AI Infrastructure Is Becoming a Business Issue

You may not be planning to build a data centre, train a large language model, or buy advanced computing equipment. Yet major changes in AI infrastructure can still affect how you run your Malaysian business.

When companies such as Nvidia, SoftBank and OpenAI commit billions to computing facilities, they are signalling that AI will increasingly sit behind everyday business tools. The impact may reach your customer service software, accounting platform, marketing tools, inventory systems and internal workflows.

That creates a practical question for you: how do you prepare for more AI without wasting time, exposing business data or forcing your team to manage complicated technology?

TL;DR

Nvidia plans to invest US$1.5 billion in SB Energy, a data-centre developer connected to SoftBank and OpenAI. The project could scale from 4.25 gigawatts to 8 gigawatts, with Nvidia supplying the computing infrastructure, according to TechCrunch.

For your SME, the lesson is not to copy large technology companies. It is to prepare your processes, data and staff so you can adopt reliable AI-enabled automation when it becomes useful.

What This Means

A data centre is a facility filled with servers that store information and run software. AI applications need substantial computing capacity because they process large amounts of data and perform complex calculations.

In the reported project, Nvidia will be the sole supplier of compute infrastructure for OpenAI’s Ports-Pike data centre near Cincinnati, Ohio. The facility is expected to begin at 4.25 gigawatts and potentially expand to 8 gigawatts, based on company documents reported by TechCrunch. Nvidia also plans to provide up to US$105 billion in credit support for construction, according to the same report.

The wider project includes a planned 9.2-gigawatt natural gas power plant. The plant is expected to cost US$33 billion, while BloombergNEF reported that natural gas power plant construction costs had risen 66% over the previous two years, as cited by TechCrunch.

These figures matter because AI is not only a software trend. It depends on physical infrastructure: servers, chips, electricity, cooling systems, networks and secure facilities. As this infrastructure expands, AI providers may offer faster, more capable and more specialised services to businesses.

The useful question for an SME is not “Should we use AI everywhere?” It is “Which repetitive process should become easier, faster or more consistent?”

How This Applies to Malaysian SMEs

1. Your existing business software may become more intelligent. Malaysian SMEs commonly use cloud accounting, customer relationship management, e-commerce, payroll, messaging and inventory systems. As providers gain access to more computing capacity, they may add features such as automatic invoice extraction, sales forecasting, customer response suggestions and stock alerts.

You do not need to build these capabilities yourself. Instead, review the systems you already use. Ask your software provider whether AI features are available, what data they use and whether a human can approve important outputs. A wholesaler in Shah Alam, for example, may use automated reminders for overdue documents while keeping final approval with the finance manager.

2. Your business data will become more valuable—and more sensitive. AI works best when information is organised. Customer names, product codes, supplier records, service histories and standard operating procedures can help automation tools produce useful results. But poorly controlled data can create incorrect answers, privacy concerns or operational mistakes.

Before adopting an AI tool, identify where your important information is stored. Check whether customer records are duplicated, whether staff use personal messaging accounts for business instructions, and whether old files contain unnecessary personal information. A Malaysian recruitment agency, for instance, should define who may upload candidate information into an AI application and how long that information is retained.

3. Small teams can benefit from workflow automation, not just chatbots. Many owners first think about an AI chatbot. That can be useful, but a more practical starting point may be connecting routine steps across your business. When a customer submits an enquiry, the system could record it, assign it to a salesperson, create a follow-up task and send a confirmation message.

For a local service company, such as an air-conditioning contractor, automation could turn a WhatsApp enquiry into a job record, remind the team about site visits and request feedback after completion. The value comes from reducing missed follow-ups and repeated data entry, rather than from producing impressive-looking AI content.

4. Dependence on outside platforms requires a backup plan. Large infrastructure investments may improve AI services, but your business can still be affected by service interruptions, policy changes or a provider removing a feature. You should know what happens if an AI function becomes unavailable.

Keep essential customer, order and financial information in systems you can access and export. Document manual alternatives for critical tasks. If an automated quotation process fails, your team should still know how to prepare and approve a quotation without guessing.

Practical Takeaways for Your Business

• Choose one repetitive process first: Start with enquiries, document filing, appointment reminders, invoice processing or internal reporting.
• Measure the current process: Record how many enquiries, documents or follow-ups your team handles each week before making changes.
• Set human approval points: Require staff to check quotations, refunds, employment decisions and customer-facing claims.
• Clean your data: Standardise customer names, product codes, supplier details and document formats.
• Check data handling: Read the provider’s privacy, retention and security terms before connecting business records.
• Keep an audit trail: Make sure you can see who changed a record and whether an automated suggestion was accepted.
• Train staff on exceptions: Your team should know what to do when an automation produces an incomplete or incorrect result.
• Review monthly: Check whether the workflow reduces delays, errors or duplicated work.

A Simple Readiness Scorecard

Area	Ready when you can answer	First action
Process	Which task repeats most often?	Map the steps from request to completion.
Data	Where is the required information stored?	Remove duplicates and standardise key fields.
People	Who approves the result?	Assign an owner and define escalation steps.
Security	Who can access the information?	Review user permissions and provider controls.
Continuity	What happens if the tool is unavailable?	Write a simple manual fallback procedure.

Use this scorecard for one workflow before expanding to other departments. A focused pilot is easier to monitor than a broad project covering every part of your company.

What to Ask Technology Providers

When a vendor presents an AI feature, avoid judging it only by a demonstration. Ask practical questions that relate to your daily operations:

1. What business information does the feature access?
2. Is our data used to train a wider model?
3. Can we restrict access by staff role?
4. Can a manager review and approve the output?
5. Can we export our records if we stop using the service?
6. What happens when the system gives an incorrect answer?
7. Which activity logs and reports are available?

These questions help you separate a useful business function from a feature that simply adds another screen for your employees to manage.

The Bigger Picture

The Nvidia-SB Energy project shows that AI growth depends on major investments in computing and power infrastructure. The reported facility could expand to 8 gigawatts, and the associated power development is planned at 9.2 gigawatts, according to TechCrunch. This scale is far beyond the needs of an individual SME, but it helps explain why AI services are becoming a permanent part of business software.

For Malaysian business owners, the long-term advantage will not come from chasing every new AI announcement. It will come from building a business that is organised enough to use dependable tools safely. Clear processes, accurate records, defined responsibilities and sensible approvals will matter just as much as the software itself.

Start with one workflow that wastes staff time every week. Document it, improve the data, test a controlled automation and measure the result. Once your team trusts that process, you can decide whether another part of the business deserves the same treatment.

Ready to Streamline Your Operations?

Your business should run itself. AutoRunBiz deploys AI agents to automate your daily operations — WhatsApp orders, invoicing, customer follow-ups, and accounting. Book a free 15-min ops audit to see where automation fits your business →