GIC is looking beyond semiconductor companies and AI developers when it evaluates AI-related investments. The more revealing question, according to its CEO, is whether a company can use AI to increase productivity and efficiency in its own operations.
That is a useful distinction for operators. The story is not that every business now needs an AI label. It is that the evidence of value is shifting closer to the workflow: where work gets faster, decisions improve, or a hard-to-copy advantage becomes stronger.
From AI exposure to operational proof
GIC has invested in AI-related businesses since around 2020, including semiconductor makers, chip supply-chain companies, Anthropic, and Vantage Data Centers. Its stated focus is now broader: large listed companies experimenting with AI and assessing how it could change their businesses and operations.
The fund's CEO pointed to proprietary data as one possible structural advantage. In plain language, a company may have information, processes, or customer relationships that competitors cannot easily reproduce. If AI helps turn that advantage into a better workflow, it may create more value than a generic chatbot layered over the same work.
That does not make proprietary data a shortcut to success. It is an investment lens, not a guarantee. The practical question remains whether the company can show a measurable improvement without creating new quality, privacy, or control problems.
What operators can take from this
Start with one operational bottleneck, not a broad AI programme. Choose a task where the current cost is visible: time spent preparing a report, handling repeated customer requests, reviewing documents, or moving information between systems.
Then establish a baseline before changing anything. How long does the task take? How often does it need rework? Who checks the output? A useful AI pilot should make at least one of those measures better while keeping the person accountable for the decision.
Next, identify what makes the workflow hard to copy. It may be internal knowledge, a well-maintained customer record, a specialised process, or a team with clear review standards. This is the part of the work that can turn a generic model into a real operating advantage.
Why the hedge-fund news still matters
The report also says GIC plans to deploy an additional USD 30 billion into hedge funds over the next three years, after tripling that allocation over the previous decade. That is a diversification decision, separate from the AI investment lens.
Together, the two points show a disciplined approach: seek upside where there is evidence of durable value, while spreading exposure when the wider market is uncertain. GIC reported a 3.4% average annual return above inflation over the past 20 years, down from 3.8% a year earlier.
For an operator, the lesson is more modest. Treat AI as a business change that needs proof, not as a category that automatically deserves budget.
A practical test for the next 30 days
Pick one repeatable workflow and write down its current time, error rate, and owner. Test an AI-assisted version with a defined review step. Keep the change only if it produces a clear improvement that the team can explain and maintain.
That is closer to the evidence GIC says it is looking for: not excitement around AI in the abstract, but a credible path from technology to a better business operation.