Grant announcements are the easiest news in this sector to ignore. A ministry, a number, thirteen companies most people have never heard of, and a photograph. I nearly skipped it too. Then I noticed that somebody had just published, with money attached, an answer to the question I spend most of my week guessing at: which security problems are going to matter next.
What Was Announced on 10 November
At CyberSG Innovation Day 2026, the Cyber Security Agency of Singapore announced that about S$5.6 million had gone to 13 cybersecurity companies under CyberCall 2025. Minister of State Rahayu Mahzam made the announcement. CSA published it as CSA Strengthens Support for Cybersecurity Innovation, From Research to Market, and AsiaOne and Tech Edition carried the detail.
Four things happened on the same day, and the funding was only the loudest:
- The 13 awards under CyberCall 2025, targeting AI-related threats, quantum risks and critical infrastructure security. One recipient, pQCee, intends to simulate quantum attacks so organisations can find weaknesses in how they protect data and decide what to upgrade first.
- CyberCall 2026 opened, which means the next prioritisation is already being written.
- The CyberSG Atlas directory of local security vendors was relaunched.
- An advisory committee of chief information security officers was established.
CyberCall is co-organised with the National University of Singapore and supported by the CyberSG Talent, Innovation and Growth Collaboration Centre. One caveat on the numbers before I build anything on them: the S$5.6 million figure and the threat categories come from CSA and the coverage of its release, so they are consistent across sources because they share one. That is normal, and it means the money is confirmed while the market effect is my inference.
Our expert take
S$5.6 million across 13 companies is roughly S$430,000 each, which is seed-sized rather than market-moving. Anyone telling you this changes Singapore security salaries is reading the wrong thing. The value is not the quantum of money, it is that a national agency had to choose, and published the choice. Prioritisation made under a budget constraint is the most honest signal an employer can get, and this one cost you nothing to read.
Signal 1: the Funded Threat Classes Are a Hiring Forecast
Research funding reaches shipping products on a cycle of roughly two to four years. So a funding list is a reasonable forward indicator of which capabilities become scarce, and therefore expensive, on that horizon. Three categories were named.
AI-related threats. This is the category I would act on now, and the one my own screens had least coverage of. Note the shape of the problem: it is not prompt injection as a curiosity, it is securing systems where an AI component takes actions in the world rather than producing text for a human to read. Agent-shaped systems have authorisation and blast-radius problems that most security interviews do not touch, because most security interviews were designed for request-response web applications.
Quantum risk. Important, and almost certainly not a hire for you this year. I will come back to why in a moment, because this is where I expect employers to waste money.
Critical infrastructure. Relevant if you operate in it, largely not if you do not, and the sector boundaries here are defined by regulation rather than by self-assessment.
The Hire I Would Refuse to Make This Quarter
Quantum risk is real and the funding is sensible. A post-quantum specialist on your payroll in November 2026 is, for almost every employer in Singapore, the wrong hire in the wrong order.
Look at what the pQCee project is actually for. It simulates attacks so an organisation can find weaknesses in how it protects data and decide what to upgrade first. The operative words are decide what to upgrade first. That is a prioritisation exercise, and it depends on knowing where your long-lived encrypted data sits and which of it would still be sensitive in ten years. That is an inventory a competent senior engineer or security architect can lead with existing staff.
Hire the specialist before the inventory exists and you pay a scarcity premium for somebody who spends their first six months building a spreadsheet they will tell you, correctly, that you should have had already. Do the inventory, get a migration scope, then decide whether the work is a hire or a vendor. The sequence is the whole point.
Our expert take
There is a pattern in how government funding announcements get converted into job adverts, and it is worth naming because it costs real money. A category gets funded, vendors and recruiters adopt the vocabulary within weeks, and employers open roles named after the category rather than after a problem they have. Six months later the role is hard to fill, the brief is vague, and nobody can say what the person would do in their first quarter. The defence is boring and reliable: write the first quarter of work before you write the job title. If you cannot, you have a research interest rather than a vacancy.
Signal 2: a Vendor Directory and a CISO Committee Are a Buy Signal
This is the part of the announcement nobody will write about, and I think it is the more useful half.
CSA relaunched a directory of local security vendors and established an advisory committee of chief information security officers on the same day it handed out the grants. Read those two together. The state is not only funding capability, it is lowering the cost of finding capability you can buy, and creating a forum where the people who make build-against-buy decisions compare notes.
For an employer, that changes the honest answer to “should we hire a security engineer”. Some of what you were about to staff internally is now easier to procure from a local vendor whose existence is catalogued for you. I am aware this is an odd argument to make on a site that helps companies hire developers. It is still the correct one, and the hiring consequence is not nothing: it moves the scarce role from the engineer who implements to the security architect who decides what you own and what you buy. That person needs to be an employee, because the decision is a conflict of interest for anyone selling you the alternative.
| What you might staff | Better answer after 10 November | Why |
|---|---|---|
| Post-quantum specialist | Inventory led by existing senior | Scope does not exist yet; premium paid for spreadsheet work |
| Another detection engineer | Check the vendor directory first | Procurement cost just fell; capability is catalogued |
| AI security generalist | Add an agent-authorisation question to existing loops | Cheap, immediate, tests the actual gap |
| Nothing yet | Security architect who owns build against buy | Only role a vendor cannot hold without a conflict |
Working out whether this is a hire or a purchase
Tell us what your security team covers today and we will tell you which of the three funded threat classes creates work for you, and whether the next move is a role or a vendor.
Let us discuss itWhat This Does Not Mean
Thirteen funded companies and S$5.6 million do not move a labour market, and the honest framing matters more than the enthusiasm.
This does not mean security salaries in Singapore rise this quarter. It does not mean post-quantum skills are suddenly employable at scale, and it does not mean the funded companies will produce usable products on any particular schedule, because most seed-stage security research does not. The retrenchment numbers in information and communications have been elevated this year, and a few million dollars of research funding does not reverse that.
What it reliably changes is the quality of information available to you when you plan headcount. A published prioritisation from a buyer with no product to sell you is worth more than a vendor threat report, and it arrives roughly once a year.
What to Watch Next
- The CyberCall 2026 problem statements. CyberCall 2026 is now open, and its published challenges are the next prioritisation. Read them when they appear; they are a cleaner forecast than the award list because they describe problems rather than companies.
- Whether the CISO committee publishes anything. If it produces guidance, it becomes the closest thing this market has to a shared standard for what an in-house security function should own. That is a free job description.
- Whether agent security shows up in the 2026 categories. If systems where AI takes actions become their own named category rather than part of AI-related threats, treat that as confirmation and staff for it properly.
- Your own interview loop. Add the agent-authorisation question to the next two security interviews you run and compare the signal to what you were getting before. That is a test you can complete this fortnight.
For the role-level mechanics, we cover the hiring process in our guide to hiring security engineers in Singapore after a disclosed vulnerability, the governance side in what the MAS AI risk guidelines changed for hiring, and the interview technique for agent-shaped systems in evaluating AI agent skills in developer interviews.
Frequently Asked Questions
What did CSA announce on 10 November 2026?
Why does a grant announcement matter to an employer who is not applying for a grant?
Should a Singapore employer hire a post-quantum specialist now?
What is the practical hiring change to make this month?
Discussing it beats guessing at it
We help Singapore employers turn a funding announcement into a defensible security headcount plan, then shortlist pre-vetted engineers against the scope rather than against a job title.
Let us discuss it