Only 3 video surveillance manufacturers do a billion dollars or more in annual revenue - Hikvision, Dahua, and Axis. Now, Anyvision plans to join that group in just 3 years.
Can a company who only started to ship product last year so quickly catapult themselves into such a rarefied segment?
Inside this note, we examine Anyvision's plan, the challenges they face and the burgeoning facial recognition market.
Yes, selling through the channel. Even if they sell at $1,000 per channel, that's a million licenses to reach a billion in revenue.
Regardless of channel cuts, the bigger issue is what price the market will settle at, over time, and correspondingly how much demand will there be at that price point.
Was there mention of additional product lines being rolled out? I would find it hard to believe a VC firm would sign off on a growth plan that estimated a million licenses of facial recognition a year (it would be possible in China but doing so would likely end any hope of sales expansion in the West), but if they intend to roll out other similar analytics (business, marketing, safety, etc) and/or hardware offerings then it would be at least a little more reasonable of a goal [at least reasonable enough for a VC to accept it]
Insights was supposed to come out in April, didn’t happen, still hasn’t happened.
Their 2.0 facial rec software was supposed to be released March 14, didn’t happen, still hasn’t happened.
There is no gun detection, just marketing brochures.
Body detection was scrapped because it doesn’t work.
The Access Control is inadequate and can’t scale.
Read the Glassdoor reviews (not the fake 5 star reviews posted by the CEO’s friends), I wish I would have. It’s probably the most frustrating, debilitating, and weirdest place to work at in the industry.
I'm guessing Genetec will hit $1B is sales long before Anyvision. Anyvision's growth projections are wildly optimistic. The big questions marks for me are:
1. Product. Do they actually deliver vale?
2. Scalability. Can they achieve AND manage this level of growth?
I've seen cool demo's but nothing to indicate that they can pull this off....
The only way they'll sniff $100M, much less a billion, is if they get some big wins in the consumer space (large handset makers, ecommerce, online banking, etc) for multi-factor authentication. And there's no shortage of competitors in this space either...
Anyvision probably has the best analytics I have seen in a live demo. It's impressive. My issue is that it almost feels like the industry is trying to shove analytics down the end user's throat when they (currently) don't see the value. I think it will get there eventually, but it will take a major shift in the market for mid to large sized customers to make that investment.
The investors have been hoodwinked. There is zero chance that this company will ever reach $1B in 10 years, never mind 1 year. They will be out of business in a few years because they have grossly underestimated how difficult the market is to penetrate.
“The investors have been hoodwinked. There is zero chance that this company will ever reach $1B in 10 years”
I disagree about the investors being hoodwinked. Their investors are professionals and are likely just making a bet in a space they expect to be very big. Of course, I’m not saying that they bet “right” but it’s a calculated gamble.
Anyvision is not Theranos. They have a product. We've been testing it for a few weeks. It's not faked.
However, whether and for whom it is worth the price they are charging for is an important question. Most startups fail for 'normal' reasons, few are Theranos types.
I definitely agree, but I do see a danger from such ambitious (obviously unreachable) goals being set forth- when VC looks into our industry and decides to throw that kind of money (70m$+ for anyvision, 40m$+ for Verkada, etc) towards companies that have a business model seemingly in contradiction to what could be considered reasonable then it may do significant future harm to the potential of newer companies getting seeded. Less VC money also likely carries with it less talent spillover from silicon valley, which could hurt the overall technical development of the industry as well- just my .2¢
Yes, definitely risks, though I see Verkada and Anyvision as fundamentally different.
There's no real technological wizardry with Verkada - it's more financial engineering - how to package video to win over end-users and then lock them in for many years. Financially speaking (i.e., from an investor's perspective) I think it's really well done.
With Anyvision, they are trying to deliver a technological breakthrough in AI, while adding on a new cost to security / surveillance systems. Financially speaking, I think that's really hard.
One big challenge for Anyvision and other analytics company is that the cost of the server infrastructure is high. They require a lot of CPU and/or a lot of GPU processing. So in addition to the cost of licensing one can need 10's or even 100's of thousands of dollars of hardware to run the applications.
One big challenge for Anyvision and other analytics company is that the cost of the server infrastructure is high. They require a lot of CPU and/or a lot of GPU processing.
I think this is true across the board right now. Of the analytics I've seen lately that do some form of facial recognition (Avigilon, Panasonic FacePro, Anyvision) they're all making use of NVidia cards for processing. Panasonic and Avigilon at least have their cameras tagging the faces/shapes and sending back to the server for processing. Anyvision is processing pretty much each frame. So yes, it is more likely to tax the server harder than the other products but also has much broader applications. Avigilon can basically only use their cameras or add a costly analytics appliance for third party cameras... it's also not really facial recognition in the traditional sense but is the least costly/overhead intensive of the bunch. Panasonic FacePro can only use a handful of very specific Panasonic cameras and is still taxing processors and Nvidia cards heavily.
AnyVision will not get to more than $100m based on their current GTM and products. NEC who has been stronger in the face recognition industry than others (not comparing to AnyVision which I think is better technology) is not making enough sales of their product. Everyone is talking about Face but in reality, it's a fracture of the VMS needs, the TCO is very high, and the ROI is questionable. $1B of SW, really??? Its seems to be more of investor relationship PR rather than real estimation.
I agree with your general concerns. I do see 2 structural differences between NEC and Anyvision:
- NEC sells consulting projects for facial recognition while Anyvision is selling a product.
- Anyvision has way more salespeople than NEC selling facial recognition.
To your point, though, I still think there are concerns in executing this but the NEC approach of bespoke custom niche projects never had a chance of generating sizeable revenue. Anyvision may fail in trying to mass sell a productized facial recognition offering but at least it addresses 2 core elements that locked NEC into its small market.
How can you mass sell at this price and such a high HW cost?
This type of application can be a success is you have a regulation that force organizations to use such technology. Like voice recording in banks and insurance, or QA in call centers that forced companies to use Verint or NICE, etc.
There are also other players with as good technology and even better, that sell to automotive which will bring the AI company more revenue stream than VMS or ACS as in automotive its becoming a standard for safety.
Worth noting that if you were to purchase facial recognition analytics from verint at least some of that money goes to anyvision as it's driven by their software.
That's my understanding as well, though Verint is no longer a big player in video surveillance. Best case scenario maybe Verint can generate a few million annually in sales for Anyvision but for a billion dollar target, it's not a significant partner.
Worth noting, Bosch invested into Anyvision and are working on a Bosch Anyvision powered offering so that too will generate revenue for Anyvision but I would still expect that to be modest relative to a billion dollar target.
One big challenge for Anyvision and other analytics company is that the cost of the server infrastructure is high. They require a lot of CPU and/or a lot of GPU processing.
Anyvision also sells the servers. So far they seem to be missing all of the capabilities to upsell features on the Dell/HP servers they are selling. iDrac/ILO, extended manufacturer warranties, and server options are all missing from their price list. It is almost as though they view hardware as a necessary evil. This is different than some other manufacturers.
At least Anyvision has not gone whitebox Foxconn or SuperMicro for servers.
Deja vu of the heady days of ObjectVideo (circa 2004). The technologies have changed but the fundamentals for a company like Anyvision have not. They have a solution in search of a problem. Or put another way; they have a thousand dollar solution to a hundred dollar problem.
There will be projects; Area 51 might be a good call this week ;-); that need advanced analytics to solve serious issues. Some of these will also have budget and proceed. To get the kind of growth Anyvision is projecting they need broad adoption though. AND they need to scale. Meaning that every deal can't involve a Sales Engineer hopping on a plane to do a site survey.
Anyvision's growth projections got them a bunch of money in the short term. That's the story VC's want to hear. It will be interesting to see how they handle reality in 12 months.
Maybe I missed it, but didn't see anyone mention this technology is typically used for about 10% of the installed cameras, for example, around entryways for casinos, transit, and retail, so I see their high valuation a challenge.
Talking with some former AnyVision employees this percentage trended with their known past installations and may have been why someone else referenced the "hoodwink" comment. The AnyVision leadership originally thought they'd get a much higher percentage of the project.
its a smart move for AnyVision to promote the IPVM article, even including the numerous scathing comments.
they’re using the pay-wall as a pay-shield.
so, what a non-member would see of the article in clear text is limited:
as for the comment sections, the random word reveals are (intentionally) not enough to convey the same tone as the full text, and in some cases may seem better than they are, for instance:
and, my favorite:
net/net: the number of people without IPVM memberships, that will subscribe just to read the AnyVision article is likely far less than those that will not subscribe and just click back out.
Well, if it was such a great idea to publicly tell people Anyvision was going to do a billion in revenue why did they not announce it themselves in the first place?
Well, if it was such a great idea to publicly tell people Anyvision was going to do a billion in revenue why did they not announce it themselves in the first place?
maybe because IPVM wasn’t reporting that they said it until this article.
the fact that a respected industry voice would even consider the claim gives it a modicum of respectability, i.e. “Look who’s taking us seriously!”
that combined with the (unintentionally) anodyne lede but without access to the full critique makes it a savvy marketing move.
Anyvision's target for 2019 is $70 million, per that slide, copied in the report. We don't know what they are on track to do for this soon to be completed year but if it's anywhere near to $70 million, we would be surprised:
One, it is just hard to ramp up revenue in this market for an add on offering like Anyvision.
The mentality at Anyvision is revenue = money raised, not sales. So my guess is that the $70m in 2019 reflects the $73M in Series A. They want to raise another $100-$120M in 2020 and have an evaluation of $1B by 2022. There is no way in H E double hockey sticks that they sold $5-10m in 2019. Of course, I’m talking legitimate commercial business. I have no clue what governments are paying them for the tech.
With Eric Holder investigating them, Microsoft debating pulling their investment, #dropanyvision trending on Twitter, insane turnover, countless terrible Glassdoor reviews and low sales, it is going to make raising more money much more difficult. Money they desperately need. They burn through cash quite impressively.
My first reaction was to call #11 out for making things up but sure enough there really is a non-trivial number of tweets with numerous likes calling for #dropanyvision, such as:
On the list of Anyvision's problems, this is probably not in the top 10, but still not a good thing to be dealing with.
The product is decent so far in our testing. However, I cannot see any way for them to reach $100 million by 2022, let alone $1 billion. They may have some humongous accounts in the hopper that we are not aware of, but I don't think even one or two mammoth accounts could reach this goal, especially in light of the increasing pressure regarding facial recognition in various states.
Of course, as a non-China company, Anyvision does not have such access but if the rest of the world adopts facial recognition even a fraction of the levels as the PRC, there could be billions made in 2022 (overall).
However, I cannot see any way for them to reach $100 million by 2022, let alone $1 billion
FWIW, I would agree with you.
A big challenge that I think they may be missing, is that as their odds for reaching $1B in revenue increase, their odds of reaching $1B in revenue decrease :)
Right now there are technically a lot of analytics players in the market, but only a relatively small number that seemingly have the core technology, sales team, and market approach to get to $1B in revenue. At this phase, Anyvision's name comes up frequently when high-quality face recognition is discussed. This is good for Anyvision, and seemingly gives them some decent odds at success and growing the market to the point that there is $1B+ being spent on analytics annually. At the same time, as the market increases and customers start purchasing analytics more frequently, we will likely see more competitors enter the space, and some number of those will be as good, or better, than Anyvision, resulting in more challenges to meeting the $1B goal.
In my experience, there is a lot more to selling analytics than just being the "best", the "fastest", "the most efficient", etc. If you sell purely on those metrics alone, there will always be someone coming along to eat your lunch and be just a little bit better/faster/more efficient.
Sales channel, integrations, support, perceptions of being open, and other factors begin to matter more, once you have crossed a certain threshold of being seen as "more than adequate" for the task. FWIW, I have not seen Anyvision attempt to tackle this part of the equation yet in a manner that puts them on track for $1B.
IPVM conducts reporting, tutorials and software funded by subscriber's payments enabling us to offer the most independent, accurate and in-depth information.
Comments (48)
Undisclosed #1
But they are selling through a channel, correct? So they are likely only collecting 50-60% of that MSRP number.
Create New Topic
Undisclosed Manufacturer #2
I'm guessing Genetec will hit $1B is sales long before Anyvision. Anyvision's growth projections are wildly optimistic. The big questions marks for me are:
1. Product. Do they actually deliver vale?
2. Scalability. Can they achieve AND manage this level of growth?
I've seen cool demo's but nothing to indicate that they can pull this off....
Create New Topic
Undisclosed Manufacturer #3
The only way they'll sniff $100M, much less a billion, is if they get some big wins in the consumer space (large handset makers, ecommerce, online banking, etc) for multi-factor authentication. And there's no shortage of competitors in this space either...
Create New Topic
Alex Wasson
Anyvision probably has the best analytics I have seen in a live demo. It's impressive. My issue is that it almost feels like the industry is trying to shove analytics down the end user's throat when they (currently) don't see the value. I think it will get there eventually, but it will take a major shift in the market for mid to large sized customers to make that investment.
Create New Topic
Undisclosed #4
The investors have been hoodwinked. There is zero chance that this company will ever reach $1B in 10 years, never mind 1 year. They will be out of business in a few years because they have grossly underestimated how difficult the market is to penetrate.
Create New Topic
Undisclosed Integrator #5
Anyvision was my pick to win the product shootout being done by IPVM. Is that still happenning?
Create New Topic
Undisclosed #6
Smart company good product, but such a claim shows they have more balls than brains...good luck to them and I hope they enjoy the ride.
Create New Topic
Undisclosed Manufacturer #7
One big challenge for Anyvision and other analytics company is that the cost of the server infrastructure is high. They require a lot of CPU and/or a lot of GPU processing. So in addition to the cost of licensing one can need 10's or even 100's of thousands of dollars of hardware to run the applications.
Create New Topic
Undisclosed Manufacturer #8
I feel Ray Davis will sell BRS Labs for a billion before these guys make it.
Create New Topic
Undisclosed Manufacturer #9
AnyVision will not get to more than $100m based on their current GTM and products. NEC who has been stronger in the face recognition industry than others (not comparing to AnyVision which I think is better technology) is not making enough sales of their product. Everyone is talking about Face but in reality, it's a fracture of the VMS needs, the TCO is very high, and the ROI is questionable. $1B of SW, really??? Its seems to be more of investor relationship PR rather than real estimation.
Create New Topic
Undisclosed Manufacturer #2
Deja vu of the heady days of ObjectVideo (circa 2004). The technologies have changed but the fundamentals for a company like Anyvision have not. They have a solution in search of a problem. Or put another way; they have a thousand dollar solution to a hundred dollar problem.
There will be projects; Area 51 might be a good call this week ;-); that need advanced analytics to solve serious issues. Some of these will also have budget and proceed. To get the kind of growth Anyvision is projecting they need broad adoption though. AND they need to scale. Meaning that every deal can't involve a Sales Engineer hopping on a plane to do a site survey.
Anyvision's growth projections got them a bunch of money in the short term. That's the story VC's want to hear. It will be interesting to see how they handle reality in 12 months.
Create New Topic
Undisclosed #13
Can someone embed a countdown timer on this thread ending at the beginning of 2022? :D
Starting tomorrow roughly 896 days until 2022 kicks off so they will at least need to make about $1,116,071.43 per day to stay on track.
Create New Topic
Undisclosed Manufacturer #14
Maybe I missed it, but didn't see anyone mention this technology is typically used for about 10% of the installed cameras, for example, around entryways for casinos, transit, and retail, so I see their high valuation a challenge.
Talking with some former AnyVision employees this percentage trended with their known past installations and may have been why someone else referenced the "hoodwink" comment. The AnyVision leadership originally thought they'd get a much higher percentage of the project.
Create New Topic
John Honovich
Looks like Anyvision is now embracing and sharing the report, despite the skepticism of commenters and voters here:
Create New Topic
Undisclosed Manufacturer #16
Did they do $75M in 2019? That would be impressive
Create New Topic
John Honovich
Well, this AnyVision employee is optimistic:
Create New Topic