Category: Top AI Startups

Anthropic AI – #9 in Most Interesting AI Companies – #9 Anthropic AI

August 30th, 2024
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
Mistral AI – #1 in Most Interesting AI Companies – #1 Mistral AI

August 26th, 2024
Kicking off our series on Most Interesting AI Companies with Mistral AI because they are a Unicorn among Unicorns

Why Mistral is a Unicorn among Unicorns
1. Insane Speed of Fundraising
2. Based in France, in a world where AI startups seem to only choose US or China
3. Going up against the Leaders. An absurdly hard task given the first mover advantage and insane amount of money the leaders already have
4. Being able to compete with the Big Boys so far (here is to hoping they can keep making things interesting). It is remarkable that a company founded on April 28th, 2023 is a frontrunner in AI services
5. Actually Open Source, not just in name
6. Very Interesting Products. Relentlessly adding new and improved products
7. No Drama. No fighting with World governments over retina scanning. No boardroom drama. It’s almost as if they are happy being an AI company
8. Undercutting the leader on pricing, albeit slightly. This is a very smart tactic as they are still priced well, just a bit lower than the current leader
All the drama and stupidity we see with the Big Technology companies makes it very refreshing when an AI startup focuses on just the technology.

Mistral’s Unique Challenges

Normal Startup Challenges
1. How to raise enough money to compete against the market leader
2. How to get to profitability and maintain profitability
3. How to do all of this in Europe which is nowhere as startup friendly as the United States
4. How to hire and retain the best, when some competitors are better funded
5. How to hire and retain the best, when some very large technology companies are offering absurdly high salaries to AI engineers
6. How to convince the best AI engineers to move to France (not counting some very good ones already in the Paris area and in France)
7. What products to build
8. How to get breakthroughs in AI research and development while still keeping the Commercial Product Pipeline going
9. How best to utilize funds that are raised
10. What products to build and how to evolve the product offering
Special AI Geopolitical Challenges
1. How to handle US interference as Mistral AI is in France and it becoming the world leader in any facet of AI would be considered a big threat to US hegemony
2. How to handle Chinese interference as, again, Mistral AI becoming the world leader in any facet of AI would be a big threat to China’s plans of world dominance
3. How to handle data related challenges. All the large countries now want to keep their citizens’ data within the country itself
4. How to handle espionage and spying as many of the employees are from the large US technology companies and might be, or could be, compromised by spy agencies
5. How to handle cyber attacks by Governments and Government funded hacker groups
6. How to keep the tecnology and trade secrets safe when many countries would consider Mistal AI technology a ‘must have for national interests’
It is interesting to note that Mistral AI, by locating in France, have opened up a Pandora’s Box of troubles for themselves

This might work, as the added pressure and threats might help them evolve faster and better and create better AI technology.

It might also backfire, as we are seeing what has been done/is being done to TikTok and Telegram for the unforgivable sin of successfully competing with US based companies (Which have always had the backing of the US Government – It is only now that is has become so painfully obvious that the game is rigged in favor of US technology companies).

The Not-So-Obvious Challenges
1. [Bringing this up again, because this is absolutely critical] From what has been happening to TikTok and Telegram’s Pavel Durov it should be obvious that any non US company will be heavily attacked by all means possible. Given the strategic importance of AI to America, it is a sort of madness that Mistral located themselves in France, instead of Silicon Valley or New York
2. Amount of Risk Mistral is taking on, is not something Europe or France are used to. Will there be investors with a higher risk threshold as Mistral AI need more money down the line?
3. Nearly all the Best AI Talent is still in America and China. It is one thing to relocate from Seattle to San Francisco. It is a much bigger move to relocate from America/China to France
  - Side Note: France isn’t really endearing itself to technology workers by arresting Telegram’s CEO on what seem to be rather flimsy charges
4. The sheer amount of money the large tech companies have. At some level, none of the AI startups can compete with Apple, Microsoft, Google in the long term. Until and unless the AI startups make some big AI technology breakthrough or hit massive profitability they are fighting against massive odds
5. The sheer amount of Governmental support the large tech companies have. We see a pattern of Governments letting large tech companies do relentless anti trust and illegal tactics. Then pay a fine and continue to do illegal things to smaller technology companies. It is not an exaggeration to say that Mistral will face many illegal attacks from the large tech companies (it might already be)
Mistral’s Insane Funding History

Mistral has raised money at an almost unprecedented rate, since being founded in April 2023
1. Seed Funding -> June 2023 -> 105 million Euros ($117 Million Dollars) raised at a valuation of roughly 240 million Euros ($267 Million Dollars)
  - Investors -> Lightspeed Venture Partners, Redpoint, Index Ventures, Eric Schmidt (former Google CEO), Xavier Niel, JCDecaux Holding, Rodolphe Saade and Motier Ventures and Bpifrance from France, La Famiglia and Headline from Germany, Exor Ventures from Italy, Sofina from Belgium, First Minute Capital and LocalGlobe from UK
2. Series A -> December 10th, 2024 -> 385 Million Euros ($428 Million Dollars) at a valuation of $2 Billion
  - Investors -> Andreessen Horowitz, BNP Paribas, Salesforce
  - Additional Investors -> Still researching
3. Series B -> June 11th, 2024 -> 600 Million Euros ($640 Million Dollars) in a mix of Equity & Debt (468 Million Euros Equity + 132 Million Euros Debt) at a valuation of $6 Billion
  - Investors -> General Catalyst led the round. The list of investors includes previous investors and some new ones such as DST (Yuri Milner), Nvidia, Samsung Venture Investment Corporation, Belfius, Bertelsmann Investment, Cisco, Eurazeo, Hanwha Asset Management’s Venture Fund, IBM, Korelya Capital, Latitude, Milennium New Horizons, Sanabil Investments, ServiceNow, and SV Angel
Please Note: There were lots of rumors of SoftBank wanting to invest in Mistral AI. So far, that has not happened. We do expect SoftBank to jump in, sooner rather than later.

Super Rough Cap Table + Funding Details

How much equity did Mistral give up on its manic fund raising spree?

Seed -> 43.46% of the company for $113 Million USD at a valuation of $260 Million
Series A -> 21.4% of the company for $428 Million at a valuation of perhaps $2 billion
Series B -> 10.67% of the company for $600 Million USD at a valuation of $6 Billion

Equity Given Up -> 75.55% for $1.141 Billion. So the super-fast fund raising has come at the cost of giving up a rather large amount of equity for the first $1.141 Billion in funding
Left -> 24.45% with Co-Founders and Employees

Please Note: These are estimates based on reports. These might be inaccurate.

The CEO, Arthur Mensch, has said that the founders still have complete control. The Founders perhaps have special class of shares with higher voting rights and likely have over 50% voting rights. This is how the Meta/Facebook founder and the Google co-founders maintained their control of their respective companies.

What makes Mistral very Interesting to DealAI

Firstly, choosing France and foregoing the protection and hidden benefits of choosing America or China

Secondly, the complete and utter lack of Drama. No trying to scan irises. No Board trying to fire CEO

Thirdly, DealAI wanting the underdog to win. All the odds seem stacked against Mistral AI and that makes them a great story

Fourthly, wanting a massive technology company in Europe. There have been close to zero massive technology companies in the last few decades that have not been American or Chinese or Indian companies. Europe has been a wasteland (sorry, Adyen)

Fifthly, they are actually Open Source and not just ‘Open’ in name

Can Mistral take on American and Chinese AI Companies & Startups?

Yes and No

Yes, Mistral AI have shown the ability to take on the US and Chinese AI Companies

No, as the US Government will ensure that an AI company in France does not become the #1 AI company. It might even attack Mistral heavily to prevent it from becoming a Top 5 player in AI

China will perhaps take the approach of hacking and stealing Mistral AI’s technology, or building upon it. In the long run, China too will go out of its way to prevent Mistral from becoming a Top 5 Player in AI

Might Mistral be the Big Winner in AI?

Yes, as crazy as it seems, we do think Mistral has a non trivial chance of becoming the Big Winner in AI

Might Mistral be one of the Top 5/Top 10 Winners in AI?

Yes, this is actually very likely.

At that stage, we will begin to see lots of attacks by both America and China on Mistral.

If Europe realizes the strategic importance of AI and puts its weight behind Mistral, then it would be a very interesting Europa Vs America and Europe vs China battle, with Mistral AI right in the middle of it.

Who are the Co Founders of Mistral?

Mistral has 3 co-founders
1. Arthur Mensch – generally the ‘face’ of the company
2. Guillaume Lample
3. Timothee Lacroix
They have roughly 10 years experience each in AI and Machine Learning. Mensch worked at Google DeepMind and Lample and Lacroix worked at Meta Platforms. They met while students in college.

Which Investment Funds and Venture Capitalists and Investors have funded Mistral?

Seed Round -> Lightspeed Venture Partners, Redpoint, Index Ventures, Eric Schmidt (former Google CEO), Xavier Niel, JCDecaux Holding, Rodolphe Saade and Motier Ventures and Bpifrancefrom France, La Famiglia and Headline from Germany, Exor Ventures from Italy, Sofina from Belgium, First Minute Capital and LocalGlobe from UK

Series A -> Andreessen Horowitz, BNP Paribas, Salesforce. Please Note: This list is incomplete

Series B -> General Catalyst led the round. The list of investors includes previous investors and some new ones such as DST (Yuri Milner), Nvidia, Samsung Venture Investment Corporation, Belfius, Bertelsmann Investment, Cisco, Eurazeo, Hanwha Asset Management’s Venture Fund, IBM, Korelya Capital, Latitude, Milennium New Horizons, Sanabil Investments, ServiceNow, and SV Angel

What does Mistral’s strategy seem to be?

To Be Updated. We are still figuring out exactly what Mistral AI’s full strategy is. We will update this post when we have figured it out

Where does the name ‘Mistral’ come from?

Mistral is a cold and dry wind in the South of France that blows through the province of Provence.

It generally leads to good weather.

It can blow any time of the year. Sometimes it blows for a few days and sometimes for up to a week. Speeds are pretty high with 30 miles per hour to 60 miles per hour.

Please see the Encyclopedia Britannica link below for some very interesting information. It’s a fascinating wind, this Mistral.

Courtesy: Wikipedia and Encyclopedia Britannica

References & Credits
1. Wikipedia -> Mistral AI – Wikipedia
2. Encyclopaedia Britannica -> Mistral | Mediterranean, Provence & France | Britannica
3. TechCrunch -> France’s Mistral AI blows in with a $113M seed round at a $260M valuation to take on OpenAI | TechCrunch and Others

Category: Top AI Startups

Anthropic AI – #9 in Most Interesting AI Companies – #9 Anthropic AI

Mistral AI – #1 in Most Interesting AI Companies – #1 Mistral AI