On October 20, 2026, McKinsey published a quantum road map for business leaders. It describes three horizons. Early commercialization is the one we’re in. A competitive position comes once fault-tolerant machines arrive, which the firm says could be within five years. A quantum-enabled economy follows in the late 2030s. For each horizon it tells leaders what to do: “secure ecosystem partnerships and access to frontier systems; pilot applications ahead of commercial readiness; build organizational familiarity through talent and IP; and explore how quantum technology could ultimately alleviate compute and energy bottlenecks in AI scaling.” Near the end sits the line someone will read aloud in a board meeting: “Quantum technology will not wait for organizations to feel ready.”
Before anyone reads it aloud, check who it was written for. The road map assumes every reader runs a company where quantum could pay. By the firm’s count, “over 300 organizations worldwide are already working with quantum technology providers.” That’s who the advice fits. If you run one of the several million other companies, you’re reading someone else’s memo.
I spent years at McKinsey, and I read the firm’s work the way I’d want mine read: with a red pen. An underline means the claim holds and you can repeat it to your board as written. A footnote means I’d send it back for more data, because the claim is true in aggregate and says nothing about what to do at one company. I did that with this piece, next to two other documents from the same month. One is the Government Accountability Office’s October 6 audit of how the 24 largest federal agencies are preparing for quantum-safe encryption. The other is Gartner’s survey of CIOs for 2027, which projects IT budgets growing an average of 3.7 percent next year while funding for agentic AI grows 31.8 percent.
Put the three documents side by side and the picture is simple. McKinsey has three horizons. You have one budget. For most companies, one of the three horizons is a security program with federal deadlines attached, the other two get a watching brief, and the security program is the line that goes in the 2027 plan.
The Short Version
- The road map tells you what playing costs. One-third of the quantum buyers McKinsey interviewed spent $10 million or more in 2025. A 3.7 percent raise on a $50 million technology budget is $1.85 million. Most companies can’t get there, and most don’t need to.
- The sentence to keep is about security. Companies that never touch a quantum computer still have to replace their encryption, and the firm calls that “a multiyear program, not a patch cycle.”
- The agencies under orders to do this haven’t started. GAO found that one of 24 federal agencies has finished its inventory of vulnerable encryption, none has identified the funding, and one has done any testing.
- Your 2027 line is three things. An inventory of where your encryption lives, a clause in every new contract, and an estimate for the full migration. Run it like Y2K. Budget it like AI.
Four claims hold up, and the price is the one to keep
The road map is better than most of what gets handed to a board on this subject. Four things in it get an underline.
The price comes first. “When we interviewed business leaders already working with quantum technology providers, one-third said their companies spent $10 million or more on quantum computing in 2025, with 7 percent of those companies spending $50 million or more.” Big-firm pieces rarely tell you what taking their advice costs. This one does, in passing, and it’s the most useful number in the document.
Put it next to your budget. Take a company with a $50 million technology budget, a reasonable size for a mid-market company or one division of a large one. A 3.7 percent increase is $1.85 million. The entry ticket McKinsey describes is more than five years of that entire increase, before a dollar goes to the AI lines Gartner says are growing more than eight times as fast as the budget (Exhibit 1). The road map was written for companies where that math works.
The security paragraph gets the second underline. “Even companies that don’t directly deploy quantum computing will need to update their security platforms,” the authors write, because sensitive data “is already exposed to ‘harvest now, decrypt later’ attacks.” Then: “This is a multiyear program, not a patch cycle.” I made the same argument on October 12. The firm gives it one paragraph inside a growth story. For most companies it’s all the quantum money there is through 2031, and I’ll come back to it.
The third underline goes to the list of where the value lands. The road map names four arenas where it expects outsize value: cybersecurity, obesity-related drugs, industrial and consumer biotechnology, and robotics. Read that list as a filter. If your company doesn’t make molecules, materials, or decisions at enormous scale, the growth story belongs to someone else.
The last one is candor about what could slow all of it down. The authors say scaling is largely a manufacturing problem, that talent is scarce, and that no hardware design has won. I’d read those three sentences to any board member who came back from a conference excited.
What to do: measure the $10 million entry price against your budget increase, never against the forecast. If the ticket is more than a year of your whole raise, you’re reading about someone else’s budget.
Four claims need a footnote before they go in a board deck
My standing objection to big-firm thought pieces, including ones I admired from the inside, is that the data gets aggregated until it stops telling any one reader what to do. A global value pool doesn’t pay for anything at your company. A survey of people who bought can’t tell you whether to buy. The road map has four numbers like that.
The value number. The firm projects “$1.3 trillion to $2.7 trillion of economic value worldwide by 2035” from quantum computing, and in the same document expects the quantum computing market “to capture between $43 billion and $71 billion” in 2035. Divide one by the other and the value created is somewhere between 18 and 63 times what the vendors collect. That’s my math on their two ranges, and the shape is reasonable. It means almost all of the value, if it arrives, lands with the companies using the machines, and the document doesn’t say which companies or how. A board funds projects with names. A share of a $2.7 trillion pool doesn’t have one.
The entry price. The one-third-spent-$10-million figure comes from interviews with leaders “already working with quantum technology providers.” As far as I can find, the public version doesn’t say how many people were interviewed. So the figure describes what participants spend. It doesn’t tell you what participation requires, and those two get confused in budget meetings.
The five-year machine. The road map says hardware road maps “are converging on the possibility of delivering a fault-tolerant quantum computer within five years,” and cites IBM’s published path to large-scale fault tolerance by 2029 and IonQ’s target of two million physical qubits by 2030. Those are vendor road maps. The last independent view I checked, the Global Risk Institute’s survey of 26 experts, put the ten-year odds of a machine that breaks today’s encryption at 28 to 49 percent. Both can be true at once. Vendors publish the optimistic path because that’s what a vendor road map is for.
The 300 organizations. “Working with quantum technology providers” could mean a paid co-development program or a cloud account somebody opened to try an algorithm. The document doesn’t draw the line, and the line decides whether 300 is a movement or a mailing list.
| What the road map says | My mark | Why |
|---|---|---|
| One-third of the buyers interviewed spent $10 million or more in 2025 | Underline | The only cost figure in the document. Measure it against your budget |
| The security migration is “a multiyear program, not a patch cycle” | Underline | True for every company that holds data, and it has federal dates attached |
| Outsize value lands in four arenas: cybersecurity, obesity drugs, biotech, robotics | Underline | Tells you in one line whether the growth story is yours |
| Scaling is a manufacturing problem, talent is scarce, no hardware design has won | Underline | The authors’ reasons it could run late |
| $1.3 trillion to $2.7 trillion of economic value by 2035 | Footnote | A global pool. It doesn’t say which companies get it or how |
| A fault-tolerant machine within five years | Footnote | Vendor road maps. Independent experts put the ten-year odds at 28 to 49 percent |
| Over 300 organizations working with providers | Footnote | ”Working with” is undefined. It could be a contract or a cloud login |
| AI will compress the quantum timeline | Footnote | Possible. If it happens, a faster machine opens your data first, so the security line comes first either way |
Source: McKinsey, “Three horizons of quantum,” October 20, 2026. The marks and the reasons are mine.
I understand why the document reads the way it does. A piece that carries a large firm’s name is reviewed by people whose job is to keep the firm out of trouble, and the safest sentence is the broad one. I’m a one-person firm. Nobody reviews this but me. So I can tell you which rows don’t apply to you and put a number next to each one, and that’s the part a large firm’s review process tends to strip out.
What to do: ask for the number that applies to one company. If the document doesn’t carry it, keep the claim out of your board deck.
One question tells you which horizon you’re on
Does your company make money by discovering molecules, designing materials, or optimizing something enormous, like a global supply chain, a trading book or a chemical process? If yes, read the road map as written and argue about the price. If no, you’re on the horizon the road map mentions in passing, and the four priorities look different from your side of the table (Exhibit 3).
| McKinsey’s priority | What it buys | What it costs | Who it’s for |
|---|---|---|---|
| Partnerships and access to frontier systems | Time on scarce machines, and a place in line when fault-tolerant capacity is rationed | Multi-year contracts with a hardware provider. The interviewees in this tier spent $10 million or more in 2025 | The roughly 300 organizations the firm says are working with providers |
| Pilots ahead of commercial readiness | Proof that quantum can touch a problem you have | A team, machine time and a problem. By the firm’s wording, the pilot runs before the technology can pay back | Companies whose product depends on molecules, materials or optimization at enormous scale |
| Talent and intellectual property | People who can write quantum algorithms, and patents | Salaries in a market the firm calls scarce. A line that recurs every year | The same few industries |
| Quantum for AI compute | An option on cheaper AI capacity in the 2030s | Little in 2027. A research watch, with nothing to show until the third horizon | Companies that run their own AI infrastructure |
| The security migration (one paragraph in the road map) | Keeping your data readable only by you after the machine arrives | An inventory, a procurement clause and a whole-job estimate in 2027. The expensive part, replacing old systems, comes later | Every company that holds data someone would want in 2035 |
Source: McKinsey, “Three horizons of quantum,” October 20, 2026. The cost column is my reading of what each priority requires, with the one number the document gives.
The fifth row is the one I’d fund. The firm and I agree it’s required. We disagree about how much of the budget it is.
The security horizon has dates and the growth story does not. Executive Order 14412 gives federal agencies until December 31, 2030 to move their most important systems to quantum-safe key establishment and until December 31, 2031 for digital signatures, and it tells the Federal Acquisition Regulatory Council to propose a rule carrying the 2030 date into contractor requirements. I laid those dates out on October 12. What arrived since is a report on how the migration is going for the people who have the deadline.
The organizations that are required to do this haven’t started
GAO’s audit, dated October 6, 2026, looked at three preparatory steps across the 24 largest federal agencies: inventory the systems that use vulnerable cryptography, identify the funding needed, and test the replacement algorithms.
In the report’s words: “Only one of the 24 agencies fully addressed all applicable activities for developing prioritized inventories of systems with vulnerable cryptography.” “None of the 24 agencies fully addressed the one activity under the preparatory practice to identify the funding needed.” On testing, one agency out of 24 had done any work at all, and that work was market research (Exhibit 4).
These are organizations with an executive order, two dated deadlines and a memo. OMB told them on June 24 to submit a migration plan within 120 days that includes “an estimate of funding and personnel resources required.” By my count that plan was due October 22, 2026. The Office of the National Cyber Director told GAO that the immature market for quantum-safe products was “the biggest factor” delaying funding estimates, and 16 agencies said testing was premature because vendors hadn’t built the new algorithms into their products yet.
A board at a private company can take three things from that audit.
- The inventory is the bottleneck, and it’s the cheap step. If the agencies with the mandate couldn’t finish the list, the list is harder than it looks, usually because nobody knows where the old systems are. Start there.
- “The vendors aren’t ready” is a sentence you’ll hear too. The answer is a procurement clause, so that everything you buy after the clause is signed arrives with the new standards or a documented upgrade path. Waiting for the vendor market to mature is how an agency ends up with no funding estimate two years after the standards were published.
- A plan without a number isn’t a plan. OMB asked for a resource estimate because the estimate is what gets a plan funded. If the organizations with the deadline can’t produce one, producing one is the work, and it belongs in the 2027 line.
What to do: fund the list and the estimate first, and write the procurement clause before the first vendor meeting.
Run it like Y2K, budget it like AI
That’s the line I’ve been giving the boards that asked me about quantum in October, and it’s the shortest version of the sequence I’d recommend.
Run it like Y2K. The Commerce Department estimated that from 1995 to 2001 American agencies and companies spent about $100 billion fixing a two-digit date field, about $365 per American. Y2K had a date, so it got an owner, an inventory, a finish line and a budget, and then it got boring and got done. The quantum migration has dates too. They’re federal dates in an order and a memo, and they’ll reach your contracts and your largest customers’ audits. Set your own finish date, with 2030 as the default because that’s the one your federal customers inherit, give the work an owner inside the security oversight you have, and report on it like a project.
Budget it like AI. The lesson of 2025 and 2026 is that the money shows up after the first step. AI started as a pilot line and became the fastest-growing line in the budget, and vendor pricing changed underneath it. The quantum line will behave the same way inside a budget growing 3.7 percent. The costs that hurt, replacing the old systems nobody wanted to touch, come after the cheap inventory. So treat it the way a careful CFO treats AI in 2026: ask for the full estimate as a range before the first step is approved, revisit it every budget cycle, and expect it to grow.
Boring is hard to invoice. It still needs a line in the 2027 plan.
The Board’s Reply to the Road Map
- Which horizon are we on? If our product doesn’t depend on molecules, materials or optimization at enormous scale, we’re on the security horizon, and the growth story gets a watching brief.
- Where is our encryption, and how long does each kind of data have to stay secret? If nobody can answer, that’s the first line item. GAO found that one of 24 agencies with a mandate had finished answering it.
- What’s the full estimate for the migration through 2031, as a range? Shown before the first step is approved, and refreshed every budget cycle.
- What’s our one trigger for funding a quantum pilot? A vendor shows, with proof, that a quantum machine beat the best conventional method on a problem we have. Until then, no pilot.
Where the road map could be right, and why the security line still comes first
The road map makes its case well. The window to build a leading position “will narrow with each horizon that passes.” Equity investment in quantum companies rose from $2.5 billion in 2024 to $13.3 billion in 2025. Quantum companies have been going public in 2026, and McKinsey names five. And the authors argue that AI and quantum reinforce each other: AI is speeding up algorithm design and error correction, which could bring fault tolerance forward.
I take that seriously. I’ve said myself that once momentum builds, these technologies show up faster than the road maps say. AI taught everyone that lesson in about eighteen months.
It doesn’t change the sequence. If the timeline compresses, the first thing a faster quantum computer does to most companies is open their data, so acceleration is one more reason to fund the security line first. And a trigger costs nothing to write down. The companies that spent $10 million in 2025 bought an option on early access. A written trigger buys most companies the same option for the price of a memo, and it fires the day a vendor proves an advantage on a problem you have.
Here’s what would move me. A sourced benchmark for what the migration costs a single company, which I’ve looked for and haven’t found, could move the first line up or down. A demonstrated quantum advantage on an ordinary business problem, a routing problem or a portfolio problem at a company that doesn’t make molecules, would move the pilot from a trigger to a line.
Funding a line with no return and no fixed date is a hard budget conversation, and AI never had to have it. The boards that have it in the 2027 cycle will do it with an inventory, a clause and an estimate, for a fraction of what the road map describes. The data your customers gave you is being copied by people who expect to read it later. Decide which horizon you’re on, and fund the one line that has a date on it.
How RLK Can Help
I brief boards and leadership teams on quantum in plain English, and I help CIOs and CFOs put a defensible first number on the migration and sequence it against the AI budget. Start a conversation.
Sources
- McKinsey, “Three horizons of quantum: A business road map to the future computing economy,” October 20, 2026
- Gartner, “CIO Planning for 2027,” August 31, 2026
- GAO-27-108740, “Quantum Computing: Federal Actions Needed to Prepare for Emerging Cyber Threat,” October 6, 2026 (revised October 14)
- Cybersecurity Dive, “US government lagging in transition to post-quantum encryption, GAO finds,” October 8, 2026
- OMB Memorandum M-26-15, “Execution of the Migration to Post-Quantum Cryptography,” June 24, 2026
- The White House, Executive Order 14412, “Securing the Nation Against Advanced Cryptographic Attacks,” June 22, 2026
- Global Risk Institute, “Quantum Threat Timeline Report 2025,” March 9, 2026
- Nextgov, “Biggest Y2K costs behind us,” Commerce Department estimate, November 16, 1999
- RLK Consulting, “Quantum Is a Bill Before It’s a Business Case,” October 12, 2026