First, the revenue isn't bad for a 10 year old business.

Eg. Google took 9 years and Facebook took 11 years to get to a $4B quarter, but of course apples and oranges on business models. Advertising has a massive margin, so Google and Facebook were very profitable at those points. AWS took about 11 years to get to a $4B Q. (You get to do the adjustment for inflation ;-)

The problem is that unless you are monitizing a private cloud (Google-classic, Facebook, Apple, etc), the Cloud revenue sweet spot is enterprise sales, and enterprise sales are hard, with long cycles and expensive sales support requirements. When you are playing catch-up there are many more ways to lose a deal than close one.

Year-over-year, Google Cloud made about $1.25B more, and lost $800M less. It sure looks like they will be in the black by the end of this year. It would be a long business cycle for a VC, but looks like a well managed business expansion.

When will the entire venture go ROI positive? Dunno, but in some ways that is a sunk cash fallacy. It looks like a pretty decent business now.

(Ex-Googler still holding GOOG in case either of those color your view of the above.)

AWS is the only cloud hosting company making any real money today. Azure technically does, but only through very creative accounting (revenues are grouped with other segments like Office 365). Anyone who is not AWS has to spend an inordinate amount of money in credits and other customer acquisition costs simply to stay competitive in the space.

Google basically hands out 100k in credits to any startup. It’s spending insane amounts of money to gain market share. But since all three clouds are doing it, the net effect is really just that software companies are getting subsidized.

AWS does this too, also to the tune of 100k iirc. You just need to show that you are funded by a known VC, and then each subsequent funding round gets you additional credits.

Correct. AWS gives you 100k credits for 1 year (or some other deal like 25k/year for 2 years). They also have tons of other incentives

* Free support and consulting (probably useful for some less technical companies)

* Free or discounted resources for POCs (ie: "This customer wants to do a 3 month POC, we're going to spend literally 0 dollars on that infrastructure")

* Other misc stuff depending on the org

worth noting that google doesn't give out 100k for free, or to "any startup". They give it in exchange for equity, and only through their VC partners and accelerators.

Nonsense - they give it away, just like AWS/IBM/Oracle do, no strings attached.

Source: received it. Not in a VC/incubator.

Citation needed.

Last I saw they definitely do not take equity.

You do however have to be qualified by applying through a partner as you say.

Google Cloud as it exists today was formed under Diane Greene (VMware) only a few years ago.

The Google Cloud Platform product (or, "Let's sell some of that extra YouTube server space") is a decade old.

Not sure why you're getting downvoted - it's a reasonable question to ask for a (supposedly) serious competitor to AWS & Azure.

E.g. if I'm leading engineering for a startup and thinking about which of the three to invest in, the long-term profitability and stability of my choice here is important. You can't just rely on the "Google" name - if anything, that signals that they're willing to just shut it down or raise prices if they can't reach profitability targets soon.

Azure raises prices all the time esp their licensing fees. I'm sure so does azure.... So even the bigger providers increase prices to meet targets . Are you expecting static prices

Revenue is up 46% to $4B for Q1 2021. But the costs to close the biggest deals must be gargantuan. Anthos dev & sales alone probably requires a substantial bet ;)

They want market share more than anything. The intention is to trade short-term profits for long-term customers. This is a reasonable business strategy, especially for a business with a relatively undiversified source of profits.

Of course the short-term profits being lost don't seem so short-term at this point, so Google is surely debating internally on how long they want to continue this strategy if it doesn't begin to pay off.

> so Google is surely debating internally on how long they want to continue this strategy if it doesn't begin to pay off

I am certain that nobody is considering shuttering a business unit that has an annualized revenue of $16 billion and is growing 46% per year. Anyone who is “surely debating” this would be fired from management at any sensible company.

It is worth noting that Google has shuttered dozens of products will multiple millions of users that nearly any other sensible company would not shutter, and which, outside of Google, would be considered wild successes. (So few are spun off as separate companies because Google's internal architecture is basically incompatible with anyone else operating their products.)

Google isn't interested in being a minority player in any market, if they can't capture it, they will eventually leave it.

Users != Customers as you well know. If a startup outside of Google had millions of users, they wouldn't be considered a wild success unless they could monetize those users, otherwise it's just an enormous cost that eventually leads to failures once they run out of VC runway.

And let's be real, a huge swath of VC funded startups recklessly engage in "growth hacking" to acquire users, without a legitimate business plan, usually, e.g.

Step 1: get users Step 2: <magic happens> Step 3: Profit!

And then it turns out their ideas for monetization don't work, and they get an exit via acquisition by one of the big players as the VCs try to get some of their money back and leave someone else holding the bag.

SV is littered with companies that had millions of users but ended up folding and selling, because ultimately, running a business isn't free, and those users have to generate enough revenue to cover costs plus ROI.

Sure, but most of these otherwise successful products never even attempted a business model. I could pick a dozen Google apps today and challenge you to determine how Google ever expected to benefit from them. Ironically, most random Google apps don't even display ads. I assume some might exist solely to collect ad data, but I am not sure if Google's ad team is investing steadily in how to mine worthwhile data out of each and every scattered new Google app.

Which is to say, there's probably a decent number of apps Google created and shut down, which probably would've been pretty successful companies on their own with a modest monetization model. But Google never attempted it, because the user count didn't hit nine figures.

Have they shuttered any with >$10b of revenue and >40% growth?

The toothpaste is out of the tube on this one. Outside of the current path, I think the only option is spinning off GCP into a separate company. I don't think this is a bad idea either.

I'm with you on the AI. If AGI is possible in the next decade, I'd argue Google has at least a 30% chance of being the company to pull it off.

> If AGI is possible in the next decade

Fortunately, AGI is almost certainly not possible in the next decade.

Because it is designed to lose money. The ROI window for digital transformation for cloud computing is pretty big. If you want to be top 3 in the industry (AWS, Azure, GCP) you've gotta spend, spend, spend. There are infinite things and features you can supply to customers. Have you seen those cloud consoles?

>Because it is designed to lose money.

Not sure I understand this response, nor do I agree. AWS drove $13.5B in _profit_ for Amazon in 2020, and has been profitable since at least 2014, _8 years_ after its founding.[1]

Why isn't reasonable to ask "where and when is the profit coming?"

[1] https://www.geekwire.com/2021/amazon-web-services-posts-reco...

AWS had a helluva head start.

Not only did GCP (and Azure, etc.) have to catch up with where AWS was 8 years ago, they've got to invest even more to catch up with where AWS is now.

It's obvious that GCP is a very long-term investment, but also one that it not just purely profit-driven for GCP in isolation, but also hugely strategic for Google as a whole.

> Not sure I understand this response, nor do I agree. AWS drove $13.5B in _profit_ for Amazon in 2020, and has been profitable since at least 2014, _8 years_ after its founding.[1]

Here's the question to ask: if Google stopped investing aggressively in marketing and building new offerings to expand the Cloud userbase could it be profitable today?

Or in other words, is it better to have 30% YoY growth and on paper profitability, or 45% YoY growth and an on paper loss, with the ability to flip the switch and move to the first option whenever you want?

If they stopped spending on product and infrastructure build out, they could maybe be profitable at an instantaneous point in time. The issue is that Azure and AWS will continue developing and that will drive the market. Google share will erode and they will not be profitable again.

Is that really the case here? Is Google's marketing holding them back from profitability? Is GCP spending more on marketing than building useful products & expanding their infrastructure?

> 30% YoY growth and on paper profitability, or 45% YoY growth and an on paper loss, with the ability to flip the switch and move to the first option whenever you want?

This sounds suspiciously like a line from the WeWork documentary on Hulu.

Well its a common refrain when discussing startups and high growth business. I'm basically quoting the (ultimately correct) argument about AWS and Amazon as a whole from a decade ago. People were concerned about Amazon not being profitable (or being just barely profitable) for a long time. And then eventually they decided to be profitable.

Exactly. The difference between WeWork and Google Cloud is that WeWork has an unproven viable market, whereas cloud computing has competing, larger players. (If you are implying that GCP could collapse like WeWork)

It is reasonable to ask, I am just giving a possible answer.

Its pretty obvious, serving the customer is not in googles DNA, it never has and never will be. This is core to why all their consumer products have fallen flat on their face. Think google glasses, google delivery, etc

Given that I expect to never see waymo brand cars shuttling people around, most likely they will license out the tech to companies who deal with customers or to the car companies themselves.

> Given that I expect to never see waymo brand cars shuttling people around, most likely they will license out the tech to companies who deal with customers or to the car companies themselves.

I'm a big supporter of what Waymo is doing, but I've always thought this is real end game for them precisely because of the reasons you mentioned. Google is incredibly shy of anything that involves physical support and running a robotaxi service requires a lot of it along with operations/logistics. They are no Amazon.

? It's core customers - businesses who are looking to advertise - are very well served by Google.

Yes! If you factor in businesses running ads, and even Google search usage for work-related tasks, you could argue that Google is skewed B2B. In the long run, I think MS and Google have the same goal: serve businesses a buffet of services they need to operate: productivity, cloud, ads.