CEO Column | C6 | Put the Risks Up Front: This Isn’t Pessimism, It’s Board Language
The companies that truly survive are the ones that dare to state their risks clearly.**
Two months ago in Singapore I gave a large fund a briefing on PSF’s cross-border compute deployment. I had assumed what they most wanted to hear was: 256 H200s, the PUE cost advantage, a five-year revenue model, local government cooperation, the energy structure. It wasn’t. After five minutes the fund’s chairman put down his pen and asked a question that reset the whole meeting: “Tell me the failure modes first.” (First tell me how this project could die.) In that second I understood immediately: in the capital market, real professionalism is not drawing the future but dismantling the risks. Because investors are not afraid you aren’t strong enough; they are afraid you don’t know where you will fall.
01 | The Investor Mindset of the AI Era: Ask How You Die First, Then How You Live
The world AI brings is not linear but a world of “amplification effects.” A risk that isn’t stated clearly:
* will be amplified by Agents
* will spread across models
* will cascade through processes
* will escalate into incidents
* will become a systemic problem
This is why when you talk about vision, investors frown; but when you talk about risk, they smile.
Because people who can state risks clearly are usually people who have done their homework.
02 | Putting Risk First Isn’t Pessimism; It’s “Turning Uncertainty into Structure”
You can describe a project in two ways:
Way A (amateur):
“Our model is strong, our market is big, we will surely succeed.”
Way B (professional):
“This project has six key risks, and our corresponding mitigation mechanisms are as follows…”
Which one will investors buy?
The answer is always **Way B**, because Way B means:
* you know what you are doing
* you have a checklist
* you have contingency plans
* you are predictable
* you don’t rely on luck
This is Board Language.
03 | What Is “Risk-First Narrative” in the AI Era? (The Most Important Part)
The following is the core template for all your future pitches / MOUs / partnership proposals.
1. Model Risk
* Version updates may cause unpredictable behavior
* Data-source contamination
* Prompt injection
* Model drift
* Training bias
The point is not the risk itself, but what your rollback strategy is.
2. Workflow Risk
* Agents misusing data
* Permission mismatches
* Inconsistent SOPs
* Automation spinning out of control
* Monitoring blind spots
Investors only want to know whether you can “circuit-break,” “downgrade” and “have a human take over.”
3. Security Risk
* Data leakage
* Supply-chain attacks
* User abuse of privileges
* API vulnerabilities
Your answer must be: zoning, least privilege, traceability.
4. Supply Risk
Especially important in the compute era:
* GPU supply cycles
* Energy stability
* Cooling-system redundancy
* Construction delays
* Government variables
PSF is always direct here: “Our backup plan is as follows.”
5. Financial Risk
Not “we will earn a lot,” but:
* What if deployment is slow?
* What if customers pay late?
* What if GPU prices fall?
* What if utilization is insufficient?
Each has a back-calculation mechanism.
6. Regulatory Risk
For cross-border projects, what matters most is:
* Can you comply?
* Can you keep complying?
* Can you adjust quickly when regulations change?
AI Agents will automatically cross-check regulations, and you must answer: “Our automated compliance check (Compliance Agent) is already integrated into the process.”
04 | True Professionalism Is Putting the “Risk Map” on the Table Before the Investor Even Speaks
This is the highest-level logic I learned in the capital market. When you voluntarily state risks:
* Investors will think you are mature
* The board will think you are steady
* Executives will think you are reliable
* Partners will think you can last a long time
And when you don’t talk about risk:
* No one will believe your growth
* No one will dare to bet on your future
* No one will want to work with you
* No one will hand the stakes to you
Because people who hide risks are people without structure.
05 | PSF’s Way: Every Proposal Opens with “Risk First”
This is our fixed format, and you can copy it directly:
1. Risk Landscape
List the top 8 risks.
2. Each risk is matched with “evidence”
Not concepts, but documents, SLAs, flowcharts and data.
3. Every risk has a “mitigation strategy”
Circuit-breaking, downgrade, zoning, rollback, human takeover.
4. Every mitigation strategy has a “chain of responsibility”
Who signs, who is responsible, who audits, who can override.
5. Every chain of responsibility has “traceability”
Logs, versions, processes.
The result? Investors usually say just one key sentence: “You should last a long time.” In the capital market,
that sentence is worth more than “you will grow fast,” because only those who can last deserve a high multiple.
06 | Conclusion: Risk Is Not Pessimism; Risk Is Strength
The cruelest and also fairest line of the AI era: vision inflates; risk does not. Vision relies on narrative; risk relies on structure. That is why future pitches, partnerships and fundraising will all be classified by machines in one sentence:
> **“Risk structure: complete → investable.”**
> **“Risk structure: unclear → rejected.”**
If you cannot master risk, you cannot master valuation.