32% YoY increase in revenue, and a 15% pop after hours, 1.3B revenue for the year, against (what is currently) a 35B market cap, for 27 trailing P/S. Okay, free cash flow was $40m, so trailing P/FCF of 220 - seems like the SaaS stocks (like CRWD, SNOW, ZS, etc) are again commanding a premium, except this time everyone’s growing at 30% instead of 60%+ (ZS at least still at 40%), am I missing something? (fwiw I sold my remaining 2% position in NET after hours at $105)
NET has been lagging the rest of the CyberSec stocks that are all hitting daily ATH’s/2yr highs the past few months. The buyside expectations were just too low. Hence the pop.
The report looks good not great.
Long NET(it is a never sell for me)
This is how I see the space (SASE growing the fastest presently)and now Cloudflare GTM is improving (?)
to be considered “single-vendor SASE” with their moves towards SD-WAN inclusion or bypassing the need for it altogether (where Zscaler and Cloudflare are both going). But a big differentiator that Cloudflare has in the single-vendor SASE market is their other related services that are a part of Cloudflare One , which includes email security from Area 1, as well as core networking services like DDoS protection, load balancing, FWaaS, Argo Smart Routing, and more. This “Connectivity Cloud” can bring all these key features to every cloud, branch, and private network that it interconnects. I hope to see them leverage these added capabilities (that the other rising SSE leaders don’t have) to gain market share and rise in the leaderboards.
Mark Anderson is from Palo Alto Networks and was instrumental in them growing like they did with all the contacts etc (?)
In isolation it wasn’t a great report. But one part of today’s move I think is that the market is looking out 6-12 months and seeing the nadir in growth decline due to “AI tailwinds”. In other words, it’s pretty likely Cloudflare’s revenue growth will show some acceleration in FY24
For instance they just reported 32% YoY growth in Q4. They guided for 3% QoQ growth for Q1. If they beat that even modestly and get 5% QoQ the YoY growth will be 32% matching Q4.
Their growth peaked in Q2 2022 at 54%. So almost 2 years later the growth slowdown is leveling off and perhaps will show some acceleration.
Seems like FY23 was the year of “optimization”, which has abated, and FY24 will see a rebound. So again, in isolation the report wasn’t amazing but it shows the growth story for NET is not broken.
In Q3 2023 CFLT reported a modest slowdown and the stock dropped 42% in one day. Yesterday they reported a modest beat and a weak FY guide of 22% and the stock was up 35%. So the market is trying to make sense of the longer term growth trajectory.
In the call Prince noted that it’s the most common cloud platform among “the hottest AI start-ups.” This matches with a link to a podcast that I posted here last year where the guest (a venture capital investor in AI) was asked if there were any public companies he was watching - and he called out Cloudflare as being one that he was seeing all of his AI companies working with.
A little AI hype might be part of the stock price boost.
I’m not sure why, but I think I read that Cloudflare’s network/technology is more well suited for AI than competitors. (Don’t quote me on that)
This is now my largest stock holding. 3.8%.
Numbers in itself are good not great. But the key to reaction I think is -
- Acceleration in RPO, bookings, pipeline
- AI Buzz
- Evidence of traction of its zero-trust solution in enterprises.
Relevant quotes from conference call:
We blew away our previous record for new ACV booked in the quarter. In Q4, new ACV booked grew nearly 40% year-over-year, making it not only our record in absolute ACV, but also the fastest percentage growth we’ve seen since 2021. Our pipeline closed rates sales force productivity, average steel size and linearity all improved markedly quarter-over-quarter. I continue to be encouraged by the early results we’re seeing from our new sales recruiting and training programs.
During the fourth quarter, the pipeline generated by the cohort under the new program was 2.1 times higher than the year ago cohort. And account engagement increased by 3.5 times compared with a year ago. I’m excited by our maturing sales execution and believe it will pay dividends in 2024 as these sales professionals fully rank.
Remaining performance obligations or RPO came in at $1.245 billion, representing an increase of 15% sequentially or 37% year-over-year. Current RPO was 73% of total RPO.
And I think that’s the biggest thing to point out. I think the second thing is that we’ve really seen strength in our Zero Trust and SASE products. Some of the biggest wins that we saw in the quarter were big Zero Trust deals. And I think that those products have caught up significantly. And you’ll see in the Analyst Day, Industry Analyst reports and others that we’re very quickly becoming one of the real leaders in that space. And I think over time, we will surpass many of the first-generation Zero Trust vendors and that really opens up another leg on the stool.
And then lastly, I just continue to be really astonished by the rate at which our developer platform is growing. AI is not yet contributing materially to revenue. But it has contributed materially to developer excitement over the platform. And we’re seeing more and more large deals, more and more interesting applications coming in using the workers’ platform and that has gone well as well. So I think operational excellence in go-to-market, I think, that really strength around our Zero Trust products and then just continued outperformance on our developer platform. And those things all came together in Q4 of 2023.
Seems like from all of the responses here that it’s a lot of hope and hype - I do admire Prince for being such a great communicator and salesman of his stock - there’s no other company that’s been able to command such a high valuation for such growth rates as low as NET. Product announcement after announcement and none of it ever translates into hyper growth in revenue. Will’s chart seems to be pretty accurate - product and tech “innovation” that nobody wants to pay for. To be at the same level as companies like Symantec and Citrix is not a good sign, imo.
The new hire to run sales, Mark Anderson, has not been at Palo Alto since 2018. You know what he actually ran for the past 3 years? Alteryx. An old favorite around these parts that increased revenue by 13%, 60% (nice), and 19% in 2021, 2022, and 2023, respectively, in the 3 years he was there, before the company went private a couple of months ago. Does that seem like someone who will re-accelerate NET to 50%+ to you? But maybe it’s AYX’s product.
Oracleoo’s point on RPO and bookings is interesting, but the numbers are still tepid no? 15% sequentially is a big number, especially annualized, but we all know RPO is bumpy and dependent on when/for how long customers sign deals - YoY is still only 37%, hardly a hyper growth amount. The pipeline speak is again nebulous. Is it 2.1 times higher because last year sucked? (i.e. when Prince publicly blasted his sale org) Or because they have actual promising soon to be customers/revenue? What does 3.5x account engagement even mean? Does it translate to more revenue or are your sales associates just expensing lunches with existing customers? Net retention rate was only 115%, meaning if you want to hit 50% growth rate you’re going to have to expand your customer base by at least 30% if they have the same revenue profile as existing customers, but most likely more since new customers are typically smaller and contribute less revenue.
A lot of talk about “AI tailwinds” - but no substance - what are they actually doing to enable AI? Is it something customers actually want? Seems like the only software companies right now benefitting from AI is either OpenAI, or the omega caps like MSFT and META. Nebulous statements about AI from companies like NET, SNOW, and many other software companies sure seem to move the needle on their stock price - I just don’t know if it will translate to actual revenue. We’ve had a year and a quarter now since the seminal AI moment when ChatGPT went live publicly, and it seems like hardware/semis will continue to reap the rewards for the foreseeable future - it could take many more years than we realize for software to start getting benefits.
I looked back at my portfolio last year and at one point I had a 15% position in NET early in the year, in the 60s. SNOW/NET/CRWD all look overextended to me at this point - the prices would be more reasonable after another year of 30-40% growth. I’ll be sitting this one out for now, good luck to all staying in the company. I really do hope NET continues to go to the moon
This is an excerpt from a MF summary on NET. It is from 1.5 years ago, so I assume their network is even greater now.
“Cloudflare’s platform is truly massive. The global network covers more than 275 cities across over 100 countries, with network capacity of 155 trillion bits per second. Cloudflare boasts 11,000 networks connecting directly to its platform, including internet and cloud service providers and larger enterprises. The company’s network is so comprehensive that 95% of the world’s population can connect to a Cloudflare data center within 50 milliseconds.”
I think NET gets such a high multiple because of the moat of this network. They have greater likelihood of longevity than a pure software company in my opinion. Software can be replicated or obsoleted more easily than an enormous physical network like Cloudflare has built. Pure software can scale faster, and it’s more profitable, but I think larger investors will pay more for a company with a greater moat even if it’s not growing as rapidly.
I do personally think it’s getting a little stretched at it’s current valuation. But at one point in the not-to-distant past I think it had a P/S of 60 or 70 - so “stretched” can always stretch farther.
This is an excerpt from Brad Freeman’s free weekly newsletter. Go to his website if you want to read his whole summary on the ER.
"Cloudflare wanted to have inference tuned Graphics Processing Units (GPUs) in 100 cities by this quarter. It reached 120 cities and reiterated plans to have these GPUs in its entire global network by the end of this year. That is a prerequisite for running GenAI use cases like Workers AI through Net’s platform. 6 years ago, Cloudflare began leaving empty slots in its servers around the world. It did this in anticipation of Worker AI GPU demand. This means it can simply plug Worker AI GPUs into its existing server infrastructure. So? GenAI proliferation will not entail hefty boosts to capital expenditures. Impressive foresight.
“Since the firm’s Workers AI launched in September, usage requests are up 900%. Notably, 33% of all Workers AI requests are from clients new to the Workers Platform. This tells leadership that GenAI will not only create new demand streams, but accelerate existing ones too. While Workers AI brings a bevy of model building and usage options, Vectorize allows these models to be visually enriched with a client’s custom first party data to drive more relevant use cases. Worker AI is driving strong, complementary demand for Vectorize thus far.”
And, as a reminder, NET is still slightly less than half of its ATH. SNOW is around 59% of its ATH, and ZS at 74%. CRWD is at its ATH right now.
Looks like NET beat the guidance by a small margin but it’s really the commentary from the leadership which is driving the price.
- Revenue was guided 352-353, actual 362M
- Operating income guided 28-29, actual 40M
- Net income per share guided 0.12, actual 0.15
Recall from previous quarters that Cloudflare is known for having a lense ahead on the SaaS market cycle, and here is what they are saying,
- We had an exceptionally strong quarter
- Largest new logo signed with over 30M
- Largest renewal with total contract 60M+
- New ACV booked grew 40% yoy making it the fastest percentage growth since 2021
- Gross margin 78.9% ahead of 75-77% target
- From launch in September to month of December, average number of daily workers AI requests increases 9x
- Added record new customers spending 500k+ and 1M+ this quarter
- 500k customers up 56% yoy, 1M+ up 39% yoy
- RPO 1.245B up 15% sequentially, +37% yoy
- Seen strength in Zero Trust and SaaS products
- AI is not yet contributing materially to revenue, but contributed to developer excitement over the platform
- Improvements in sales and to to market, pipeline deal size improving, quarter over quarter RPO growth
- We have all the different components to build a full featured AI application
- Zero Trust products can have traffic flow both ways without additional CapEx
- Network CapEx was 8% of revenue
- Ahead of GPU delivery to the Edge, at 120 out ahead of 100 for the plan
- Guidance of 10-12% of CapEx is setting up GPU capacity for every location
- Cloudflare is the most common platform used across AI startups and companies
- Workers surpassing downloads over AWS serverless solution
- Workers a big part of the deals they are signing
- As AI applications turn into real products will see revenue show up that is meaningful
- Pipeline close rates, sales force productivity, average deal size, and linearity all improved qoq
My take is they are building out their AI platform really well and it’s picking up marketshare as the go to solution for building AI applications. While it’s being run at cost right now, their will be significant operating leverage later on. Right now it’s more important they attract the developers and get the Workers platform ubiquitous before emphasizing profitability.
It sounds like their Zero Trust solution is picking up momentum and a large part of many deals. I’d expect Cloudflare to beat their upcoming numbers based on the optimism from the call.
Enable AI? You can already run inference on their global network with a few lines of code, or a simple API call. Or simply try it out at https://ai.cloudflare.com/.
During birthday week they laid out an ambitious target to scale GPU deployment from practically zero to 100 cities in just three months, and reached it one month ahead of schedule. What’s more, somehow they’re doing this while simultaneously sustaining or improving their margins.
In my view, they’ve been pretty clear about what they’re doing and what to expect. The tricky part is to temper expectations terms of translating it to near term revenue, since it depends not only on demand, but also on pricing and how fast they can move on an ambitious roadmap.
As of now, you can run about 20 select popular models on Workers AI - typically in the 7-13B parameter range, and that’s where the focus has been and probably is. I.e. developers are begging for more, and the initial focus is on adding popular models. Looks like full “bring your own model” is further down the road, and the intermediate step is support for bringing your own fine tune to some of the base models.
Notably, Workers AI is not just about low latency inference. For example, a call to one of the text-to-image models takes about 15 seconds. There’s little reason to run that type of workload as close to the user as possible, and this is reflected in the two-tiered pricing model, where you have the choice of letting inference tasks run wherever there’s spare capacity - at lower cost.
Although I still look to SASE/ZT when it comes to revenue, this is a significant infrastructure investment that goes way beyond the vague “almost every AI company uses us for something” tailwind a few quarters ago.
All this being said, I too feel that valuations have gone up quite a bit, and it would feel irresponsible to anyone who happen to read this board not to mention it.
My quibble is not if their products work, it’s if people are willing to pay for it.
If you’re running an AI startup, you’re likely not using Workers AI for your deployment, you’re most likely calling an OpenAI api, or training your own model from one of the open source ones and deploying that on AWS/Azure/GCP. It’s nice that Cloudflare has a product, but near all of their releases in the past 3 years has reeked of “throw something at the wall and see what sticks”/let’s get coverage to say that we have a product for that, rather than, here’s what paying customers want and will pay for more of. Notice the company never breaks out where their revenues are coming from, or puts a percentage to press release descriptions of “strength” in AI/SASE/ZT, only a list of “wins” which sums up to be single percentage points of their quarterly revenue - I’m willing to bet it’s because the vast majority is still unsexy/boring CDN. Companies like MDB or ESTC started breaking out revenues and growth rates for their managed services because it was actually making a difference - I suspect if SASE/ZT/AI revenue was more than just immaterial and growth rates were tremendous that NET would do the same
Remember $5b by 2025? This was the tagline just 18 months ago. At current rate the company will need to grow 96% compounded over the next 2 years to reach that. That’s how far the growth rates (and base expectations along with it) have fallen in the past 2 years. IMO 50% is the minimum growth rate to deserve their current valuation. Doubtful they’ll even get to 2.5b at their current pace. I just don’t think it prudent to fall in love with product and roadmap announcements - we need to see the numbers follow the story that Prince is attempting to craft.
To be clear, I’m not saying NET is a bad company, if it drops down to the 60s again I would likely buy it - just that at current prices, I don’t think the reward to risk ratio is in our favor.
That was NOT the target. Prince said they can get to $5B revenue in 5 years in Nov’2022 so that would be in ~2028 sometime. He would never guide for that type of growth by 2025.
I am assuming he means $5B quarterly revenue run rate (~$1.25B per Q). They just reported $362.5m or $1.25B rr. So ~36% CAGR until the end of 2027. Is this likely? I don’t know. But if any company can do it I believe a generational biz with a CEO like Prince can.
NET provided color on this during investor day last year. The counterpart to the unsexy/boring part of the business crossed 25% of ACV at the end of 2022. It’s hardly immaterial, and most certainly even less so a year later.
While I can’t dig up the transcript at the moment, the CFO has acknowledged the need for more transparency as they grow larger, while still being mindful of the fact that there are competitors out there, and I honestly don’t see anything unusual about it.
As an example, a while ago I tried breaking out the size and growth rate of the various components that PANW includes in their “Next Generation Security” ARR. Sixteen quarters worth of presentations and transcripts netted just a few data points. E.g. Cortex ARR was ~$400 million in Q2 CY21.
One could do the same exercise with other SaaS/IaaS/PaaS companies - including those discussed here, and the result would be similar.
Cloudflare One was introduced a little more than three years ago and has since then matured into a full-blown single-vendor SASE solution, capable of displacing competitors, with RBI, CASB, DLP, email security, “branch connector” appliances (Magic WAN Connector) and more.
I’ll refrain from trying to list everything that goes into the what customers want bucket, but it’s a lot.
You do have a point about throwing something at the wall, however. Cloudflare runs almost like two separate developer organizations in parallel, with one of them being “Emerging Technology and Incubation” (ETI). That’s the “throw something at the wall and see what sticks” department. So you have something akin to one organization focused on innovation and making lots small bets, and another one focused on meeting customers’ needs and improving and maintaining products. It’s an approach on tackling the innovator’s dilemma.
The other thing to know is that Cloudflare’s “innovation weeks” aren’t just for fun or publicity. It’s a way to motivate and rally teams behind a deadline. E.g. actually shipping products or getting them out of beta. Blog posts are often written by the lead developer(s) or team lead responsible for the product/feature. In a way, you could see it as a reward - you’ve worked hard on something and get to write a blog post about it.
The problem here - with innovation weeks and two organizations running in parallel, is that you’re facing a huge amount of announcements, and financial media and investors will quite naturally highlight the fancy stuff. For example, I find the Magic WAN Connector important and pleased to see that it’s generally available, but I doubt there’s much discussion from an investment perspective about this particular product/on-ramp. (It’s a physical or virtual appliance that you can simply plug in at e.g a branch office, that in many cases greatly simplifies the transition to a SASE/ZT architecture. It’s something that customers’ been asking for.)
(From your prior post, that I feel is related to the above.) I wouldn’t expect hypergrowth. I’d expect a stacking of S-curves to translate into growth durability rather than hypergrowth. To illustrate, here’s an example of the sum (red) two sigmoid curves (green and blue).
In an environment where growth is slowing across the board in the SaaS/IaaS/PaaS space, one could argue that it’s starting to show. (With the caveat that we’re still in earnings season.)
But product announcements aren’t just about pushing growth. To use one of my favorite quotes from monday.com’s CFO, "There is a lot of innovation going on because we know that without innovation, you die eventually.. Cloudflare does indeed come out with a lot of announcements. For my part, I tend to sort them into buckets.
You’re right, you wouldn’t use it for deployment. Not yet. It’s currently in (free) beta, and - most importantly - pricing hasn’t been released. I’d expect a pricing announcement within a few weeks.