The threat of quantum computing disrupting current encryption is no longer theoretical, with estimates of qubits needed to break RSA-2048 dropping from millions to tens of thousands in just a few years. Simon Petkovic, a veteran of cybersecurity and quantum security, emphasizes that this transition requires more than technical upgrades—it demands coordinated action across four misaligned timelines: regulatory deadlines, adversary actions, vendor roadmaps, and internal organizational capacity. He argues that traditional patching fails due to the root-of-trust problem, where hardware-based keys cannot be easily updated. Instead, organizations must adopt cryptographic agility, allowing a single system to roll out new security standards across thousands of devices. This is best achieved by combining public-key cryptography (PQC) with quantum key distribution (QKD), which provides untappable, physics-based security for high-risk network corridors. Real-world deployments in countries like South Korea and Romania demonstrate that large-scale, production-grade quantum networks are now viable. However, a critical gap remains in verification—the "prove" step—where organizations must validate that their PQC implementations are truly secure. Simon’s four-step framework—discover, govern, protect, prove—provides a structured approach to managing quantum risk, with the most urgent challenge being the need for continuous observability, governance, and independent proof of implementation. Without this, even well-intentioned migrations will leave systems vulnerable to future quantum attacks.
This is The New Quantum Era. I'm your host, Sebastian Hassinger.
I have a confession to make before we start. For a long time, I didn't pay much attention
to post-quantum cryptography. Part of that was that PQC isn't actually quantum. It's
new classical math designed to hold up against a quantum computer. And part of that was that
a machine big enough and fast enough to run Shor's algorithm against real encryption
seemed so far off. Honestly, some of the people warning about Q-Day
said,
seemed a little like snake oil salesmen to me. I've changed my mind, because the estimates
for what it takes to break RSA 2048 have been dropping fast. In 2019, Craig Gidney and Martin
Akira at Google put the number at about 20 million noisy physical qubits. That number
stood for years. Then, in 2025, Gidney pulled together a set of algorithmic improvements
and brought it down to under a million qubits, running for less than a week.
In February of this year, a new Australian startup called Iceberg Quantum published an
architecture they call Pinnacle. It uses quantum LTPC codes instead of surface codes, and puts
the number at under 100,000 physical qubits. That's another order of magnitude. And it's
right around where IBM's public roadmap lands in the early 2030s.
Then, in March, Oratama came out of stealth. It's a Caltech spin-out, with Dolev Bluestein
as CEO and Manuel Endres and John Presco on the team, and they published estimates for
neutral atom machines. Their headline is Shor's algorithm with as few as 10,000 atoms.
Their more realistic scenarios are a bit bigger: about 26,000 qubits to break 256-bit elliptic
curve cryptography in about 10 days, and about 100,000 qubits to break RSA 2048 in about
three months. A few months later, they raised a $300 million series average.
And, just in the five weeks since I recorded this conversation with my guest at Quantum
World Congress, Alice and Bob published a new estimate using their cat qubits: about
23,000 physical qubits to factor RSA 2048 in a little over three days. In 2023, their
own estimate was around 350,000.
To be clear, all of these are theoretical resource estimates. Nobody has built these
machines.
And every one of these numbers depends on assumptions about error rates and speed that
still have to be proven out in hardware.
But the trend is unmistakable. In seven years, the number has gone from tens of millions
of qubits to tens of thousands.
Q-day no longer feels like science fiction to me.
So I was glad to sit down with someone who has been working on this problem for a long
time. Simon Petkovic is the Quantum Security Chief Officer at IDQuantique, which is now
part of IonQ. He started working in the field around 2018, when he left BlackBerry to join
a Canadian company called Isara and built what were then called PQC toolkits. NIST hadn't
even finalized its first draft standards yet. Before that, he spent more than a decade at
BlackBerry as VP of Security Advisory. That included negotiating with intelligence agencies
in the Middle East and Asia over lawful intercept access to BlackBerry's encrypted communications,
and working on a project that followed the news that Angela Merkel's phone had been compromised.
Simon didn't come to security through Quantum. He came to Quantum after a long career in
security.
What I like about the way Simon thinks is that he doesn't see PQC migration as a technical
problem that just needs a deployment plan. He sees it as a coordination problem across
four timelines that don't line up, which he calls the four clocks. We'll talk about his
framework and how he communicates that to customers in our conversation.
We also talk about the root of trust problem, which is why you can't just patch your way to
Quantum's safety. Simon makes the case for layering QKD with PQC rather than treating them as rivals,
and he describes what production QKD networks look like when you're running thousands of
kilometers of them in South Korea, Singapore, Romania, and Germany. Simon also has a four-step
execution framework that matches his four clocks, discover, govern, protect, and prove.
The prove step is the one that stuck with me.
Checking that your PQC implementation actually holds up is a lot harder than checking a config
file, and most organizations haven't gotten that far yet.
If you're in the Quantum world and want to understand the security side, this is a great
place to start. And if you work in enterprise security and you're trying to figure out how
seriously to take the Quantum threat, I think you'll find this useful too. Simon makes it
easy to follow without dumbing it down.
Before we start, I'll be in Chicago December 8th through 10th for Q2B26 and the Chicago
Quantum Summit. It's the first time in its 10-year history that Q2B's US conferences left
Silicon Valley, and it's co-organized by QCWare and the Chicago Quantum Exchange.
Speakers will include John Martinez, John Preskill, Scott Aronson, and many other luminaries from the
quantum industry. The new quantum era is a media partner, and you can get 20% off registration with
the code TNQE-20-CHI. If you run a startup, Duality's pitch competition is taking applications
until November 20th. Links are in the show notes.
And now, my conversation with Simon Pekovic.
Simon, thank you very much for joining me. I hope you've been enjoying the Quantum
Summit.
Thank you.
Thank you.
Thank you.
Thank you.
at the Quantum World Congress. I am really, I've been looking forward to this conversation because I think, you know, you're through the ID Quantique acquisition, you're now a part of INQ, you're now in charge of quantum security at INQ. And I think you're really unique in that not only have you got a track record of almost a decade in quantum security, PQC, QKD at ID Quantique, but you actually have a very long career in security where you've
grappled with other novel and emerging sort of threat vectors to the enterprise and to the overall security context. So how does that long track record and vast amount of experience in the broader context of security sort of inform the way you look at the quantum threat to security today?
Happy to be here, Sebastian. Thank you. Just to clarify, my role at IronQ is Field Quantum Security Officer. It's one of those new roles, new title.
Yeah.
Because everything's changing.
Yes.
Everything's, we're getting into a new age where things are accelerating. We're going to get to that. But back in the day where probably you started as well.
Yes.
In the 90s, late 90s, during the dot-com boom, I was at IBM. I was at a Department of Defense before that at IBM. And I had to learn it, all of the tech from the ground up, you know, Unix, Windows, servers, databases, everything.
Every single building block. And the most exciting part of it back in the day was that invisible thing that kept it all secure. And that caught my interest, cryptography, encryption.
Yeah.
How does it look on packet level? What can you do with it? Why you cannot open it? How the keys get exchanged? How the internet works? What the value of Diffie-Hellman protocol is? And how, you know, if that never happened, if we definitely never came up with a public encryption,
we'd never be able to have internet in the digital age. And now we're looking at this and it's the end of that golden age. I call it, in the blog, I call this 30, I don't know, is it 35 or so, 40 years almost, in 74 when he came up with it, the big cryptographic calm. Maybe there's a better marketing term for it, but it's a calm.
Right.
Right.
Or any type of public key encryption that requires, you know, the factor of large primes is now vulnerable or will be vulnerable in the near future. So you mentioned the blog. I really want to jump right to that because I think it was a really great framing. You had the four clocks of quantum crypto, right? That was the PQC migration. That was it. So what are the four clocks?
I put it that way because I believe everybody has to zoom out before they zoom in.
And maybe just worth mentioning, when I started looking into this problem, it was between 2018 and 19, and it was with a Canadian company called Aysara.
When I left BlackBerry, I went to Aysara and we started looking at PQC toolkit when there was just some very early drafts of PQC back in the day.
And what became very obvious to us very immediately was, because all of us in that team, and you had Scott Dotsky and Mike Brown and myself and a couple of other great guys, Al Panazic, mostly all BlackBerry security guys.
What was obvious after we spoke with a couple of our friends who were CISOs and big banks is that it's just, it's too much of an effort of a forklift that nobody will be able to do.
And we immediately started looking at how do we decouple cryptography out of everything where it sits invisible.
It's got to become visible.
And we knew back then the new math may not have another 40 years of calm.
Right.
Because of the quantum threat.
Right.
And it's so unpredictable, you know, how AI will run on quantum and what new attack vectors will emerge.
And it was so.
Back then we had like three or four bullet points that, um, that have held.
And one of them was we got to decouple it.
People will underestimate it.
And there was another one where it's coupled into the infrastructure and infrastructure comes from vendors you work with, which is so normal.
Everybody's very heterogeneous, but then this cuts both ways.
It's a, it's a blessing and a curse.
You know, it will help you, but it will also potentially slow you down.
Mm-hmm.
And when I was drafting the block, I, I zoomed out to, you know, stratosphere and try to recompile all of these thoughts over the past eight years for a reader who may now have this problem in front of them in that organization with a government enterprise, critical infrastructure, financial, and they're exposed to so much noise Sebastian right now, if you go to LinkedIn and logs, you get a lot of, uh, fast.
Yeah.
I don't want to call it FOD, but you get a lot of noise where people zoom in, in particular, uh, area pro problems, problems of that that need to be tackled.
But it may not make sense to somebody who just started.
Right.
And this is why I came up with, with breaking it down to four clocks on one axis and another axis, four things in a framework that somebody needs to do.
And the four clocks are maybe worth mentioning.
Yeah, please.
The regulator clock.
You gotta be aware of it.
We can discuss it.
The adversary clock, which so debatable and people have opinions on it, but it's real and it's real.
And the next one obviously is the vendor clock.
Also one that people may underestimate and, and not fully understand how it will impact them.
And the last one is the internal, um, institutional clock and that one is your ability to have the right people with the right skills.
And the time and the money and the logistics and the change management to actually overhaul everything and do the migration.
Right.
And there are more clubs than that.
Yeah.
We can come up with 20, but these four, I think sum it up.
Yeah.
Yeah.
So there's so much to unpack there, but, but the first thing I want to go back to before we dig into the, the particulars of the clocks one at a time, because I think that's a really good framing, but I just want to emphasize the reason you thought decoupling.
Was critical back when you started thinking about this eight or nine years ago was because we can't presume there's going to be any led up any moment of calm from now on that, that may have been a one-off that, that quantum calm or sorry, that crypto calm when we could rely on, on public key encryption, um, uh, and RSA and the rest of it.
So essentially you're saying the combination of quantum computing shores, algorithm, AI.
And unknown new technologies may mean that we're never in a rest state ever again for security.
That's terrifying.
That is, you know, I think as an it defender, uh, and that, that is terrifying.
And here is a terrifying thought that most people do not probably dare asking themselves, but I have to, because I've been thinking about this problem for, for a long time is it will take most companies.
If everything goes well.
Well, about 10 years, uh, a lot of people are a lot of money and about 10 years and a lot of coordination to get to saying, okay, we are now PQC fully PQC and we, we have reduced the threat and most of, if not all of our systems are now compliant.
Right.
What happens?
Here's the first question.
What happens if halfway through.
Right.
This huge forklift.
Right.
And this releases another set of drafts.
Right.
For whatever reason that may be.
Right.
I mean, these are the current PQC standards are in, in the scope of things, basically brand new cryptography algorithms.
And the history of cryptography is that as soon as you expose an algorithm to a wide audience, people start finding holes.
Yes.
And yes.
And I hope it holds for another 50 years and everybody does.
We just don't know because the advancement of compute and IonQ is the forefront of it.
So I have.
That's an advantage from my point of view, because internally I have insights into the compute roadmap and how fast things are accelerating.
We released recently a very strong paper on, on the estimates for breaking one of the elliptic curves.
Right.
Use in blockchain and IOT and so on in under 20,000 physical qubits.
That's on the horizon.
In a reasonable timeframe.
Yeah.
And if you look at our roadmap, we predict.
To have this number of qubits in two years.
Okay.
So believer or non-believer, um, going back to the clocks, the regulator clock says 2035.
Right.
I can say that a large organization, heterogeneous organization that has not potentially started yet is going to be late to that.
I don't think they can do it in under nine years.
I don't think.
10 years sounds reasonable.
I don't think they can do it in nine years.
Uh, and why, because we have to look at the history of it, it projects and history of some cryptographic changes in the past when Shaw one got deprecated.
Right.
It took over 10 years for just one small bit in cryptography, and now you have to overhaul everything.
So history, history is a great teacher and yeah, we can learn a lot.
Yeah.
Yeah.
I wanted to circle back to, uh, your Blackberry experience.
Because I think that's a really unique moment in time because as you said, there was sort of the, the mainframe to distributed evolution of security, but then personal smartphone devices, um, became this, this huge wildcard in the, the it landscape because they were personal devices acquired personally that came into the enterprise that start enterprise data started dropping into those machines and there are those devices.
And then they were out in the wild.
And so what, what.
The, what lessons did you learn from sort of the, the, uh, the transition that the adaptation of enterprise it security to mobile devices when you're at Blackberry?
That was a huge learning curve and, and you're right.
Um, the ability to have that security carry over from your it environment all the way to a mobile device was the main thing.
In addition to usability, obviously being user friendly, right.
That made it to the enterprise end of the nineties, early.
2000s.
Um, it's, it was also something that caused a lot of trouble.
If you Google Blackberry versus, you know, intelligence services globally, you'll see that the same technology that made it popular cause a lot of, a lot of issues with, with governments that were, um, unable to conduct their surveillance problems on, on digital communications because of that Blackberry security.
And I had a fortune or misfortune to sit in all of those rooms.
Uh, you know, across the world and in the middle East and in Asia, uh, I'm not going to name any country, but sometimes I went in, you know, 10 times over and flew on 24 hour notice because of this.
And so you learn it both ways, uh, what, what it can do for you in terms of privacy, in terms of enterprise security and what, what it does to, to national security programs on the flip side.
And you've got to be aware of it and, and it's a real thing.
When you look at quantum security, uh, it's just a part of cybersecurity, but you also have to be aware of that, that we have to protect our secrets and our privacy and on the enterprise side of things.
And what we do, we cannot necessarily always think what that may do to national security.
It will strengthen it, but you know, it's just, you know, on the flip side, yes, we're getting into a very uncertain world.
Yeah.
Yeah.
So that.
Brings us very nicely to the second clock, which was the adversary clock.
Do you think that the concerns about harvest now and, and decrypt later are, are warranted?
oriented or is that sort of in that category of FUD?
It's one of the most frequently asked questions.
If somebody, let's, let's go through it very quickly.
So, um, uh, harvesting wireless data is trivial.
Yeah.
Anybody can do it.
Yeah.
Okay.
Harvesting data from a wire is almost as trivial.
Yeah.
Um, dumping the data in the, in the storage is very inexpensive.
You and I can conduct harvesting.
Yeah.
I've run Wireshark myself with two hour, two hours of training, a shovel, a device
that costs a couple of thousand dollars, and we can dump probably a petabyte of
data from, from a fiber line, um, for $10,000 a month and wait those couple of
years and open it up and so, but it won't, it cannot be in the, like, nobody
will harvest data and put it in the news.
You see it in the news only.
Once it's open once on the dark, when, what, what's is being sold.
Mm-hmm this one of those attacks where you see no traces of the breach.
Right.
And this is why, this is why it's being, you know, discussed and, and, and, you
know, some people may not believe that this can bite them, but it's real.
And here's the thing.
If you do going back to the clocks, if you do the migration over those, say 10 years,
between now, and when you're fully PQC compliant, you are exposed mm-hmm in, in
different order, as you move through the migration, you reduce it over time, but
you're essentially exposed at different, at different spots in your infrastructure.
The question that most people have not asked their boards is whether they're okay
to accept that exposure between now and then, and that exposure is the exposure to.
Right.
I'm guessing the answer is no.
And here comes the question, which is those clocks are not synchronized.
What your regulator says.
Right.
What the adversary can do.
Right.
They're not aligned, aligned.
What vendors will deliver to you.
Right.
Over the next few years is not aligned with the first two.
And then you have your own ability to implement all of that, which will compound
the time, meaning that change management and all of that, when people, I
asked in the blog for people to look at their own historical timelines or how long
it took them to replace a set of routers, a set of switches, set of servers, to get
an idea of how much time they will need to add on top of the vendor clock.
And then the question is, do you want to run your risk management program mm-hmm
on your vendor clock mm-hmm or your own mm-hmm in internal institutional risk
clock and, and, and the two are not aligned when you say replace the switch,
replace router.
Do you think this is, this is gonna necessitate hardware upgrades,
hardware replacement, or is this swapping one protocol for another in
firmware or in, in operating systems or other, uh, framework libraries?
Basically.
It, it's one of those topics that's not very open and public mm-hmm it's been
for a couple of years, almost a taboo mm-hmm because in many instances, eventually
you will have to replace everything.
Mm-hmm and that's because of the root of trust and the root of trust mm-hmm is the
Mm-hmm public key.
That's.
Yeah.
Burn into the hardware.
Right.
When the phone boots, it needs to validate the OS.
Yeah.
The signature of the OS.
So you can update and patch software on your phone.
Yeah.
To have TLS that's PQC compliant.
Yeah.
For connectivity, which is good.
Yeah.
And you can upgrade a lot of libraries, but essentially your hardware, your
phone will be untrusted.
Right.
Because somebody could forge the OS.
Or router or switch.
Or router or switch or server.
Oh my God.
So software will do a lot, but eventually.
The question remains how the root of trust problem will be solved.
And that varies between different vendors and manufacturers and so on.
Yeah.
Cause some are in, some aren't able to update that root of trust over the
era, essentially depending on the architecture, but in general, one should
not expect that they will be able to get away with a lot of patching.
Because it just won't work that way.
The scale.
When you start thinking about the scale, it really does make your mind real.
Um, but, but, okay, so, um, that's kind of the, the classical side of things.
You're now, um, part of IonQ, which is going through this acquisition strategy.
You've got this vertically integrated stack now with quantum networking,
quantum security, quantum computing.
Does that make your life easier?
Thinking about the, the holistically, uh, the, the, the quantum security side of
things, is that sort of the way you think of that is, is through that whole stack?
Yes, it does.
It does help.
And, um, if you look at it from an IT defender point of view, from a CISO point
of view, and I keep wearing those, you know, lenses all the time to solve their
migration problem, uh, not just from the regulatory perspective and I, and I
split organizations into two, regulation driven, tick a box, do the minimum and
risk driven, risk driven are those who are talking to their boards around our
organization, are we going to, uh, accept the exposure or not and how we can
mitigate it and they may move faster than the risk clock.
Right.
Um, and so when you look at it from that perspective to solve the migration
problem, and you look at the cryptography living in so many spots in this, across
the whole estate today, if you look at the tooling and the tooling could be a
whole other podcast we can do, you cannot do a lot of it without very solid tooling
because today's tooling is.
Pretty fragmented and you get pieces of truth from cloud, from, um, what's on the
wire, typically TLS sessions, you see some of that, but these are all fragments of
the truth, 3%, 7%, 12%.
What is missing is having it covered a hundred percent.
And then you have to govern it.
You have to reconcile it somewhere.
So when a digital certificate flies on the wire, you need to understand when it was
cut, who it was used by.
When it changed and you have to see it across the time and where it
passed along the way and so on.
And no system, it's like a system of record for cryptography.
If you think of it that way, this is where the industry needs to get to and look at
it from a CISO procurement point of view.
Will every CISO go and buy 20 different tools to cover that?
How are they going to solve this problem?
And so what's a blessing for me in IOMQ is that we now have so much of Brad's
path in, in what we do that we can cover so much ground and help cluster these
different problem sets and value adds into a logical procurement where you don't
have to buy 20 tools and this is what we're doing and that's the strategy and
approach.
That's really interesting.
And I mean, when you say decoupling, um, does, do you anticipate quantum, uh,
encryption methods?
Somehow being part of that decoupling and, and also the other part of the question
is when you say decoupling, we just talked about sort of the, the hardware, uh, uh,
proof of trust, uh, in a phone or a router or a switch, does that decoupling turn that
into a zero trust environment where you, you don't know who owns the box?
Actually, you have to create a trusted environment somehow within an untrusted
space.
Yeah.
The age of great decoupling, decoupling is not easy to do.
Yeah.
So it will move different ways for different classes of problems.
Let me give you an example for the network estate.
Uh, there's a different example for the application estate.
Every company has a couple of hundred, a couple of thousand applications that have
coupled cryptography.
Every network asset has cryptography deeply coupled and the PQC version of what
needs to be in that device may not even have enough resources, CPU, power memory,
and so on to run, which.
Even without, you know, thinking of root of trust, you may need to upgrade it anyway,
but say you don't have to, the question is then when the next migration comes, you have
to do the forklift again.
And here's the agility part.
What if you could, um, for say 10,000 network assets, you have touch only one spot to roll
to the next, to roll to this instance of ML, 10 key exchange.
Mm-hmm.
Okay.
And if it changes.
If it changes to whatever is the next key exchange, you touch it once and all of
these network assets accept the keys from this trusted quantum safe platform.
This is the cryptographic agility.
This is what we've built.
And this is what essentially cheats the clock, but with a new architecture.
So you don't have to weigh those 10 years to move every asset on the network.
You touch one software platform that spans your whole.
The whole network infrastructure and you roll to the next generation of keys.
And, and the beauty of is that it uses PQC by itself and, or it can also use QKD, which
is physics based cryptography and where you see that playing a role is for the companies
where, you know, I was told once a line that, that I memorized and somebody very smart told
me why would.
somebody run a trillion dollar franchise, which most of them are,
on a single set of assumption so typically in a in a sound risk management platform you would not
uh framework you would not do that and so what they do now they look at combining math
with physics and they do use physics where it matters and where it matters is really in risk
profiling and threat profiling is you have some corridors or a network where thousands of apps
traverse and move through the same corridor and this is where your surface of attack is the highest
for harvesting yeah and so protecting these with physics is is becoming increasingly um
popular as a risk management strategy yeah let's talk some more about qkd because i mean pqc now
you know there's standardized algorithms coming from from nist uh which are the sort of the first
wave of of adoption there's obviously questions about
how robust they are we'll find out that we need to be agile there's there's skepticism about qkd
i think um on the operational side primarily where uh it's it's seen to be impractical or not
adding much for the level of effort but you you have a different opinion about that
this is in fact all correct okay it has to be seen in a pragmatic way
the way i see it and there's been a long divide and and and chasm between those two
pqc and qkd camps yes i'm not part of that i'm looking at as looking at it as an it defender
so qkd is physics it cannot be tapped in a traditional way it holds no matter how much
quantum compute advances or ai advances you cannot defeat it by advancement right you can
defeat it by a poor implementation if it's not done right right if it hasn't been assessed
but it comes with limitations one of them is the distance and and typically they go up to 30
decibels which is typically 80 90 miles between alice and bob and then you have to repeat that
setup so it's number of point to points right but we have overcome those limitations by putting that
cryptographically agile software layer on top i said that leverage is pqc to ensure you can do
this point to point in a very large meshed network environment
so we we've built networks that go through ring topologies and daisy chains and meshes
where qkd serves many many data centers and this is a way to manage the risk and why it is because
then you have flexibility on those late on those corridors where you have it and when the keys are
served to say ipsec ipsec devices encryptors and routers or magsec you have the flexibility on
when you're gonna do the pqc
native vendor pqc and when you're going to roll it again if there is an exchange i see because the
physics layer holds no matter what happens i see with a matter you're essentially protecting the
the gaps between those 80 90 mile uh qkd links with pqc essentially just the hops yes just the
hops essentially yeah because every 90 miles you cannot tap right and then when when this fit you
know this connection is qubit qubit path right it's in a device
in a rack yeah it has to move to a traditional as key yeah which is digital over say a 10 inch cable
and this is the classical right path to move to the next qkd and then it goes again as qubits so
our shovel and our two thousand dollar device doesn't help with that 10 inch cable you cannot
along the way yes yeah and then you need physical access that's and then you have to break through
traditional uh classical
encryption if you want to attack over a patch cable right including pqc and classical encryption
and so on right right yeah i'm i'm also curious because of the id quantique um uh sort of
deployment specifically in romania this there's more um infrastructure deployed in romania thanks
to id quantique than any other country in the world is that has that served as sort of a lab
for you for for seeing what the
what the deployment looks like what the implications are operationally
the biggest learning on on scaling qkd networks and and just to be clear we don't favor
neither qkd or pqc so we are very pragmatic we want to do what makes sense for the enterprise
and have them leave them with a lot of choices to mitigate their risk but the biggest learnings we
had was korea korea was first to go with a super large network north to south the south korean
network with sk broadband and with uh government's interest in the project and and it was built as a
three ring network uh that are interconnected and it's operating as high as la and has been
for the for for a number of years and then we moved to many other projects one of them is with
singtel in singapore done in collaboration with a with a telecom and the regulator where we build
a large ring and on every network we build a large ring and we build a large ring and we build a large ring
in the data center where you have that trusted relay which you and i just described there is a
hub and spoke where the keys move then in spoke uh connections to many enterprises that can
leverage this as a service interesting and then romania happened and poland happened we built
born berlin link with deutsche telekom and there are many other installations we've done one in
canada uh in quebec and we have many users in the u.s right uh both in
financial sector uh government agencies who have purchased it some for lab testing some for
experimentation but overall we run um over i would say five six thousand miles of this today
about seven to eight thousand kilometers in production grade high sla qkd networks that's
really interesting and those those networks would be that what you described before where you can
if you needed to update
uh because they're they're fully decoupled you'd be able to sort of touch once and enroll a new
a new uh cryptography standard through through those networks exactly so if you look at that
network it's built on bb84 protocol it's a prepare and measure qkd um which means that
it it holds the way it is and you can do whatever you want above it it will serve
quantum safe as keys exchange with an untappable method
to encryptors and then you can have the fallback in the encryptors done traditional way you already
your encryptors to pqc so what happens in the encryptor is kind of interesting you don't fully
replace what the encryptor does encryptor keeps doing exactly what it's been doing forever doing
as encryption authenticating and then it can also do traditional key exchange classical or pqc
but what happens with that etsy 014 interface which is a key key retrieval interface it requires the
key
from underneath from our solution and our solution can be either pqc or qkd so what happens in our
solution qrng quantum random number generator right generates a key the key is exchanged over
qkd or pqc given to the encryptor and then encryptor literally merges that key combines it with its own
and does the as encryption meaning that if somebody taps that line and gets even that in-band key
it cannot break it because has no access to out-of-band key right or qkd which is required for
for decrypting and breaking interesting publicly set up interesting so as a way of closing out i'm
really curious because i mean it sounds like not i mean there's huge challenges i mean decoupling
i think is a fascinating and sounds like really necessary uh evolution but it's daunting um we
have an uncertain sort of time frame for decoupling and decoupling is a really important part of the
but let's say 2030 a cryptographically relevant quantum computer appears and uh the there's a
failure to deploy um you know a quantum safe uh set of technologies in some at some scale do you
think those four clocks those four drivers like do you think the the regulator clock the the
adversary clock the vendor clock or the the institutional clock which is sort of the
the the one that's to blame for that fail that prospective future failure if that happens
it that's gonna that's a really tough question uh i would say the the you know we can we can always
say the regulators always have to do a little bit more to push forward but we saw some of that
through the executive order yeah that moved the deadline push pull the deadline in from
2035 to 2013-31 which is a very strong signal yeah and the key exchanges are
earlier yeah than digital signatures which inexplicably says harvest now decrypt later
right should be treated early right so that that was a very good signal for the market to move
but ultimately is the institutional call is a risk management we know the dependencies
dependencies we have on vendors it's it's not a secret right but if the customers demand it
they'll accelerate i mean in other words that the board says not acceptable the cso says i'm
going to need you guys to upgrade this stuff faster the vendors are going to respond and
alternative how do we do cryptographic agility can we use out of band with pqc or qkd which
ionq provides today to mitigate that ultimately sebastian we have four clocks and and maybe one
other framework that that i'm very big on in and after all these years is everybody will have to
do four things continually from now on until the migration is done one is discover the other is
govern the third is protect and the fourth is prove discover govern protect and prove and when
you look at discover we discussed it but it has to be continuous observability things change
overnight yeah so it's not one time c-bomb and we get c-bombed to death i said that on a pal
yesterday c-bomb everywhere just look it up on linkedin just every second pause the c-bomb
that c-bomb is cryptographic truth from many places that changes every day
it has to be continuous it's a bit of a tooling issue there yeah and we we we have a platform that
addresses that as well so it has to be continuous governance can i prioritize can i see it over time
right can i see net can i see what regressed new apps come in and you know right while you're
migrating yeah so there's regression so somebody vibe coded something yes needs to manage the net
progress needs to also manage the next exposure reduction
ultimately if you agree every caesar should be able to say to the board on a quarterly basis
this is how much the exposure has reduced right right so you need that ability yeah so you need
that one source of of truth and control across the estate protect is upgrade patch as well as
move early on some of these corridors with cryptographic agility and and other solutions
that's important and the last one is proof and that one is
tricky because it's a lot of work and it's a lot of work and it's a lot of work and it's a lot
tricky prove is means many things it means reporting it means a testifying that actually
pqc was well implemented for instance if you did it that it holds and you you can do it by you know
looking at the traffic you can look at by in the config on both endpoints but you don't know
until you run it against shore and brute force yeah so you don't know so just think of think of
that one i'm going to leave you with that yeah
yeah and ultimately you have to report and you have a need to have a proof for audit right so
these are the four buckets that sit on the process axis and we have four clocks that i think everybody
should be aware of yeah well simon this has been a fascinating conversation i think the framework is
really robust and i mean whenever a problem is this big you need to have those kinds of
intellectual kind of frameworks to keep things managed so thank you very much for being with me
and thanks for talking to me today and i'll see you next time
my pleasure sebastian
thanks again to simon for joining me the line i keep coming back to is the one he heard from
risk manager years ago and never forgot why would somebody run a trillion dollar franchise on a
single set of assumptions that's the whole case for layer defense in one sentence if you want to
go deeper on the four clocks the link is in the show notes at newquantumera.com along with links
to id quantique and the link is in the show notes at newquantumera.com and the link is in the show notes
coverage of the g7 evian summit and a few earlier episodes on quantum risk resource estimation around
error correction and enterprise readiness that go well with this one if you're not getting the
newsletter yet sign up at newquantumera.com it's the best way to keep up with what's happening in
quantum and if you work in security risk management or enterprise it and found this useful please share
it with a colleague the people who need most to hear this conversation are often the least likely
to subscribe to a quantum podcast
so word of mouth really matters here you can subscribe on apple podcast spotify youtube or
amazon music or anywhere you listen to podcasts the links are in the show notes thanks for
listening i'm sebastian hasinger and this has been the new quantum era theme music by och see you next
time
Podcast Summary
Key Points:
The estimated number of physical qubits needed to break current encryption has dropped from 20 million in 2019 to under 100,000, indicating rapid advancements in quantum computing feasibility.
Simon Petkovic, with decades of security experience, frames the quantum threat as a coordination challenge across four misaligned timelines: regulator, adversary, vendor, and institutional clocks.
The root of trust problem—where hardware keys are immutable and cannot be updated—means patching alone is insufficient; full system decoupling and cryptographic agility are essential.
Quantum key distribution (QKD) offers untappable, physics-based security, especially in high-risk network corridors, while pairing with PQC provides layered, flexible defense.
Real-world deployments in South Korea, Singapore, Romania, and Germany show that scalable, production-grade QKD networks with cryptographic agility are already operational.
The "prove" step in migration—ensuring actual implementation resilience—is critical and currently underdeveloped in most organizations.
Harvest-now, decrypt-later threats are real and operational, with data harvesting being cost-effective and undetectable, making exposure during migration a major risk.
A four-step framework—discover, govern, protect, prove—enables continuous monitoring, risk management, and verification, essential for long-term quantum security readiness.
Summary:
The threat of quantum computing disrupting current encryption is no longer theoretical, with estimates of qubits needed to break RSA-2048 dropping from millions to tens of thousands in just a few years. Simon Petkovic, a veteran of cybersecurity and quantum security, emphasizes that this transition requires more than technical upgrades—it demands coordinated action across four misaligned timelines: regulatory deadlines, adversary actions, vendor roadmaps, and internal organizational capacity. He argues that traditional patching fails due to the root-of-trust problem, where hardware-based keys cannot be easily updated.
Instead, organizations must adopt cryptographic agility, allowing a single system to roll out new security standards across thousands of devices. This is best achieved by combining public-key cryptography (PQC) with quantum key distribution (QKD), which provides untappable, physics-based security for high-risk network corridors. Real-world deployments in countries like South Korea and Romania demonstrate that large-scale, production-grade quantum networks are now viable.
However, a critical gap remains in verification—the "prove" step—where organizations must validate that their PQC implementations are truly secure. Simon’s four-step framework—discover, govern, protect, prove—provides a structured approach to managing quantum risk, with the most urgent challenge being the need for continuous observability, governance, and independent proof of implementation. Without this, even well-intentioned migrations will leave systems vulnerable to future quantum attacks.
FAQs
Recent estimates have dropped significantly, with some studies suggesting as few as 10,000 to 100,000 physical qubits could be sufficient to break RSA 2048, depending on error rates and machine efficiency.
Estimates for the number of qubits required to break encryption have decreased dramatically over the past few years—from tens of millions to under 100,000—making the threat more realistic and imminent.
The four clocks are the regulator clock, adversary clock, vendor clock, and institutional clock—representing different timelines and pressures that organizations face during PQC migration.
The root of trust is deeply embedded in hardware and cannot be easily patched; most systems will require hardware upgrades or new architectures to maintain trust after a quantum transition.
He advocates for layering PQC and QKD—using QKD for high-risk, high-traffic corridors where physical security is critical, and PQC as a fallback for broader network protection and key exchange.
Proving a PQC implementation is secure requires actual testing against quantum attacks, which is difficult and costly, and most organizations have not yet validated their systems in this way.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.