Last month, Google announced the hiring of the founders of Character.ai, Noam Shazeer and Daniel de Freitas, both former Googlers. The movement drew attention for a few reasons.
One of them is that, in addition to hiring, the companies signed a million-dollar licensing contract. The other is that this was not the first move with this face in the AI arena. In recent months, big tech companies Microsoft and Amazon have also announced similar deals. The reason? These companies have a lot of market power and resources that place them in an extremely privileged position in the AI race. Acquisition movements by startups that allow them to accelerate even further in this territory place them in the spotlight of regulatory bodies.
Hiring key people for this, therefore, is a safer move. At least it would be, if it were a mere hiring.
In this article we will understand how and why these most recent movements went well beyond traditional acqui-hires.
Detail of recent movements
Let's understand the recent deals in a little more detail.
Google + Character.ai
Google re-hired in August 2024 the two founders of Character.ai along with a "small" key team whose number of members was not disclosed.
The startup had already raised 197 million dollars, reaching unicorn status ($1 billion valuation). Negotiation includes the purchase of investors' stake in the startup for a valuation of US$2.5 billion, but the total transaction values were not disclosed to the media.
Microsoft + Inflection
Microsoft paid Inflection $650 million for a licensing deal to offer its models through Azure. The deal includes the hiring of many of the company's 70 employees and most of the money will be used to buy back stakes from investors (including Greylock Ventures and Dragoneer) at 1.5x the valuation at which they invested. Inflection had already raised $1.3 billion, of which Microsoft was the main investor.
Two reasons are given for the movement. Firstly, it is believed that Microsoft wanted to reduce its dependence on OpenAI, even though it invested more than 13 billion in the company. Secondly, the format of the deal aimed to reduce scrutiny from regulatory bodies.
Amazon + Adept
Amazon also made a similar deal with Adept, which has already raised more than US$415 million with a valuation of approximately US$1 billion.
Amazon has hired about two-thirds of its employees and is also licensing the startup's technology for an estimated value of around US$440 million, just enough to return the funds' investments and leave US$25 million in the company to cover cash burn of the remaining team.
Now let's understand why this is happening.
Acqui-hire is about people. The asset is knowledge
An acqui-hire is, in fact, hiring. The fusion of words originates from the strategic motivation behind the business. But in practical terms, this often does not involve acquiring the legal entity of which this team is a part.
This does not mean that this legal entity is not rewarded for this in some way, such as through compensation. This can be especially important if there are partners who are not being hired in this movement, such as investors.
But the important thing here is that this strategic motivation is justified by the value of the knowledge that is being acquired. After all, knowledge is basically all that is being purchased in this case. If there is software or another type of intellectual property where most of the value is concentrated, an acqui-hire format should hardly make sense.
In the case of the first acquisition of Nubank, Plataformatec, we studied a typical acqui-hire. We are talking about the acquisition of a consulting company, whose main asset is brains.
Can it happen that a company that develops products, technology, has its team incorporated through an acqui-hire and its product (and business) is "discarded" by the buyer? Yes, it can, but this is why I put this type of deal at the beginning of the value ruler. In other words, if this is happening to a company whose main asset is not the team's knowledge, a large part of the potential value of an M&A is being left aside.
But when it comes to artificial intelligence, the situation is more complex than that.
In AI the asset is knowledge + data
The battle of artificial intelligence is very heated. Since OpenAI launched ChatGPT, the entire technology world has moved quickly to not be left behind. And that involves a lot of venture capital investment and acquisitions.
Between the hype and the promise of delivering real value, positioning itself as a powerhouse in AI has been seen as what can make any company the winner in its sector, justifying acquisitions such as Hyperplane by Nubank.
At the heart of the competition are the big techs (Google, Amazon, Microsoft, Apple, Meta) and the challengers (OpenAI, Anthropic, Perplexity, X.AI), startups with a lot of funding to go head-to-head with the giants.
Every month we watch the launch of new models, whose performance is included in comparative graphs, making a herculean effort seem like just another colored dot on a white background. However, all the investment in this competition does not translate into a large and lasting value in intellectual property.
This is because if one of these companies stagnates in its development pace, its models will be left behind and its value will drop precipitously.
Therefore, the greatest value is not in the model itself, but in the ability to continue developing technology capable of competing in the race, head for head. And for that what is needed is: data and brains.
All new generative AI algorithms emerged from the "Transformer" created within Google, but which did not receive as much attention within the company as outside it. That is, until two years ago. The brains that carried this knowledge outside the company were able to evolve the algorithms and the rest is history.
But as new as this transformer that gave rise to generative AI is, it is still part of the machine learning discipline, which means that to train them you need a good quantity and quality of training data.
When we talk about data, you already know who is on top of the story. Oh no? So try talking about anything else to the person next to you. Like "I need to buy new sneakers." Then open your cell phone and see if sneaker ads start appearing for you.
Jokes aside, Big Techs have access to massive amounts of data. History is showing that all they need to stay in the AI competition are the best brains.
That's why these acqui-hire movements became popular for them. Solve the problem, fall under the scrutiny of regulators. Or at least so they thought.
But how does this make sense to the other party?
The alignment of interests
The saying goes: "when one doesn't want to, two don't fight". So for a movement like these between Google and Character.ai, Amazon and Adept, Microsoft and Inflection actually happen, it has to make sense to the other party. This means that it needs to make sense for the founders and other holders of their social capital, such as their investors.
Why it makes sense for the founder
When Noam and Daniel left Google and decided to start their own business, they were probably looking for a few things. On the more idealistic side, the founder seeks to leave his mark on the world. Do something relevant that changes the course of history, even if just a little.
Before, as Google executives, they didn't feel like they were succeeding at it. That's why they needed to go out and make things happen. The same happened with the founding team of companies like OpenAI and the result is that Bigtechs are now willing to invest massively in the topic. Very different from when they left the company, now they can find in the mother ship the intention and power of transformation that they were exercising outside of it.
The other thing that drives entrepreneurs to leave a corporation and found companies is the search for a financial reward disproportionate to an executive salary. Deals like these materialize this financial gain by checking both boxes at once.
But there is also the side of what they can avoid with a movement like this. When competition becomes a big man's game, as we are seeing now, the risk increases considerably. This means that one false step can make the startup fail and all the value built be thrown away along with the effort spent on it. Knowing how to recognize windows of opportunity and jump into them in a scenario like this is an art.
Why it makes sense for the investor
It can be argued that the investor typically has less altruistic ambitions when it comes to changing the world. But when it comes to financial aspects, both opportunities are also true for investors. An acqui-hire in such a dynamic scenario that has made brains the biggest asset of the company they invested in could be the best exit opportunity they will have access to.
And the risk of everything when trillion-dollar companies make AI their number one priority is also very real for the investor. But in practice, how will he get his hands on this money? That's where the "weird" part of these businesses comes in. Licensing contracts are quite possibly just justifications to make an exit possible for investors as well. Or does anyone here see Google reselling an AI-powered virtual girlfriend from Character.ai?
How this strategy can benefit you
Keep reading with a 7-day free trial
Subscribe to Exit Strategy to keep reading this post and get 7 days of free access to the full post archives.