Darktrace (DARK): Buffettology buying at the IPO


After saying I would hold on, I sold out completely a month or so ago and put the proceeds into Diageo.

Darktrace seemed like a ‘black box’ to me that not many understand, especially after your earlier posts.

I feel I understand Diageo as a consumer of some of their products!



1 Like

I was taking a look at this and reading the prospectus, news articles and history on the company, and was happy to see Maynard had covered it and to read the comments to date.

It is a very interesting company as the idea of AI-enabled cyber security seems very intuitive to me. In a field where threats need to be spotted and neutralised rapidly, it is ideal to have a defence that is able to self-learn quickly from vast unstructured amounts of information. It makes total sense.

It also makes it an area totally vulnerable to exploitation by shysters.

You never know for sure if your cyber security is protecting you. There are two layers to this.

First, cyber threats are a negative knowledge area. Once you know of a threat you know to avoid it. However, the real threats are the unknown unknowns - these are the ones that really worry you. And by definition, you can never ever resolve this uncertainty. It’s a ‘black swan’ nest (although Taleb would take issue with me calling it a black swan when it is really a grey one).

Second, even with traditional cyber solutions or algorithmic AI, if these are very sophisticated, you - ie senior management including senior IT management - may be very reliant on some key individuals either inside or outside the organisation that have the specialised knowledge to understand ‘the code’ and can assure that the cyber security is the latest and greatest - ie it incorporates in its code or algorithms all known threats, and is effective at using this knowledge to detect, prevent and destroy threats promptly.

However, with self-learning AI such as Darktrace, there is an additional layer of uncertainty. You are being told that the machine is effective at teaching itself. This is the great leap of faith with self-learning AI. Now I am not being a Luddite. I think machine learning will make the world better. But in some areas the feedback loop is shorter, more visible and less consequential than it would be in cyber security.

All this makes a business like this a tough call as an investor - a black box as someone above put it. If ever there was a business I would want to see led by an owner-founder who had developed the product and where I had no question marks around integrity in the the organisation, this is it. But it is not in that category for me.


Hi Jerry,

Thanks for the post and welcome to the forum!

From the prospectus the two key AI developers appear to be:

David Palmer

Mr. Palmer joined the Group at founding, and is the Chief Product Officer of the Company, overseeing the mathematics and engineering teams and project strategies. With over 13 years’ experience at the forefront of government intelligence operations, Mr. Palmer has worked across UK intelligence agencies GCHQ and MI5, where he was responsible for delivering mission-critical infrastructure services, including replacing and securing entire global networks, the development of operational internet capabilities and the management of critical disaster recovery incidents. Mr. Palmer acts as an advisor to cyber security start-ups and growth-stage companies from the UK Government’s Cyber Security Accelerator and CyLon. Mr. Palmer’s insights on AI and the future of cyber security are also regularly featured in the UK media. Mr. Palmer holds a first-class degree in Computer Science and Software Engineering from the University of Birmingham.

Jack Stockdale

Mr. Stockdale joined the Group at founding, and is the Chief Technology Officer of the Company. With over 20 years’ experience of software engineering, Mr. Stockdale is responsible for overseeing the development of Bayesian mathematical models and artificial intelligence algorithms that underpin the Company’s award-winning technology. Mr. Stockdale and his development team in Cambridge, UK were recognised for their outstanding contribution to engineering by the Royal Academy of Engineering MacRobert Innovation Award Committee in 2017 and again in 2019. Prior to joining the Group, Mr. Stockdale has served as Chief Architect at Invoke from 2012 to 2013 and at blinkx from 2006 to 2011, and was also a technical director at Autonomy from 2002 to 2006. Mr. Stockdale has a degree in Computer Science from Lancaster University. Mr Stockdale was awarded an OBE (Officer of the Order of the British Empire) in 2019 in recognition of his services to cyber security.

Prospectus implies Mr Palmer has a £18m holding and Mr Stockdale a £30m holding at £5 a share:

So there are owner-founders of sorts in place, although they are below board level. But to be honest they are better off away from the stock-market’s glare with their skills.

The black-box issue is key I think. If you actually know something about AI, Bayes theory, how this software works and what effectively is in the black box, then you can be ahead of 99.99% of typical investors with this company. That will give you a great advantage when not-so-informed opinion moves the price.

Otherwise you will be at the whim of other people’s opinions on DARK’s positioning, which is not ideal given the valuation. You won’t know if a sales wobble is the start of competition having finally caught up or just a typical ‘speed bump’ that fast-growing companies occasionally suffer.

What I don’t understand is how DARK’s algorithms start working immediately on deployment. If an attack is happening at the time of deployment, how can the algorithm know the attack is ‘unusual’ behaviour and not the norm without any ‘training’ data?



I guess I would wonder whether there are generic patterns to cyber attacks, which Dark ought to be aware of, or whether specific companies attract specific attacks, possibly based on the type of business. Either way, Dark ought to be up to speed on the kind of threats. Thoughts?


From twitter, on the DARK matter of share options;

There is an RNS today that DARK is buying back up to 4m shares by 31 Dec 22, partly to satisfy share options which consist of:

3.3m shares to PG (Poppy Guftassen) OBE (CEO)
2.4m shares to CG (Catherine Graham/CFO)

The share options appear to made up of 2 grants/awards:

1. Share options of 2.8m shares and 2.1m shares to the CEO and CFO. These are the legacy incentives, due to Vest May '22. Vesting price between £2.50 (the float price) and £5.00 for the maximum award.

The twitterer, Lewis Robinson, questions the firm paying 72% more than the ipo price, just 8 months ago, and whether it is a good use of capital. I see what he is saying - presumably the company could alternatively just issue new shares and satisfy the obligation that way, keeping the cash to use to expand the business…hmmmm that would not be good for the share price, the cynical side of me thinks, and the ‘total return’ which the next tranche of share options vests on.

I also question the purchase of these shares, because post-ipo investor’s cash is being used to acquire these shares at an inflated price to satisfy pre-ipo ‘legacy’ incentives. Is there something wrong with my thinking? Has a liability from pre-ipo been passed onto the new owners?! It must have been disclosed in the admission document so that is alright then!?

2. From admission to 30 June 24, additional new share options have been granted, vesting June '24

475k to PG and
300k to CG.

Oh, but hold on, what are these additional 325k vested but unexercised share options to PG ?
Oh! these are more pre-admission legacy share options. I just can’t keep up!

It’s alright though, the remuneration committee compare everything to similar executives and organisations.

Maybe that is why the base salary is relatively low in the annual report to June '21, to compensate for the share options? Uh oh, no, it is just 3 months worth of salary, from float April to June 21 in the annual report…or is it 3 months - DARK floated on 30 April - sounds more be more like 2 month’s salary to me!(so the 77k for the CEO is £462k salary p.a.)

But, hold on, there is no bonus, is there? So this might be ok? Oh , sugar, there is a bonus. It’s only 100% of salary, though and one-third goes to charity. No need for a guilty conscience then…

Words fail me.



1 Like

some other, more cynical tweets have initmated the reason for buying now is to prop up the price and reduce the risk of DARK falling out of UKX

1 Like

Hi Ben,

Ok, so £30m spent buying shares at today’s 415p = 7.2m shares.

Share count is 699m, so a c1% ‘buyback’.

2021 AR shows 55m options outstanding. So not just the execs have options it seems. And this buyback will still leave c48m options.

Can’t see this £30m matter being of significant importance in the long run. What really counts is whether DARK’s ‘black box’ algorithm is actually the real deal or not. But that is not quite the easiest subject to tweet about though.



Hi Maynard,

Yes, I agree. BUT it does raise the point about the Director’s fiduciary responsibility/irresponsibility to current shareholders, and by extension the Director’s attitude to shareholders.

Pre ipo liabilities shouldn’t be the responsibility of post ipo owners - just a personal opinion.

I do appreciate the options scheme does not make or break an investment case - however the investment case remains unproven either way at the moment for DARK.

It’s not just the quantum of the share awards- it is their criteria. If the criteria were right I wouldn’t be so bothered by the monetary amounts.

I would feel more comfortable with share option criteria based on ROCE improvement, gross margin improvement, turnover per employee, organic turnover growth or ROE or employee engagement.

Options based on total shareholder return or the quartile the company resides within the ftse 350, effectively based on share price isn’t objective,. It doesn’t ecessarily reflect any business improvement over the period involved. The managers are there to improve the business, not produce relatively short term PR campaigns.

From what I have seen,I just don’t believe this company is run for the benefit of ordinary shareholders. Of course I may be wrong.




The small stake I bought in late Oct/ early Nov based primarily on KAL’s research and bullish view was quickly ditched a few weeks later as the share price simply continued to drift downwards for no apparent reason (at least to me).

Have they got a superior cybersecurity model that competitors will find hard to replicate?
Not proven. And in my opinion, less and less likely.

Does management give me confidence that they have the interest of ordinary shareholders at heart. given the latest shenanegans with options?
Not a bit.

So I am going to remain firml;y on the sidelines and watch this company’s development at a safe distant.

Best, Ram

1 Like

It’s looking more and more as if the IPO was priced at bargain basement level and we are seeing initial market excitement building and then rapidly ebbing. With no news expected until February 2022, the best thing to do is probably to sit on the sidelines