There’s way too much content out there on the Internet today.

Correction: Too much unhelpful, "me-too" content that wastes everyone's time. 

Everyone is publishing similar content about the same things, pushing their own agendas, and trying to sell someone something.

It’s tiresome, boring, and readers have had enough of it.

At With Content, we want to change the way content is perceived.

We don’t believe in churning out bits of content that have obviously been thrown together as an afterthought or copy-pasted-rephrased.

We believe in meticulously crafting credible, authoritative longform content that your potential customers will actually love.

Over the years, we've partnered with industry leaders to do just that, including household names like Singtel, DBS, Grab, Randstad, and TÜV SÜD.

Let us help you become the go-to authority in your space with content: https://withcontent.co/ or daniel@withcontent.co

Daniel Tay

New year, new opportunity to join the best company on the planet (I'm biased of course):

Raising the bar for content marketing in Asia | Co-founder at With Content

Storytelling is the lowest hanging fruit.

B2B marketing is getting harder than ever. After the shift from earned media to owned media, brand content is undergoing yet another paradigm shift in the era of generative AI. Simply generating content with an LLM isn’t cutting it, and brands are at a loss.

How do you avoid having your content sound like everyone else? Crucially, how do you attract the right audience and earn their trust enough to do business?

The way forward isn't more content – it's better storytelling. Stories that meet your audience where they are, say something worth hearing, and sound unmistakably like you.

I work with brands to uncover those stories and tell them well.

---

What I bring to the table:

✅ Technical fluency from a prior career in IT.
✅ 18 years of editorial judgement as an independent writer and editor.
✅ Perspective shaped across editorial, agency, and brand roles.
✅ A track record in audience growth, brand building, and narrative shaping.

---

Here’s how I can work with you: 

Work with me directly
👉 Consulting on technology trends.
👉 Content strategy and ghostwriting.
👉 Speaking and moderation.

Reach my audience
👉 Sponsored posts.
👉 Brand partnerships.

---

If you would like to read more of my content:

🔸Follow me on LinkedIn.
🔸Tech Stories Blog: https://www.techstoryteller.co
🔸Weekly newsletter: https://www.techstories.co/updates
🔸Clearly Tech Substack: https://www.clearlytech.co
🔸WhatsApp Channel: https://whatsapp.com/channel/0029VbB6pYr9hXF32Ycd4t2f

---

You can contact me via LinkedIn direct message or email at: paul@techstoryteller.co

Paul Mah

Received an email earlier this week from a reader in Australia that made my day: "I read [your] newsletter every week."

It sounds like a small thing, but to have a space in someone's cluttered inbox is not a compliment I take lightly.

It has been almost 3 years since I started writing online under my own platforms in 2023, beginning with daily LinkedIn posts, followed by weekly newsletters, and finally Substack.

You probably won't believe it, but I struggled with writing in the conversational style of LinkedIn and my newsletter commentaries. It was tough, and I sometimes felt they weren't good enough.

Anyway, I persevered with only one objective: I don't want to say the same things as already mentioned in a hundred other publications or to recycle tired talking points.

What I write must be sharp and offer unique value. Or I don't write it.

In case you wondered:

🔸My posts are all typed out by hand. Every sentence is crafted for an economy of expression. No nonsense about training an AI model to sound like me or on what I know.

🔸I don't hate AI; I do subscribe to multiple GenAI services and use them heavily. But I'm clear about not producing AI slop, which pushes the cognitive load to readers.

I want to say I am immensely thankful for the connections, conversations, and also opportunities that came my way.

As always, thank you for reading and engaging.

𝗣𝗦: Link to sign up for my newsletter in comments.

---
My name is Paul Mah and I offer sharp takes on AI, cybersecurity, data centres, enterprise IT, and sustainability. Follow me here for regular updates.

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Paul Mah. My first 'pointless selfie' of 2026 taken at Vegas earlier this month - Ian Yong Hoe Tan.

Helping growing tech companies with their content and strategy

Digital Edge has announced a US$4.5 billion investment to build a 500MW data centre campus in Indonesia, its largest infrastructure commitment to date.

✅ Join my WhatsApp Channel (one-click): https://lnkd.in/gfZUWPSe

𝟭/ 𝟱𝟬𝟬𝗠𝗪 𝗰𝗮𝗺𝗽𝘂𝘀

Digital Edge's CGK data centre campus will be built in the GIIC Industrial Estate in Bekasi, located within the Greater Jakarta area. The 500MW campus has "scalability up to 1GW."

The data centre campus will be engineered for AI workloads and designed for operational efficiency and sustainability, says Digital Edge in its release.

Some key details:

- Target annualised PUE of 1.25.
- Feature direct-to-chip liquid cooling.
- Renewable energy integration.
- Recycled water systems.

Phase 1 will consist of 3 buildings, with RFS dates of Q4 2026, Q1 2027, and Q2 2027. The capacity of phase 1 wasn't mentioned.

The CGK Campus is 40 kilometres from Digital Edge's downtown EDGE1 and EDGE2 data centres, which offer 6MW and 23MW IT load, respectively.

𝟮/ 𝗜𝗻𝗱𝘂𝘀𝘁𝗿𝗶𝗮𝗹 𝗲𝘀𝘁𝗮𝘁𝗲

Like many data centres in Indonesia, the CGK Campus is located in an industrial estate. In this case, the GIIC Industrial Estate spans over a massive 2,100ha (1 hectare is ~1.5 soccer fields).

GIIC provides infrastructure, including electricity supply (500kV substation), water and wastewater treatment facilities, and fiber optic connectivity.

It is not clear if the CGK Campus will have its own water treatment plants or is relying on GIIC's facilities. But clearly, a lot of the heavy lifting is already done by the industrial estate.

It does seem like GIIC was chosen because that's where electricity is available - and appears to be well suited for AI data centre deployments.

𝟯/ 𝗥𝗮𝗽𝗶𝗱 𝗴𝗿𝗼𝘄𝘁𝗵

Established in 2020, Digital Edge is headquartered in Singapore and has seen remarkable growth since its inception. It is backed by Stonepeak Infrastructure Partners.

Today, the company operates 24 data centres across seven countries: China, Japan, Korea, India, Malaysia, Indonesia, and the Philippines. These are a mixture of acquired assets and greenfield developments.
​
📆 𝗚𝗲𝘁 𝗺𝘆 𝗰𝗼𝗻𝘁𝗲𝗻𝘁

✅ Get my weekly newsletter: www.techstories.co/updates
✅ Join my WhatsApp Channel: www.techstories.co/wa

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Digital Edge. Digital Edge CGK Campus.

In anticipation of the growing role of HVDC in AI workloads and data centres of the future, STT GDC today launched a testbed to support its adoption.

✅ Join my WhatsApp Channel (one-click): https://lnkd.in/gfZUWPSe

𝟭/ 𝗙𝘂𝘁𝘂𝗿𝗲𝗚𝗿𝗶𝗱 𝗔𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗼𝗿 𝘁𝗲𝘀𝘁𝗯𝗲𝗱

Located in Jurong Island, the FutureGrid Accelerator is a testbed to test and refine HVDC (High Voltage Direct Current) electrical systems using real AI servers.

By integrating HVDC with actual AI workloads, the goal is to gain competency in managing DC power systems to eventually develop new standards for data centres.

At the media event held in Jurong Island today, STT GDC's Mingcheng Lim noted that AI infrastructure is pressing against the limits of traditional AC electrical systems and how HVDC offers the answer.

STT GDC also signed MOUs with the ITE, SP, NTU, and NUS to provide internship placements, as well as various training on AI, sustainable energy systems, and HVDC.

𝟮/ 𝗛𝗩𝗗𝗖 𝗯𝗲𝗻𝗲𝗳𝗶𝘁𝘀

Though AC electrical systems have served us well, they are reaching their limits for ultra-dense AI infrastructure, explained Mingcheng.

Namely, multiple power conversion stages introduce inefficiencies, with enormous infrastructure complexity in the form of electrical switching, transformers, and redundant electrical pathways. 

On its part, HVDC can:

- Simplify electrical distribution system.
- Eliminates multiple conversion stages.

The outcomes are nothing to scoff at: Around 30% improvement in overall energy efficiency, and major savings in space and materials with less copper and fewer components required.

𝟯/ 𝗧𝗵𝗲 𝗗𝗖 𝗱𝗮𝘁𝗮 𝗰𝗲𝗻𝘁𝗿𝗲

Nvidia's upcoming Rubin Ultra GPU has cast HVDC into the limelight. Beyond cutting-edge AI systems, is there room for it in the data centres of the future?

The main challenge for HVDC adoption could be the lack of standards and awareness, Teong Chuan Y., who heads R&D at STT GDC, told me.

Another could be a lack of skillset.

- HVDC skillset.
- Operational know-how.

The second is easy to overlook but is a critical aspect that the testbed hopes to bring in the practical knowledge to be shared.

Will data centres of the future turn to HVDC? Hybrid deployments are likelier than a pure HVDC facility, says Teong Chuan.

In my view, here are two things going for HVDC:

🔸Even without AI workloads, HVDC can help ever-larger data centres cut down significantly on energy losses.

🔸When one considers the fact that renewables such as solar photovoltaics produces DC, HVDC makes even more sense.

What do you think of HVDC in the data centre?

𝗣𝗦: I also looked at the Green Data Centre Park plot, which I'll write about another day. Do sign up for my newsletter so you don't miss it.

📆 𝗚𝗲𝘁 𝗺𝘆 𝗰𝗼𝗻𝘁𝗲𝗻𝘁

✅ Get my weekly newsletter: www.techstories.co/updates
✅ Join my WhatsApp Channel: www.techstories.co/wa

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Paul Mah. Mingcheng Lim speaking. HVDC solutions from some partners.

Micron today broke ground on a new advanced wafer fabrication facility at its North Coast Drive manufacturing compound. Here's what it means.

✅ Join my WhatsApp Channel (one-click): https://lnkd.in/gfZUWPSe

𝟭/ 𝗧𝗵𝗲 𝗻𝗲𝘄𝘀

First, the news. Micron Technology today broke ground on an advanced wafer fabrication facility that will produce NAND wafers for high-performance storage.

Micron president and CEO Sanjay Mehrotra was in town for the groundbreaking, as was DPM Gan Kim Yong, MTI's Beh Swan Gin, EDB's Jermaine Loy, and JTC's Jacqueline Poh.

Micron says the upcoming facility represents a planned investment of US$24 billion over 10 years. Wafer output is scheduled to begin in the second half of 2028.

Some notable facts:

- First double-storey wafer fab here.
- To co-locate manufacturing with R&D.

It's worth noting that Micron just last year announced an investment of US$7 billion on a HBM advanced packaging facility. They are located side-by-side.

𝟮/ 𝗡𝗔𝗡𝗗 𝗮𝗻𝗱 𝗗𝗥𝗔𝗠

According to a 2024 report, Micron gives Singapore a 10% share in the high-end global memory market. What is memory anyway?

Modern electronics relies heavily on two main categories of memory: DRAM and NAND.

🔸GPUs used for AI require large amounts of HBM which is an advanced, extremely high-performance form of DRAM.

🔸NAND is used to make modern SSD storage drives used by consumers and businesses.

In AI data centres, one could rightly say that one feeds data to GPUs at extreme speed, while the other stores it.

𝟯/ 𝗠𝗮𝘀𝘀𝗶𝘃𝗲 𝗲𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺

Memory production is a complex endeavour that requires a massive ecosystem.

- Materials, chemicals, and tooling.
- Wafer fabrication.
- Packaging.
- Testing.
- Integration.

This means working with a tightly interlinked network of manufacturing facilities along with a plethora of suppliers and partners.

𝟰/ 𝗠𝗶𝗰𝗿𝗼𝗻 𝗶𝗻 𝗦𝗶𝗻𝗴𝗮𝗽𝗼𝗿𝗲

Micron has dozens of facilities around the world, including in Malaysia. In Singapore, today's announcements would bring its investments to more than US$60 billion since 1998, says CEO Mehrotra.

Micron's investments in Singapore have accelerated in monetary terms. Using publicly announced figures:

- 𝟮𝟬𝟮𝟲: US$24 bil.
- 𝟮𝟬𝟮𝟱: US$7 bil.
- 𝟭𝟵𝟵𝟴-𝟮𝟬𝟮𝟰: More than US$29 bil.

One observation: the new NAND facility will be built up in phases in line with market demand.

In its press release, Micron says it will "maintain flexibility in managing the pace of capacity ramps to align with market demand."

𝗣𝗵𝗼𝘁𝗼𝘀:
4. Micron's existing compound being expanded. Under construction: (Top left) HBM facility (Top middle) NAND facility.

📆 𝗚𝗲𝘁 𝗺𝘆 𝗰𝗼𝗻𝘁𝗲𝗻𝘁

✅ Get my weekly newsletter: www.techstories.co/updates
✅ Join my WhatsApp Channel: www.techstories.co/wa

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Paul Mah

Do you aspire to be a Top Voice? LinkedIn has finally published its requirements and best practices. And oh, you can even nominate yourself. Here's how.

Earlier this week, Admond Lee shared in a group chat a new LinkedIn FAQ none of us had seen before. Here's what you need to know.

✅ Join my WhatsApp Channel (one-click): https://lnkd.in/gfZUWPSe

𝟭/ 𝗟𝗶𝗻𝗸𝗲𝗱𝗜𝗻 𝗧𝗼𝗽 𝗩𝗼𝗶𝗰𝗲

According to LinkedIn, Top Voice is an invitation-only programme designed to recognise members from a select group of experts, leaders, changemakers, public figures, and innovators.

For the longest time, scant details were available from LinkedIn; almost all information came from influencers who shared tips on how to get it. I'll also note that most of them are no longer a Top Voice.

I admit I originally made getting a Top Voice a goal when I started writing daily in 2023. However, the desire faded over time. Why? I realised all rankings are inherently biased by default.

In fact, when I received my Top Voice somewhere in August 2024, I never wrote a post about it. Accolades are fleeting, after all. Crucially, I realised I gained great joy in writing and sharing.

But if it's your cup of tea, here's what matters.

𝟮/ 𝗪𝗵𝗮𝘁 𝗟𝗶𝗻𝗸𝗲𝗱𝗜𝗻 𝗹𝗼𝗼𝗸𝘀 𝗳𝗼𝗿

Here are 3 key points that caught my attention.

(𝗮) Not perpetual

LinkedIn Top Voice badges are assessed and renewed twice a year. If you are invited, you will remain for at least 6 months, after which it will be reviewed and renewed. Top Voice status can and is removed; many have lost it over the years.

(𝗯) Originality

A distinctive voice and unique perspectives matter. How I see this is: If your post can be copied and shared verbatim by just about everyone else on the platform, then it's probably not very distinctive or unique.

(𝗰) Insights

LinkedIn looks for the sharing of expertise and insights. It doesn't matter if it refers to recent news, is a commentary, or an analysis. The key is whether it adds value to an industry or the community.

𝟯/ 𝗠𝘆 𝗰𝗹𝗼𝘀𝗶𝗻𝗴 𝘁𝗵𝗼𝘂𝗴𝗵𝘁𝘀

It's worth noting that the FAQ mentions "prominence" as a consideration. So don't compare yourself to the CEO of a large firm or a prime minister and beat yourself up.

Can you use fully AI-generated content and become a Top Voice? While this isn't addressed, readers generally don't like obvious AI slop - and I suspect LinkedIn takes that into consideration. But I don't think using AI to edit will compromise your chances.

I've written for almost 1,000 days now, and it is my belief that real-life influence matters more than metrics, rankings, or accolades.

If Top Voice comes, great. If not, keep writing anyway.

🔸LinkedIn Top Voices FAQ: https://lnkd.in/g9T7HT3E
🔸Nominate yourself: go.techstories.co/KqWA

📆 𝗚𝗲𝘁 𝗺𝘆 𝗰𝗼𝗻𝘁𝗲𝗻𝘁

✅ Get my weekly newsletter: www.techstories.co/updates
✅ Join my WhatsApp Channel: www.techstories.co/wa

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Screenshot

Walk into a data centre campus and you will invariably see massive storage tanks for water. Here's why data centres are so thirsty and why this won't stop overnight.

✅ Join my WhatsApp Channel (one-click): https://lnkd.in/gfZUWPSe

𝟭/ 𝗛𝗼𝘄 𝗱𝗮𝘁𝗮 𝗰𝗲𝗻𝘁𝗿𝗲𝘀 𝗹𝗼𝘀𝗲 𝘄𝗮𝘁𝗲𝗿

Data centres rely heavily on water. Inside, a vast network of pipes routes water through cooling towers, chillers, pumps, heat exchangers, and computer room air handlers.

Water is lost through two primary ways:

- Evaporative cooling to expel heat.
- As mineral builds up, water is discarded (blowdown).

This isn't new. What is new is the scale of data centres. Data centres of 5-10MW were considered large a decade ago. Today, data centre campuses are topping hundreds of MWs.

Even allowing for better economies of scale and increased efficiency with new systems, a 300MW data centre would consume almost 30x the water of a 10MW facility.

And it is impacting communities. In Johor, there is a temporary water moratorium for new data centres until 2027, as it races to divert water and build new water treatment plants.

𝟮/ 𝗖𝘂𝘁𝘁𝗶𝗻𝗴 𝗱𝗼𝘄𝗻 𝘄𝗮𝘁𝗲𝗿 𝘂𝘀𝗲

Are there ways to reduce water or even do without it? There are. But there are trade-offs.

(𝗮) 𝗧𝗿𝗮𝗱𝗲-𝗼𝗳𝗳𝘀

Dry coolers reject heat without water, but this typically comes at the cost of higher energy consumption compared with evaporative cooling systems.

(𝗯) 𝗭𝗲𝗿𝗼-𝘄𝗮𝘁𝗲𝗿

Microsoft has talked about its zero-water evaporation design for data centres. I couldn't find technical details for it, but what I found suggests using elevated water temperatures and direct-to-chip cooling for everything (which costs more).

More tellingly, a New York Times report earlier this week noted that water consumption by Microsoft data centres will continue to rise as it builds new data centres. So it's clearly not using this system for all its new data centres.

Crucially, it's not possible to do either (a) or (b) without retrofitting existing data centres, a costly endeavour that few operators will embark on.

𝟯/ 𝗥𝗲𝗱𝘂𝗰𝗶𝗻𝗴 𝗯𝗹𝗼𝘄𝗱𝗼𝘄𝗻

Another way to save water is to minimise blowdown water. This is water that must be discarded as minute traces of minerals and impurities build up as water is evaporated.

Failure to do that in a timely manner can result in scale buildup, which can reduce cooling efficiency and ultimately damage equipment.

What can data centre operators do on this front?

- Treatment to improve quality of incoming water.
- Use chemicals as inhibitors.
- Filtration and other treatments.

Note that this presupposes the continued use of water, albeit with less wastage for much better water usage efficiency (WUE).

I'll talk more about this in another post.

📆 𝗚𝗲𝘁 𝗺𝘆 𝗰𝗼𝗻𝘁𝗲𝗻𝘁

✅ Get my weekly newsletter: www.techstories.co/updates
✅ Join my WhatsApp Channel: www.techstories.co/wa

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Paul Mah. Water storage tanks in a data centre in Johor.

Humans are starting to sound like AI, even in spoken conversations, says a study that evaluated 360K YouTube videos and 771K podcasts. 

Ever felt like AI is changing how we write and talk? It's not your imagination, according to a study that concluded LLMs are changing how we speak.

𝟭/ 𝗜𝗻𝗳𝗹𝘂𝗲𝗻𝗰𝗲 𝗼𝗳 𝗟𝗟𝗠

The paper "Empirical evidence of Large Language Model’s influence on human spoken communication" was first published in Sep 2024 and substantially updated in July 2025.

The paper found a "measurable and abrupt increase" in the use of words preferentially generated by ChatGPT after its release.

The report noted: "These findings suggest a scenario where machines, originally trained on human data and subsequently exhibiting their own cultural traits, can, in turn, measurably reshape human culture."

Or put simply, our use of LLMs has resulted in a "closed cultural feedback loop" where cultural traits circulate bidirectionally between humans and machines.

Indeed, v1 of the report noted how a random sample of 50 manually reviewed videos found that the majority who spoke the word "delve" were not looking at a script.

𝟮/ 𝗖𝗹𝗲𝗮𝗿 𝗰𝗼𝗺𝗺𝘂𝗻𝗶𝗰𝗮𝘁𝗶𝗼𝗻

In the same vein, AI use is pervading written content and social media platforms. To me, the issue goes beyond sounding like AI, but about offering genuine insights and clarity of communication.

Depending on how it's applied, AI has a tendency to muddy or produce averaged-out content that is fluff. And those who are most reliant on it are unlikely to notice because they simply don't read at that level.

This is something B2B firms must pay special attention to. Just because you don't notice the AI slop doesn't mean the experts you're marketing to won't. Some will, and it can harm the brand.

On LinkedIn, I personally use an AI-detector I trust to give me an additional signal about whether a creator is worth paying attention to.

Anyway, there were initial concerns about potential AI model collapse. Now, I can't help but wonder if it is human communication that will 'collapse' instead.

---
My name is Paul Mah and I offer sharp takes on AI, cybersecurity, data centres, enterprise IT, and sustainability. Follow me on LinkedIn for regular updates (direct link): https://lnkd.in/gu5EMKQg.

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Screenshot from report.

The Huawei MatePad Pro 12.2-inch tablet is a premium tablet complete with a full-sized keyboard at an almost-bargain price.

I had the chance to try out one of the latest devices from Huawei's premium tablet over several weeks. My thoughts? It's a stellar device that offers outsized value.

𝟭/ 𝗚𝗲𝗻𝘁𝗹𝗲 𝗼𝗻 𝘁𝗵𝗲 𝗲𝘆𝗲𝘀

The MatePad Pro is the premium line of tablets from Huawei, and this particular model comes with a 12.2-inch tandem OLED "PaperMatte" display.

The tandem OLED is gorgeous with up to 2000 nits of peak brightness, while the display is matte through nano-level etching, and is very pleasant and gentle on your eyes.

The tablet itself weighs 512 grams, and when paired with the bundled keyboard case and stylus, brings it to just 850grams.

𝟮/ 𝗞𝗲𝘆𝗯𝗼𝗮𝗿𝗱 𝗮𝗻𝗱 𝘀𝘁𝘆𝗹𝘂𝘀

Unlike other brands, Huawei has been bundling keyboard case and stylus with its tablets, which I used together with the MatePad Pro 12.2.

As a touch typist, I really loved the keyboard, which offers full-size key pitch and great key travel. I find the typing experience akin to a laptop, which says a lot.

The Huawei M-Pencil (3rd gen) stylus was adequate for writing, and Huawei Notes is a very polished app designed just for that. "Feel" cannot compare to my Remarkable Pro but it does feel less plasticky than other styluses that I've tried.

For handwriting across apps, Huawei has its FreeScipt as the equivalent to the iPad's Scribble capability. FreeScript worked well in my experience.

Both peripherals are charged wirelessly. The stylus attaches to the side iPad-like where it charges. The keyboard also snaps on magnetically and charges its built-in batteries automatically.

They communicate with Nearlink wireless, which is distinct from Bluetooth. Pairing is as simple as attaching them magnetically to the tablet and agreeing to pair. That's it.

𝟯/ 𝗚𝗿𝗲𝗮𝘁 𝘃𝗮𝗹𝘂𝗲

Like other Huawei devices, accessing Android apps require jumping through a couple of initial hoops. You could either use the open-source MicroG app or the free GBox app.

- MicroG: Substitute for Google Play Services.
- GBox: Virtual machine environment.

In my opinion, the Huawei MatePro line is an underappreciated product range. Priced at ~SG$1,000 (can be lower during promotions) inclusive of keyboard and stylus, it offers a premium tablet experience without a premium-level price tag.

Of course, it's not perfect. Windowing experience trails behind the iPadOS 26 - which is exceptional. And there are no mobile-enabled variants.

The Huawei MatePad Pro 12.2 is sold out in Singapore right now, though a new model should replace it soon. I expect new models to offer similar value.

---
My name is Paul Mah and I offer sharp takes on AI, cybersecurity, data centres, enterprise IT, and sustainability. Follow me on LinkedIn for regular updates (direct link): https://lnkd.in/gu5EMKQg.

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Paul Mah.

Each time we think we have AI figured out, everything changes with new models or ways to use them. Maybe we are all thinking about AI wrong.

We keep trying to put AI into a box. But perhaps what really matters is not what AI can or cannot do, but how we choose to use it.

I had fun putting this together. You can read the full Substack piece here (1,000 words): go.clearlytech.co/E2DY

Here's the quick read version:

𝟭/ 𝗧𝗵𝗲 𝗯𝗹𝗶𝗻𝗱 𝗺𝗲𝗻 𝗮𝗻𝗱 𝗔𝗜

In the classic blind men and the elephant parable, each person touches a different part of the animal and draws wildly different conclusions. None of them are wrong, yet they all completely miss the bigger picture.

Each time I write about the use of generative AI in writing, I’ve received wildly different responses. The reason? Everyone is using it differently.

𝟮/ 𝗧𝗵𝗲 𝗹𝗶𝗻𝗴𝘂𝗶𝘀𝘁𝗶𝗰 𝗰𝗮𝗺𝗼𝘂𝗳𝗹𝗮𝗴𝗲

AI can effectively beautify poor or shallow writing, making it appear insightful until you actually read it. But the cracks appear quickly once you look beyond the synthetic coherence.

I personally think the best content will continue to be human written for a long time to come. Because AI cannot replace the craftsmanship and discernment needed to frame and write exceptional stories.

𝟯/ 𝗧𝗵𝗲 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗮𝗹 𝗽𝗮𝘁𝗵

Here are some ways I use AI to help me get better at writing, not replace me.

- Quickly extract quotes from a transcript.
- Critique my long-form content.
- Fine-tune text when copywriting.

𝘁𝗹𝗱𝗿;

A calculator doesn’t make someone a mathematician; a spell-checker doesn’t make someone a writer. AI amplifies existing capability rather than replacing it.

When it comes to writing, the question isn’t whether to use AI. It’s whether you have something to say.

What do you say?

𝗽𝘀: Sign up at clearlytech.co to get my latest Substack content delivered to your inbox.

📆 𝗚𝗲𝘁 𝗺𝘆 𝗰𝗼𝗻𝘁𝗲𝗻𝘁

✅ Get my weekly newsletter: www.techstories.co/updates
✅ Join my WhatsApp Channel: www.techstories.co/wa

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: Unsplash/Randy Laybourne.

A consortium led by KKR and Singtel is in advanced talks to buy STT GDC, which has a network of over 100 data centres worldwide. Here's what the deal is about.

✅ Join my WhatsApp Channel (one-click): https://lnkd.in/gfZUWPSe

𝟭/ 𝗟𝗮𝗿𝗴𝗲𝘀𝘁 𝗱𝗲𝗮𝗹 𝗶𝗻 𝗔𝘀𝗶𝗮

This news was first reported in November, with KKR and Singtel said to be seeking to borrow SG$5 billion for a proposed acquisition.

On Saturday, a new report by the Wall Street Journal noted that talks are in an "advanced" stage for a deal that would value STT GDC at more than $13 billion (US$10 billion).

According to Bloomberg on Sunday, Singapore's GIC and Abu Dhabi's Mubadala are in discussions to join KKR and Singtel as minority co-investors.

If completed, this acquisition would rank among Asia's biggest data centre transactions.

𝟮/ 𝗦𝗧𝗧 𝗚𝗗𝗖

ST Telemedia currently owns 82% of STT GDC, which it established in 2014 to push into the data centre space. An earlier SG$1.75 billion investment in 2024 means KKR already owns around 14% and Singtel around 4%.

ST Telemedia is wholly owned by Temasek Holdings.

STT GDC has grown rapidly over the years, through a combination of acquisitions, joint ventures, and greenfield data centres. It moved early and adopted a diversified strategy to enter multiple markets, long before generative AI entered the spotlight.

Today, STT GDC has a combined capacity of more than 2.3GW spread over more than 100 data centres located in a dozen countries: six in Southeast Asia, India, South Korea, Japan, and three others in Europe.

𝟯/ 𝗪𝗵𝗮𝘁 𝗶𝘀 𝘁𝗵𝗲 𝗱𝗲𝗮𝗹 𝗮𝗯𝗼𝘂𝘁

As I wrote in November, an acquisition would help build Singtel Nxera into a leading, global data centre player.

Singtel Nxera is currently developing new data centres in Johor, Batam, and Thailand. However, these take time to construct and build up. An acquisition would propel Nxera into a powerhouse overnight.

Moreover, Nxera's consolidation in Singapore with the closure of five old data centres leaves it with just three facilities. STT GDC's six in Singapore would greatly bolster its geographical coverage in the important data centre hub.

There is one more consideration. While the AI frenzy has fuelled interest in data centres, the reality is that many key regions require more digital infrastructure for non-AI deployments.

On this front, STT GDC’s substantial footprint will serve as a crucial resource whether in terms of familiarity with the region, training new data centre specialists, or doing R&D.

📆 𝗚𝗲𝘁 𝗺𝘆 𝗰𝗼𝗻𝘁𝗲𝗻𝘁

✅ Get my weekly newsletter: www.techstories.co/updates
✅ Join my WhatsApp Channel: www.techstories.co/wa

𝗣𝗵𝗼𝘁𝗼 𝗖𝗿𝗲𝗱𝗶𝘁: STT GDC. Upcoming STT VNG Ho Chi Minh City 2 in Vietnam.

taplio

The Battle

Daniel Tay

Paul Mah

Followers

Posts

Engagements

Likes

Performance ScoreDaniel Tay's LinkedIn Performance

Performance ScorePaul Mah's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Daniel Tay

Paul Mah

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Daniel Tay's LinkedIn Performance

Performance ScorePaul Mah's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreDaniel Tay's LinkedIn Performance