Meta's AI Strategy, Databricks new round & What to consume?

1/ What is Meta’s AI strategy?

In 2016 Zuck spent a year on building a home AI, his personal exercise to learn about AI by building a Jarvis like assistant.

Like all the big tech companies Meta has been working on AI projects and using AI in existing products. But for some reason Zuck thought Metaverse is a bigger opportunity than AI for the company. Now that the Metaverse hangover is gone and took a back seat, AI is now at the forefront of what Meta is doing.

But what is Meta’s AI strategy?

It looks like Meta is taking an open-source approach to get developer adoption for its large language models while doing clever partnerships for distribution with the main distribution player in AI, Microsoft.

One of the underlying strategies that’s playing out whether its AI or just new tech in general is the 3 cloud players (Azure, AWS, GCP) are platform to distribute software.

For Meta to make a real business out of AI it should have a path to reach to every Enterprise customer in the world. Which means partnering with cloud players and also making sure developers use them vs using Open AI or Google or any other models. These are two sides of the strategy.

Meta’s AI strategy boils down to two things:

• make everything open source so developers adopt Meta’s LLMs to build next gen AI products

• leverage Microsoft to distribute models to every Enterprise customers (expect Meta to distribute AI models on AWS & Google too in coming years)

2/ Even good companies can be bad investments (Databricks)

Databricks raised another 500M at new valuation of $43B. The round is mostly done by strategic pre-ipo investors & not Softbank or Tiger Global like venture growth firms. Why is Databricks raising more money when they recently raised $1.6B in Aug 2021 and are making $1.5B in ARR? They should have plenty of cash to get them ready for an IPO. I think Databricks is wasting time staying public more time than it should. It reminds me of Stripe, one of the biggest VC success in the last decade that has not yet gone public. In the last 3 years Stripe went from being the best payments company to a competitor to Adyen. There is now even a question whether Stripe & Adyen are great businesses compared to Visa & Mastercard. The narrative is as important as the numbers and Stripe lost the narrative game (not forever but for now it did).

It looks like the investors who led this round’s best case scenario is Databricks goes public at the same earnings multiple as Snowflake in next couple of years. Even n that scenario the IPO valuation would be at 52B-55B range. That’s only an upside of 23% in couple of years. So in this case the investors have to be long term orientated and hope that Databricks would be a generation defining company and hold it for next 5-8 years to exit at a much larger valuation.

3/ What to consume?

You are what you consume, be it food or content. I proud myself finding niche and unique content in my areas of interest.

If you are interested in business, here is a podcast you must check out. It’s called After Hours, hosted by 3 Harvard professors talking weekly about interesting things happening in the world of business. They are currently off for the summer but will be back soon, so keep an eye out.

Till next week, Stay above average!

Nataraj