Strategy Decisions
27+ recent decisions
Claim the Core42 Dubai build-out for engineering outright - there is no sales engineering function, so if it gets built it gets built by engineering, and every scope increase still arrives with headcount
Aug 21, 2026 · strategy · high93% confidence
In a DM with Nathan on 8/20, Nathan was worrying about the scale of what Core42 is asking CIQ to stand up. Peter walked the argument to its end rather than reassuring him. He said CIQ does not have a sales engineering function, maybe one day it will, but today if something is going to get built at this company it will be handled by engineering - and that does not mean the people who work here today. He then named the pattern: we have seen over and over for the last year that requests for additional scope have come with a demand for headcount, and me and Bjorn be aligned on that. He asked Nathan directly whether Jimmy, Larry or Patrick were going to build out a datacenter of that scale, and if not, whether Nathan wanted Bjorn managing it or Peter. He closed by absorbing it: all the resources I listed will report to me, so it is not a Nathan problem, it is 100 percent a Peter problem - and asked how do we get you to a place where you never spend another brain cell worrying about that kind of thing. The staffing shape he named was consult where needed, heavy lifting from the Shin brothers plus one or two from Ryan world, then see if another contractor or two is needed.
People: Nathan Blackham, Bjorn Hovland, Ryan Smith
Pre-empt a three-week delivery commitment to Core42 before the deal closes - hand Bjorn the real sequencing mid-call so the number dies before it is said
Aug 19, 2026 · strategy · medium82% confidence
During the Aug 19 Core42 gap-analysis call, with the customer having just widened the scope, Peter went to Bjorn privately in DM to stop the three-week delivery figure from being repeated to Core42. He gave the decomposition rather than an instruction: setting all of this up will take dramatically longer than the 3 weeks we have been talking about. Lots of this cannot even start till we have done the 3 weeks we were planning of getting OSs set up and networked, etc. Then it is setting up identity managment, that could take two weeks on its own (likely one). And then other monitoring etc on top of that. He then named the intent outright - I want it to be a problem for after the deal closes... I am trying to pre-empt a Bjorn special we will have this all done in 3 wks comment made before the deal closes. Bjorn produced the correctly caveated version unprompted: To be clear, I will say that provided there are no external dependencies, that we could deliver original SOW is 3-4 weeks. The new scope is larger. Peter: super comfortable.
People: Bjorn Hovland
Fork the Core42 engagement into two SOWs and let nothing delay the first - land the GB300 paper now, put the expanded security architecture in a separate SOW
Aug 19, 2026 · strategy · high94% confidence
On the Aug 19 CIQ-C42 Gap Analysis Part 2 call, Alex Trafton (Core42 CISO, roughly 10 weeks into the role) opened a scope far wider than the existing SOW - PAM consolidation across 4-plus tools, continuous vulnerability management, AI-specific security, physical and OT telemetry correlation, and future AMD MI350X deployments. The room read it as a much larger opportunity. Peter, facilitating, held the immediate GB300 deployment SOW as a closeable unit on its own and pushed the expanded scope into a separate second SOW. In the debrief immediately after the call he put it to Adam Jackson directly: But you agree, Adam, that for the next couple of days, it is just about getting that first SOW across. Adam: Yeah, 100 percent, Peter, 100 percent. Peter: I just do not want to put anything in the way of that. The same instinct had been set the night before in the pre-call, where the agreed play was to propose a second SOW or a simple addendum rather than expand the current paper.
People: Bjorn Hovland, Adam Jackson, Brian Dawson, Michael Shinn, Nathan Blackham, Alex Trafton
Redirect the entire Core42 solution-review effort away from the people on the call and toward whoever watches the recording afterward - and make every answer technical, never personal
Aug 13, 2026 · strategy · high92% confidence
Peter reframed the Core42 engagement in his Nathan 1:1 and again in the group prep. His read: the Core42 security team is married to a Red Hat solution and is using the calls to collect disqualifying gaps, so there is no answer that converts the room. He therefore instructed the team to assume the people on the call are antagonistic, to stop optimizing answers for them, and to tailor everything to the decision-makers who will review the recording. Concretely: stay calm, concise and confident; give less detail rather than more; say yes we can do that repeatedly; and convert every rebuttal from a personal challenge into a technical statement - the FreeIPA docs confirm session recording is possible, not why are you saying it cannot. He also killed Bjorn framing that the win depends on having the best prepared answers.
People: Nathan Blackham, Bjorn Hovland, Brian Dawson, Adam Jackson, Jonathon Anderson
Treat prototypes crossing from Product into Engineering as a feature rather than a burden - the code is worth two days, the market learning is the payload, and every one arrives with headcount
Aug 11, 2026 · strategy · high86% confidence
Justin raised Elm - a prototype somebody had been building for about three weeks and which was apparently launching the following week - and assumed it would eventually become his problem without the bandwidth to carry it. Nathan had objected in stronger terms that we cannot launch things without engineering being deep into it. Peter sided against both of them and with Bjorn, and gave the reasoning in two parts. On the code: it is a prototype, nothing more than a prototype, the amount of code there is not more valuable than two days, it would take two days to recreate it from nothing, so when it does come over to our side of the fence it is not like we are burdened with a pile of architecture that is terrible and horrible to manage. On what actually transfers: all we are burdened with is what we have learned in the market about what people want from it, which is an awesome thing to be burdened with - because now we are burdened with these two customers wanted to do these things, instead of Brady woke up one day and had a brain fart and decided that this is what a product should do. He then removed the capacity objection three separate times: it will come with headcount. Two, and you are seeing Ascender come your way with headcount, period. Everfox desktop with headcount, Kubernetes with headcount. So it will come with headcount, period. He named the contrast with their shared past - it is a very different model than anything we would have done at Amazon - and pushed the logic further with Bjorn, suggesting Satellite be announced with a go-to-market and a feature list without actually making the software available, to see who rings our front doorbell. He also accepted Justins counter-constraint on feature bloat and added his own: we can cut features too, I do not mind us removing features and seeing if nobody ever notices.
People: Justin Haynes, Bjorn Hovland, Nathan Blackham, Brady Dibble, Chris Wolford
Price the Google live-kernel-patching ask at 2 HC for 3.5 months plus 1 ongoing, and take the honest slower timeline back to Tissa rather than fit their October date
Aug 11, 2026 · strategy · high89% confidence
Tissa asked for a quote on live kernel patching as paid work in addition to the existing monthly full-kernel contract: evaluate whether an incoming CVE patch can be delivered as a live patch, deliver it that way when it can, and still roll it into the next monthly drop. Volume unpredictable - might be 70 percent of a months patches, might be zero. Google wants something working by October so they can test before a January go-live. Peter opened the request to Nathan and Justin with a pre-commitment that removes the usual objection - assume we get headcount to cover this work - then interrogated the shape rather than the number: 2 engineers for 3-4 months and less ongoing, or 2 forever? He landed on 2 HC for 3.5 months and 1 ongoing after that. He then did the arithmetic out loud against Googles date - lets assume a month to hire, and 4 after that to have something, as the reasonable case - concluded I think that gets it to them too late and they will say no, and chose to go back with it anyway: but lemme go back to tissa with that and see where it lands. He closed with lets go fishing and see what we can catch. He also flagged that Tissa appeared unaware of internal discussion on the topic - Tissa seemed in the dark - and asked whether that discussion had ever reached Google.
People: Nathan Blackham, Justin Haynes, Tissa Senevirathne
Core 42 delivery will not land on Nathans org - it needs a purpose-built deployment team, likely based in the Middle East, under Bjorns forward-deployed structure
Jul 30, 2026 · strategy · high91% confidence
Nathan called an impromptu Zoom to make sure the Core 42 work was not silently assigned to him. His read: the work is a data-center build-out, a white-glove deployment similar to how Fuzzball is delivered, requiring a standard-operations deployment skill set he does not hire for - and CIQ has no equivalent of Wolfgang doing Rocky deployments. Peter agreed and made the routing call: if we get across the line on this, we need to be making clear to the account team we are going to need to put together a team to go do this, and it is not the people currently in Nathans org - they may be able to help. He accepted that some of the work will land on Nathans team as guidance, architecture and planning, but not execution. Then he named the likely shape and geography: there is a good chance the small team that does this wants to be located in the Middle East - a team in Qatar flying to the UAE, to Saudi, doing these deployments across the region. Nathan twice offered to own that team and Peter twice declined on his behalf: I think this is going to fit in with Bjorns forward-deployed engineering structure that he is trying to set up, and I do not think you want to own that. I get that you are happy to, but I do not think it fits properly there. Peter added his expectation that it ends up under sales eventually.
People: Nathan Blackham, Bjorn Hovland
Took the Fuzzball pitch to Axiom Space himself rather than letting Sales carry it - and set the closing path at NDA then contract
Jul 29, 2026 · strategy · high89% confidence
On the 7/28 Axiom Space CTO introductory call (Mark Seigle and Ethan Leas, with Tom Shingler, Ramesh Srinivasan and Bjorn on the CIQ side), Peter learned Axiom is building an orchestration platform for sovereigns and immediately concluded CIQ needs to get Fuzzball in front of them. When Bjorn asked whether Tom was pitching it, Peter answered I did. He read out the technical fit in real time - they will need post-quantum work which is already on the roadmap and they seemed happy with that, and they will need Ascender plus updates to run well over low-connectivity satellite links - set the next step as sign an NDA and start talking contract, and named the strategic window: they are just starting, so they are in a position to choose OS cleanly right now. Mark said he wants to close in weeks. Peter told Bjorn yeah if we lose this we suck and posted in the deal channel sounds like it is ours to lose. Ramesh confirmed Axiom is moving forward with the NDA.
People: Bjorn Hovland, Tom Shingler, Ramesh Srinivasan, Mark Seigle, Ethan Leas
Google ELTS: pay or I stop the work - and CIQ eats the disputed 150k out of the existing discount to remove the excuse
Jul 29, 2026 · strategy · critical92% confidence
In the 7/28 CIQ ELTS payment discussion with Google (Madhu, Alyssa, Arun, Katur and Kelly Hall), Peter went in deliberately hot and said so. Madhu was trying to hold Kelly to account for who made a technical decision back in February, with the implication that if CIQ made it Google would not pay for work since then. Peter refused the premise - these are technical decisions, they get made by the team, they get recorded by your people as the decisions of the consolidated team, and then we move forward - and then removed the money as an obstacle by offering to absorb it against an existing concession: I gave you a 250,000 dollar discount over here, so you are bitching about 150,000. I will shake your hand right now that that is on us. That 250,000 discount just turned into a 100,000 discount. With the excuse gone he put the actual question: are we getting paid or are we not getting paid, and if we are not getting paid I am taking my toys and going home. Madhus boss and others stopped the meeting and committed to closing it out by the next day. The call also ended abruptly, which Peter read as someone telling Madhu to stand down.
People: Kelly Hall, Chris Baek, Bjorn Hovland, Nathan Blackham
Sensitive Decision
Everfox desktop - CIQ is an integrator, not a builder of novel desktop technology; convert the ambiguity into integrator-shaped requirements
Jul 27, 2026 · strategy · medium78% confidence
Max raised that the Everfox desktop requirement had reached him third-hand through Nathan as something about launching multiple GUI programs with some running in a special memory-protected mode. Max said CIQ can be great as an integrator pulling patches together and writing tests, but cannot build brand new security-critical technology into Linux. Peter agreed - not with the tiny team that we have - and made that the operating constraint. He tasked Max, with Bjorn and possibly Brady, to convert whatever Everfox is actually asking for into a series of requirements that fit an integrator posture, and stated he does not want CIQ designing a new desktop. He offered the shape he hopes the answer takes: which desktop are we shipping, which variant, and is there a set of a few applications we ensure run.
People: Max Spevack, Bjorn Hovland, Brady Dibble, Nathan Blackham, Ramesh Srinivasan
Definitive no to LG U+ cross-distro package support - the test is whether we would ship it down the middle for everyone
Jul 27, 2026 · strategy · high95% confidence
LG U+ wants to run DNF update on RHEL servers pointed only at CIQ/Rocky repos, which touches bootloader packages and breaks the system on reboot. They rejected the exclude-packages workaround. In the LG - Next Steps meeting Peter refused to let the discussion be about whether the ask is a bug or whether it closes the deal, and reframed it to a single question for Justin: is this something we would do down the middle for everybody, or is it one-off work for LG. Justin answered we cannot support it. Peter then gave the definitive verdict: no, this is not something CIQ Engineering is going to support. He posted the same conclusion in the temp-lg channel after aligning with Bjorn, and handed customer messaging to Bjorn with Tommy and Art.
People: Justin Haynes, Ryan Smith, Brady Dibble, Bjorn Hovland, Arsalan Zaidi, Melissa Kivisto
Clear Greg to proceed on the open-source AI-training-data initiative with no CTO gate
Jul 24, 2026 · strategy · medium78% confidence
On the morning of 7/24 Greg told Peter by DM that the conference had reinforced for him that an open source source of AI training data is one of the most important things needed right now and that he has to hurry. Peter cleared him within about 24 seconds and made a point of saying the analysis was already done: get on it, I have been thinking about it and I do not have any concerns. Greg replied On it. Peter attached no metrics, no owner, and no scoping ask - he cleared a lane rather than taking the work.
People: Gregory Kurtzer, Steve Wallace
Reclassify the GDC MIP-listing disclosure question as a contract-timing call and route it to Kelly Hall with a decision criterion
Jul 24, 2026 · strategy · medium80% confidence
On 7/23 Brady Dibble escalated to Peter, Bjorn, and Kelly that CIQ still had not told GDC the 8/8 MIP listing was at risk because the ESV certificate had not arrived from NIST - and that atsec had flagged NIST not asking for review or comment, typically meaning they had not reached the certificate. Justin Haynes pushed back directly with why would we not let them know. Peter did not answer the disclosure question. He assigned it: this is a Kelly Hall call, and supplied the criterion - whatever path is most likely to get a contract signed in the next couple of days. The thread resolved as non-disclosure, with Bjorn stating he would not flag it and Justin later confirming they did not bring it up and neither did we.
People: Kelly Hall, Brady Dibble, Justin Haynes, Bjorn Hovland
Slow-roll Googles request for kernel-build documentation until the contract is signed
Jul 24, 2026 · strategy · medium87% confidence
On 7/23 Google asked CIQ for documentation on how it builds its kernels. Peter decided to slow-roll any response until after the contract is signed, and said so in both a group DM and directly to Justin Haynes - correct, at least until contract is signed, after that I am fine. He read the motive out loud: they are asking so they can build their own.
People: Justin Haynes, Bjorn Hovland, Chris Baek
Release the Google 6.18 build; hold the line on future deliverables until Google resolves the LTS question
Jul 24, 2026 · strategy · high96% confidence
On 7/22 Peter lifted his own 7/21 hold on the Google/GDC 6.18 image. He connected with Bjorn first, then told Justin Haynes in #google-partnership-governance and in DM to Kelly Hall: give Google the build. He emailed Tissa Senevirathne at Google the same morning confirming the release, restating a firm stance that the LTS work has already been provided to Google, and stripping Mahdu and Alyssa off the thread. He paired the release with an explicit condition: future deliverables are held until Google sorts out the LTS commitment.
People: Bjorn Hovland, Kelly Hall, Justin Haynes, Tissa Senevirathne, Brady Dibble
Pull the business side (Kelly) into the FIPS 6.18 CAVP-retrigger vs fork call
Jul 21, 2026 · strategy · medium82% confidence
Justin flagged that upstream 6.18 LT landed hundreds of CVEs/fixes touching the crypto code being FIPS-certified, forcing a choice between re-triggering CAVP (risking the 8/8 Google delivery) or forking/reverting the kernel (a growing maintenance burden). Justin was inclined to retrigger CAVP and risk the timeline. Peter did not make the technical call himself - he decided the business side must be convened, with Kelly Hall involved, because the choice puts the Google 8/8 delivery at risk.
People: Peter Nelson, Justin Haynes, Kelly Hall, Bjorn Hovland
Pause the AI Data Management policy approval and route it to Greg
Jul 21, 2026 · strategy · medium83% confidence
Steve Wallace asked for Peter Google-Doc approval on the AI Data Management Policy (an effective-date fix). Rather than approve it, Peter flagged that Greg is now exploring training open-source models and said the whole approach may need to be revisited, telling Steve to get the doc in front of Greg to ask what he wants to do. Peter declined to ratify a policy that an in-flight strategic pivot could immediately obsolete.
People: Peter Nelson, Steve Wallace, Greg Kurtzer
Approve GPU hardware spend and drive hours-not-days urgency for the Fuzzball/Arcee test cluster
Jul 21, 2026 · strategy · medium85% confidence
Peter approved procuring and standing up GPU hardware for a Fuzzball test cluster - routing ~240K of GB200 cards he had in hand to the Reno office, approving a ~60K server plus InfiniBand switch (and more if needed), and greenlighting installing cards into the office servers Brady was using. He pushed hard on speed: hours matter, not days. The goal is to get a cluster online fast for Chris Wolford Arcee/Fuzzball work and the board/Nvidia demos.
People: Peter Nelson, Chris Wolford, Stephen Moody, Greg Kurtzer, Bjorn Hovland
Hold the Google 6.18 image as contract leverage - nothing ships to Google without Peter say-so
Jul 21, 2026 · strategy · high86% confidence
Both 6.18 kernels (default and minimal) were built and validated by 7/17, but Peter instituted a delivery-control gate: no kernels go to Google without his explicit say-so, and delivery-timing questions route to Kelly Hall. The build is done; the only remaining gate is contractual, so CIQ holds the finished image as leverage until GDC signs and resolves the outstanding payment. The ultimate release decision Peter deferred to Bjorn and Kelly.
People: Peter Nelson, Kelly Hall, Bjorn Hovland, Justin Haynes, Nathan Blackham
Support Nathan CVE-first prioritization over the Google minimal-kernel delivery; own the customer comms
Jul 9, 2026 · strategy · medium80% confidence
Nathan flagged that in-flight critical local-privilege-escalation CVEs collided with the committed 6.18 minimal-kernel delivery to Google and called for CVE remediation to come first. Peter backed that call rather than making it: agreed to a bounded slip (no more than a week, not three), was comfortable asking the team to work a weekend to verify already-built patches, and took personal ownership of communicating the slip to Google - framed as value-add (surface only the CVEs Google benefits from), not an apology. Peter wrote and sent the explanatory email to Tissa and the Google team the same day.
People: Peter Nelson, Nathan Blackham, Justin Haynes, Bjorn Hovland, Kelly Hall
Reinforce that an automated Linux CI/CD pipeline is one of the top-3 H2 company objectives
Jul 7, 2026 · strategy · medium72% confidence
Peter had already set an automated Linux CI/CD pipeline as one of CIQ three top company objectives for H2 (base decision not previously captured). In the Max 1:1 and Engineering Weekly he reinforced to people that it is there and that it is critical to Linux delivery - a company objective, not just a Nathan-team objective. Logged as reinforcement.
People: Peter Nelson, Max Spevack, Nathan Blackham, Justin Haynes
Gate Rakuten 8.6 / RT-kernel support work on a signed 1.6 to 1.8M expansion that funds dedicated headcount
Jul 6, 2026 · strategy · medium80% confidence
Decided that the risky, currently unsupportable Rakuten 8.6 / RT-kernel support work will either not be done at all or only be delivered tied to a signed contract expansion in the 1.6 to 1.8M range that funds dedicated Rakuten headcount on CIQ side. The bar is enough to cover roughly two dedicated engineers, not an arbitrary large number. Peter forwarded the engineering teams infrastructure objections to Bjorn to arm the customer conversation.
People: Nathan Blackham, Bjorn Hovland, Jonathan Maple
Everfox desktop-OS: require explicitly scoped and funded eng-investment before supporting the deal
Jun 30, 2026 · strategy · medium66% confidence
On the Everfox desktop-OS deal, Peter framed it as a fundamentally different business (be like Ubuntu while also being like RedHat, not add desktop support to Rocky) and insisted the unknowns around hardware enablement / driver support and a dedicated lab be made explicit. His engineering-side conditions: the MSA must fix the hardware scope with out-of-scope hardware priced separately, and CIQ should proceed only with an explicit commitment to fund the engineering regardless of revenue. He tasked Max to assess technical feasibility with Nathan before committing. Relates to the prior logged Everfox decision d1ea8ed9; this is the engineering-stewardship condition layer.
People: Bjorn Hovland, Nathan Blackham, Ramesh Srinivasan, Max Spevack, Suzanne Spencer-Purcell
Champion Kubernetes-by-CIQ as an H2 engineering deliverable (build on the standard)
Jun 30, 2026 · strategy · medium55% confidence
Peter is actively pushing for a CIQ-branded Kubernetes offering, built on the standard rather than reinvented, to be an H2 engineering deliverable that eliminates a recurring objection in sales conversations. He green-lit Chris Wolford putting Kubernetes by CIQ on his H2 doc. It fits the broader turnkey/easy-button vision (single installable image combining Rocky Hardened plus Fuzzball substrate). Peter acknowledges final product prioritization authority sits with Bjorn/Product; his action here is advocacy and engineering-roadmap steering, not a prioritization decree.
People: Max Spevack, Chris Wolford, Bjorn Hovland
EU Cyber Resilience Act - scope hinges on legal definitions; Bjorn to engage lawyers; aim to publish our own definitions
Jun 18, 2026 · strategy · medium74% confidence
On the CRA thread (raised by Leigh Hennig re a Sept 11 deadline for RESF/Rocky/CIQ), Peter directed that the entire scope hinges on two definitions - vulnerability and actively exploited - and delegated Bjorn to engage lawyers to weigh in. Strategic aim: ideally CIQ/RESF publishes its own definitions, reframing CRA compliance as living up to our stated market promises rather than being exposed to outside interpretation. Until legal responds, nothing to do.
People: Bjorn Hovland, Greg Kurtzer, Leigh Hennig, Brady Dibble
Hold the margin line - willing to walk from bad-economics deals (currently applied to Rakuten)
Jun 18, 2026 · strategy · medium72% confidence
In his 1:1 with Baek, Peter articulated a generalized stance and named its current target: stop playing the tell-the-customer-yes-to-anything game, and hold a hard line even if it means losing the deal. He will not spend 2M to capture 400K. He confirmed this is a generalized principle that at this moment absolutely applies to Rakuten as those negotiations finalize.
People: Chris Baek, Ramesh, Bjorn Hovland