Anthropic is one of the biggest AI Startups. It has an excellent set of products, including Claude – A very strong competitor to ChatGPT.

Amazon announced on August 30th, 2024 that ‘Remarkable Alexa’ (wonder who chose that remarkably bad name) will be powered by Anthropic’s Claude.

Anthropic and Claude are considered by many to be the biggest rivals to OpenAI and OpenAI’s ChatGPT.

Why Anthropic AI is a super interesting AI Unicorn

1. Crazy funding that is perhaps even crazier than Mistral AI’s super-fast funding
2. Claude is pretty darn good. Some good arguments to be made that Claude is evolving faster than ChatGPT
3. Interesting Mix of pushing Security and Good Principles and having a solid repuation, while getting sued for using other people’s content without permission
   • A real 50% Hero 50% Anti-Hero vibe
4. The Founders are a brother and sister pair – How often do you see that happen?
5. Very interesting website that does perhaps the best job of selling the company story of any $1 billion or higher valuation company anywhere (yes, better than Apple and other ‘story’ magicians)
6. Claude’s new Artifacts feature is very cool. Here is a good ZDNet article on exactly why -> Why Claude’s Artifacts is the coolest feature I’ve seen in generative AI so far | ZDNET
7. They published Claude’s system prompts. This is an unprecedented move and adds some great detail and transparency -> Anthropic publishes the ‘system prompts’ that make Claude tick | TechCrunch
8. Founded and run by sister and brother pair of Daniela Amodei and Dario Amodei. Quite interesting how everyone forgets Daniela exists (thanks to Mark for pointing this out)

Really, we could go on and on. Anthropic is a very interesting company that stands out – Even in a marketplace (AI startups) overflowing with peculiar and fascinating companies.

Anthropic’s Unique Challenges

1. The amount of money required for Anthropic to keep growing and evolving. Anthropic has already raised $7.8 Billion, and it needs more
2. Possible Loss of Control. The co-founders have already handed over 66.4% or so Equity for the first $3.8 billion out of total $7.8 billion of funding. On top of that, Amazon has $4 billion worth of convertible notes (debt) which might be another 20% to 40% (???) of Equity
3. The lawsuits are a very tough challenge. Just don’t see a future where Generative AI companies don’t have very heavy legal costs and have to pay a significant amount as damages. They are not yet big enough to bribe (sorry, we meant ‘lobby’) Congress
4. Some very impressive competitors. OpenAI is the leader and massively funded and already generating very high revenue. Mistral AI is open source and moving very fast
5. Large Tech Companies such as Google and Facebook/Meta are also mounting their own challenges
6. Is Anthropic too nice and balanced? When some companies in your space are basically run by high-functioning psychopaths and others are run by Open Source altruist saints – Is it a mistake to be a true neutral and well-balanced?

Anthropic is a company that has raised $7.8 Billion in 3 years and yet, somehow, it still faces massive extinction level risks.

Anthropic’s Geopolitical Challenges

1. Anthropic is based in the US. The main challenges it will face are China doing obvious things to block it in China, and Europe doing subtle things to block it in Europe
2. Within the US, it would really depend on which AI startup gives the most access to the US Government. If OpenAI is giving a lot more access, then it is conceivable that Anthropic gets taken out to shift more users and more data to OpenAI
3. Quite frankly, the best way to minimize Geopolitical risks is to found and base your company in the US and then sell out to the US Government. So Anthropic have the first part right and we’re sure they will figure out the second part sooner or later

Very few Geopolitical risks, to be fair.

Anthropic’s Not So Obvious Challenges

1. Raising so much money and from so many high-risk investors. Amazon’s convertible notes (debt). Raising money from the trifecta of Bankman-Fried and Ellison and Singh from FTX. These sort of deals bring their own unique challenges
2. Anthropic’s revenue run rate is $8 million a month in 2023 and is expected to increase to $64 million a month by end of 2024
   • Does that warrant a $15 billion valuation? Perhaps it does in a super hot area like Generative AI/Frontier AI
3. Anthropic’s story is Too Perfect. A somewhat unpolished story (most AI startups), or a very flawed story (OpenAI) are easier to believe. There is literally zero negative press about Anthropic and that is peculiar
4. Effective Altruism Connections. Daniela Amodei is married to the founder of an Effective Altruism Non Profit
   • Anthropic’s Series B had funding from Sam Bankman Fried, Caroline Ellison, and Nishad Singh. Those are the top 3 FTX executives and one or more are reported to be Effective Altruism proponents (unconfirmed)
5. Anthropic is ‘very balanced’ and that is tough when some competitors are very cut-throat and some are ‘open source’. At the end of the day, it comes down to whether the Amodei siblings have the ruthlessness to compete against devils and saints

Anthropic is super interesting because it has chosen the most ‘palatable’ strategy – Neither Evil, nor Noble. Is that a strength or a weakness?

Anthropic’s Crazy Funding History

If you think Mistral are funding savants, wait till you read about Anthropic and their absolutely bonkers funding history (including a Series B led by Sam Bankman-Fried)

1. Across 10 funding rounds (Pitchbook says 10 funding rounds) they have raised billions
   • New York Times says Anthropic have raised $7.3 billion in 2023 and 2024
2. The specifics of their funding rounds are rather difficult to find. We have listed what we could find below, in a later section
3. They had their Series B in April 29th, 2022 led by Sam Bankman-Friend, with Caroline Ellison (FTX) and Nishad Singh (FTC CTO) as other investors
4. The Center for Emerging Risk Research (https://emergingrisk.ch/) invested in Anthropic’s Series A and Series B rounds. It hasn’t invested in any other companies (according to Crunchbase) and information on it is hard to find

Details on Anthropic’s Funding Rounds –

1. Series A on May 28th, 2021 -> $124 Million raised at an unknown valuation. Led by Jaan Tallinn (Co Founder of Skype)
   • Investors -> Jaan Tallinn, James McClave, Duskin Moskovitz (co-founder of Facebook), Center for Emerging Risk Research, Eric Schmidt (former CEO of Google), others.
2. Series B on April 29th, 2022 -> $580 Million raised at an unknown valuation. Led by Sam Bankman-Fried.
   • Investors -> Sam Bankman-Fried (CEO of FTX), Caroline Ellison (FTX), Jim McClave, Nishad Singh (FTX CTO), Jaan Tallinn, the Center for Emerging Risk (CERR – https://emergingrisk.ch/).
3. [Unconfirmed] Round X in early 2023 -> $400 million investment at an unknown valuation. Led by Google
   • Investors -> Google, unknown if others
4. Series C on May 23rd 2023 -> $300 million investment at a $4.1 billion valuation. Led by Spark Capital
   • Investors -> Spark Capital, Google, Salesforce Ventures, Sound Ventures, Zoom Ventures
   • Spark Capital’s Yasmin Razavi joined Anthropic’s Board of Directors
5. Series D in August 2023 -> $100 Million at an Unknown Valuation
   • Investors -> 2 Asian telecoms
6. Series E in End 2023 -> $4 Billion at an Unknown Valuation
   • Investors -> Amazon
   • Investment -> Convertible Notes (Debt that becomes equity when Anthropic hits certain milestones)
7. Series F in End 2023 -> $2 Billion at an Unknown Valuation
   • Investors -> Google
8. Series G in February 2024 -> $750 Million at a $15 Billion (according to the New York Times)
   • Investors -> Menlo Ventures
   • Reference -> Inside the Funding Frenzy at Anthropic, One of A.I.’s Hottest Start-Ups – The New York Times (nytimes.com)

Would never have thought that Mistral AI’s funding rounds would seem sane. The combination of raising $7.854 Billion in 3 years and having $4 billion as convertible notes (debt) is a sort of madness we don’t get to see and appreciate often.

Anthropic Super Rough Cap Table + Super Rough Equity Ownership

Please Note: This is based on estimates and assumptions. Do not take this as Gospel

The only known figure we have is that FTX’s stake in Anthropic is 8%. We know this from FTX bankruptcy court. Everything else is speculation and guesstimates.

Series A -> $124 Million. Assuming $1 Billion Valuation -> 12.4% Equity. Most likely Equity would be higher than 12.5% as this was a Series A

Series B -> $580 Million. Very hard to say percentages as FTX share of Anthropic is 8%. It is not clear what share the other investors have. Let’s say $2.9 Billion Valuation and 20% Equity.

Series X -> Unconfirmed $400 Million from Google. Leaving this out as Unconfirmed.

Series C -> $300 Million at $4.1 Billion Valuation. That’s 7% Equity approximately

Series D -> $100 Million at $5 Billion perhaps (unconfirmed). 2% Equity.

Series E -> $4 Billion Convertible Notes (Debt) at perhaps $10 billion valuation. Can’t say equity percentages until the notes convert. Perhaps 20% to 40% Equity

Series F -> $2 Billion from Google at perhaps $10 Billion Valuation. So 20% Equity.

Series G -> $750 Million at $15 Billion Valuation. So 5% Equity.

Obviously, the numbers don’t really add up. It’s a total of $7.854 billion raised and equity of 66.4% given up. Additionally, Amazon has convertible notes that might be anywhere from 20% to 40% Equity.

In theory, Anthropic gave up 66.4% for $3.854 Billion and another 20% to 40% as convertible notes (debt) to Amazon for 20% to 40%.

One wild card is that Amazon might use that convertible notes (debt) to nuke Anthropic like they nuked Future Retail in India. So that is a big, big risk Anthropic took.

FYI -> If you are a hot AI startup, don’t ever take on Debt.

What makes Anthropic Very Interesting to Deal AI

Firstly, the sheer amount of ammunition Anthropic has.

Secondly, how good Claude is.

Thirdly, the implicit and explicit backing of the US Government and of US Venture Capitalists and US Institutions.

Fourthly, Anthropic has dodged all the bad PR some AI companies have been inundated with.

Fifthly, Anthropic has started a $100 Million fund for AI startups. It provides checks of $100K and also $25K in Anthropic credit. This is quite interesting, for a number of reasons we will discuss in a separate post.

Sixthly, Anthropic seems an insanely well run company. Getting a lot of Nvidia vibes where everyone knew Nvidia were doing beautiful work in Graphic Cards. No one realized they would become #1 in AI chips and AI servers. Anthropic has a lot of the same ‘we get stuff done excellently’ vibes

Seventhly, they already have a very, very strong stable of Clients. Perhaps second only to Open AI.

Eighthly, their PR and Press Outreach is very strong and very professional. Notice how all the articles about them seem to fit into the Anthropic ‘story’ seamlessly. What a remarkable coincidence, eh.

Ninthly, so far they seem to be following the ‘Fast Follower’ strategy excellently. The first mover sometimes wins. At other times, they take all the hits of being the pioneer and make a lot of mistakes. Then the ‘fast follower’ becomes the winner. Anthropic are setting themselves up perfectly to be ‘The Fast Follower that learned from the First Mover’s mistakes and then win the race to be #1’.

Tenthly, Daniela Amodei is a very smart cookie who communicates superbly. Watch her interview and tell us who is the best communicator out of all the AI Startup founders -> Anthropic Co-Founder Daniela Amodei on AI adoption, Claude 3 and impact on payments | Watch (msn.com). Watch this too, if you have the time -> First Block: Interview with Daniela Amodei, Co-Founder & President of Anthropic (youtube.com)

Might Anthropic be The Winner in AI?

Yes, Anthropic has a pretty decent chance of becoming the Big Winner in AI. This is, of course, not guaranteed in any way, shape or form.

We do feel very strongly that OpenAI will get overtaken because of OpenAI’s constant drama and rollercoaster leadership. When that happens, Anthropic has as good a chance as any AI company to become the frontrunner.

Might Anthropic be a Top 5/Top 10 Winner in AI?

Yes, there is a very high chance Anthropic will be a Top 10 Winner in AI. There is also a decently high chance Anthropic will be a Top 3/Top 5 Winner in AI.

Who are the Co Founders of Anthropic?

Brother Sister pair, the Amodeis.

Dario Amodei -> Dario Amodei – Wikipedia

Daniela Amodei -> Daniela Amodei – Wikipedia

Which Venture Capitalists & Investment Funds & Investors have invested in Anthropic?

Series A Investors -> Jaan Tallinn (Co-Founder of Skype), James McClave, Duskin Moskovitz (Co-Founder of Facebook), Center for Emerging Risk Research, Eric Schmidt (former CEO of Google), others.

Series B Investors -> Sam Bankman-Fried, CEO of FTX. Caroline Ellison (FTX), Jim McClave, Nishad Singh (FTX CTO), Jaan Tallinn (Co-Founder Skype and Lead Investor in Series A), the Center for Emerging Risk (CERR – https://emergingrisk.ch/).

Series X Investors -> (Unknown, rumored to be $400 Million from Google) -> Google, unknown if others.

Series C Investors -> Spark Capital, Google, Salesforce Ventures, Sound Ventures, Zoom Ventures

Series D Investors -> 2 Asian Telecoms

Series E Investors -> Amazon

Series F Investors -> Google

Series G Investors -> Menlo Ventures

DealAI’s Thoughts on Anthropic’s Future

Very Bright. Incredibly Bright.

To be quite frank, Anthropic are doing almost everything right –

1. Choice of Location
2. Amount of Funding.
3. Funding when the market is hot
4. Breaking the rules just enough
5. Selling their story very well

After Nvidia, Anthropic might be the best run AI company right now.

DealAI’s Thoughts on Anthropic’s Strategy

To Be Updated. While we have figured out a lot of Anthropic’s strategy there are some aspects we still don’t understand. We will update this Strategy section when we have more clarity on Anthropic’s complete Strategy.

References & Credits

1. Anthropic company website -> Home \ Anthropic
2. Wikipedia -> Anthropic – Wikipedia
3. New York Times -> Inside the Funding Frenzy at Anthropic, One of A.I.’s Hottest Start-Ups – The New York Times (nytimes.com)
4. Techncrunch -> Anthropic publishes the ‘system prompts’ that make Claude tick | TechCrunch
5. ZDNet -> Why Claude’s Artifacts is the coolest feature I’ve seen in generative AI so far | ZDNET
6. The Information (Subscription Wall) -> OpenAI Rival Anthropic Raises Funding at $4.1 Billion Valuation — The Information