term, there are concerns that prices may rise in the future as the market stabilizes.
- Current discounts are aimed at facilitating a shift to usage-based pricing.
- There is speculation about potential price increases in the future.
- The market dynamics suggest a transitional phase that could lead to more stable pricing models.
Transcript
0:00 The information published exclusive reporting on the discounts Anthropic is making to try to gain traction in many of its products outside of Claude Code. My colleagues Kevin McLolin and Laura Bratton wrote a story about that this week. I want to bring on Laura to share more about what they learned. Laura, welcome back to the show. It's great to have you here. >> Good to be here. >> Okay, so we know that Claude Code has been a huge success for Anthropic. what you and Kevin really got into in your story were the other products that it's going to have to make as stunning a success to really diversify its business a bit. And you know, I I want to go through each of these bets they're making. I mean, Co-Work is I think maybe the other well-known product that people know very well. There's been a lot of competition since Co-Work first came out from other companies. So, you've talked to customers. How do they find co-work compared to the other options?
0:59 >> Yeah, co-work appears to be really popular. Anthropic itself said in its S1 filing ahead of its IPO that you know co-work has seen a higher level of adoption in its first 9 months since being released in research preview than Claude Code saw at a similar point following its debut. So I think that that's a clear sign of codework success. And then you know among the consultants that I talked to they said that their customer spending was roughly split probably about 60% to engineering use cases you know cloud code and then 40% to co-work for knowledge workers to use cloud co-work essentially as a personal AI agent to help them build slide decks conduct research build out dashboards that can help them with things like accounting recruit new folks. though it seems to be catching on.
1:52 >> Okay. So, cowork is catching on. cyber is another application they're pushing into. I mean, this was an interesting point in the story. You said that customers are shifting up to 20% of their anthropic spending in some cases to cyber. That sounds pretty good. Are there caveats here? >> Yeah, definitely. I mean, the consultant I spoke to who told me that said that, you know, they're using the underlying models with their own harnesses for cyber use cases and that that's what he sees as the next biggest use case beyond coding. and I think especially what we saw with the capabilities of Mythos and Babel, that definitely seems like something I think could could be true.
2:35 one thing that I'll note that I found interesting is that customers are using their own software to use the models for cyber capabilities rather than using Anthropic's own tool called cloud security. and I think it just speaks to the tension in which customers are developing their own software. They can call anthropics models for different use cases rather than going to the applications that Anthropic is developing. And I think that whether Anthropic can get customers to use those applications directly as it has with co-work will be a key tell of whether it can succeed beyond coding use cases.
3:15 >> And the the last product category I want to ask you about are clawed managed agents. What are these managed agents? How big of a business is that for for Anthropic right now? >> Totally. So my colleague Kevin got a lot of the reporting on cloud managed agents. He really led the charge there. but what I can say is that cloud managed agents is essentially a service for building and running AI agents. and that seems to be another massive use case for customers that are trying to expand beyond cloud code. So really, you know, I think that you can think of cloud managed agents as more personalized than cloud co-work which is more of a a generalized agent. and you know, Anthropic's biggest use case is going to be agents and and providing agents, personalized agents. whether that's to knowledge workers through cloud co-work or specific customized agents through cloud manage agents to its customers. And that's really what we see as its next biggest use case beyond coding.
4:17 >> Okay. So, these these are these are three categories of products that that you know they're they're growing quickly. They're a big focus for the company. I I want to talk about now how Anthropic is incentivizing people to use all of these products because >> you outlined a pretty ambitious strategy from from the company. >> How does it plan to meet it? What is it giving customers here? Yeah. So, we've seen Anthropic offering higher discounts than previously.
4:46 and in some cases offering usage credits to try new products, usage credits when customers shift from individual or team accounts to enterprise plans with annual spend commitments. and the discounts can range up to 30% or higher. in some cases, we've heard of a Fortune50 customer that was offered a much higher discount than 30%. and to get the 30% or higher discounts. We've heard of some customers sort of being told that they can only get that discount if they promise that a certain percentage of their AI budget will go toward Anthropics products rather than rivals.
5:26 which you know is called a share of wallet provision and is pretty interesting. It's something we've also heard of OpenAI doing. So sorry that the companies do they have to literally open their books and and show their their over but this their bills to at this seems kind of crazy. Yeah, it it it is you know in in the opinion of many customers I talked to something crazy and that they hadn't seen but you know what Kevin found in and his you know research into this was that Anthropic in some cases has asked customers to show invoices or documents to prove that they're living up to the promise of you know diverting a certain percentage of their AI budget to anthropics products. so yeah, >> this is is and is this is this is this at all standard in in software or is this pretty novel here?
6:25 >> you know, I hear mixed opinions on that. Some people have said they've heard of hyperscalers doing this before. Others have said there's no way that companies and regulated industries are going to open up their books and let an AI provider essentially audit them to ensure that they're living up to their share bullet provisions because those are essentially promises. >> right. And then they and then maybe they'd say, "Hey, just upload into cloud. And by the way, the data, you know, it's zero data retention anyway, so don't worry about it." And then they say, "Well, >> we're also concerned about that right now." I mean, this is this is this is like a >> an uncomfortable level of transparency.
7:01 I don't know. I mean, but maybe it works. >> It seems like it's it's unlikely this will become a a widespread practice because I don't think customers would put up with that. and I, you know, I've heard of many cases in which Anthropic will offer the higher discounts 30% when customers up their spending levels and they don't necessarily have to guarantee these promises. So I think that especially as customers begin to use their own coding harnesses or essentially software that lets them use different models like coding agents. anthropics is going to face more pressure and you know not necessarily be able to get away with behavior like this.
7:42 >> Right. Okay. Okay. So then last question for you, Laura. With these discounts, I mean, how long do you think these discounts prevail for? I mean, are people concerned that this is all going to really have a reckoning in a couple years and the company's going to raise prices back up? In my opinion, the discounts are mainly to shepherd customers into usage based enterprise contracts where, you know, they don't have usage included in a $200 or whatever per month account and and we've seen a widespread shift to customers in the last 6 months. Like yes, of course, Anthropic made the shift towards usage based pricing way earlier this year, but a lot of companies are just now shifting to those enterprise plans. And you know, it seems Anthropic is from my reporting mostly providing these discounts to try and help with that transition and to get customers to sign up for for that annual commit. And I think as customers begin to be able to predict how much they're going to spend on AI annually, sure, maybe they'll be able to get a discount. And I think we should know that hyperscalers have long offered discounts for enterprise contracts, which is why the so cost of software is so variable for different companies.
9:01 and you know, taken together, actually, Anthropic and OpenAI's discounting efforts are pretty modest compared to that of the hyperscalers. but I think that what what I'm trying to get out here is that, you know, this is a transitional period and eventually we're going to get to a point where the cost of tokens comes down a lot. Anthropic has already said that, you know, Opus 55, its latest model is so much more costefficient than a previous version of Opus and I think that's going to continue to happen. which is why the frontier model providers need to shift into the application layer so that they can, you know, grow margins and sustain revenue growth, >> right? Okay. Well, Laura, I want to thank you for coming on. That is Laura Bratton, author of our applied AI newsletter here at the information
Summary
- Claude Code has been a major success for Anthropic, but the company is pushing other products like Co-Work, which has seen rapid adoption.
- Co-Work is popular among customers, with a significant portion of spending directed towards it for knowledge work applications.
- Anthropic is seeing a shift of customer spending towards its cyber applications, although many are using their own software rather than Anthropic's tools.
- Cloud managed agents are another key focus, providing personalized AI agents for various customer needs.
- Anthropic is offering discounts of up to 30% and usage credits to incentivize customers to adopt its products, particularly for enterprise plans.
- Some customers are required to demonstrate a commitment to using Anthropic products, raising concerns about transparency and privacy.
- The current discounts are seen as a transitional strategy to encourage customers to move to usage-based enterprise contracts.
- As the AI market evolves, Anthropic may face pressure to maintain competitive pricing while ensuring sustainable revenue growth.
Questions Answered
What is Anthropic doing to diversify its product offerings?
Anthropic is focusing on expanding its product lineup beyond Claude Code, with notable successes in Co-Work and plans for other applications like Cyber and Managed Agents. Co-Work has seen significant adoption, indicating a positive reception from customers.
How are customers allocating their spending on Anthropic's products?
Customers are increasingly shifting their spending towards Cyber applications, with some reports indicating up to 20% of their budget being redirected. However, many are using their own software rather than Anthropic's tools, which presents a challenge for the company.
What role do Cloud Managed Agents play in Anthropic's strategy?
Cloud Managed Agents are seen as a significant growth area for Anthropic, providing personalized AI solutions for customers. This category is expected to expand alongside Co-Work, indicating a strategic focus on personalized AI services.
How is Anthropic incentivizing customers to use its products?
Anthropic is offering significant discounts and usage credits to encourage customers to adopt its products, particularly for enterprise plans. However, this comes with expectations for customers to allocate a certain percentage of their AI budget to Anthropic, raising concerns about transparency and compliance.
What are the implications of Anthropic's current pricing strategy?
The current discounts are seen as a way to transition customers to usage-based enterprise contracts. While this strategy may benefit customers in the short term, there are concerns that prices may rise in the future as the market stabilizes.