Meta’s new AI chips will begin production in September | TechCrunch

Meta's new AI chips will begin production in September | TechCrunch — featured image

by

`. Only HTML. No `

`.
* **Structure:**
* Hook (Why care?)
* What Happened (2-3 paragraphs)
* Why This Matters (for Malaysian SMEs)
* The Bigger Picture (long-term trend)
* CTA (Book a free 15-min call -> https://autorunbiz.com)
* **Quality Rules:**
* Source links for data points.
* Qualifying language for opinions.
* NO AI clichés.
* Second-person (“you”).
* Conversational/exciting.
* Grounded, no hype.
* NO cost/price mentions in body (wait, “cost” as in operational cost is okay, but “price/money” is tricky. “saving money” vs “$X”. “trillions of dollars” is a quote from the source. I can talk about strategy without pricing. The rule says “No cost/price/money mentions in the content”. *Wait*, the source explicitly talks about capital expenditure and saving. “Save on buying GPUs”. “Capital expenditures between $125 billion and $145 billion”. I need to strictly obey. “No cost/price/money mentions in the content.” This means I cannot say “saves money” or “costs billions”. I can talk about *strategy*, *competition*, *dependence*, *access to technology*, *efficiency*, *control*. Let’s focus on *power*, *control*, *supply chain*, *access to best tech*, *competition*, *innovation*. The article is explicitly about saving money. How do I write about this without mentioning cost/price/money? “Lower its GPU costs” is the core of the news. Instead of saying “Meta wants to lower costs/ save money”, I can say “Meta is working to reduce its dependence on external suppliers”, or “Meta is taking control of its hardware needs”. The user must be obeyed strictly. Let’s frame it around *control*, *supply chain security*, *technology access*, *competitive advantage*, *reducing reliance on a single supplier*.
* Wait, the user says “No cost/price/money mentions in the content”. I will completely avoid any mention of capital expenditures, cost, price, money, dollars, etc., in the body text. The CTA link is fine; that’s a service offering, not a price in the content.
* **Clichés to avoid:** revolutionize, game-changer, unlock, leverage, paradigm shift, in today’s landscape, it’s important to note, at the end of the day, delve, navigate.

2. **Deconstruct the Source News (TechCrunch):**
* **What:** Meta’s in-house AI chips (MTIA program) starting production in September. Made with Broadcom, TSMC, Samsung, Sandisk, Sumitomo.
* **Why:** Competing with Nvidia/AMD, securing supply chain, huge computing needs for AI (recommendations, ranking, inference for apps).
* **Context:** Meta has been making its own AI chips since 2023. Others doing it: OpenAI, Amazon, Google, Anthropic.
* **Key facts for me to link:**
* “Meta is on track to start making the latest versions of its AI-specific chip in September” (source link)
* “Working with Broadcom… TSMC to manufacture” (source link)
* “Meta detailed the four new chips… in March” (source link)
* “producing its own AI chips since 2023” (source link)
* “deploy 7 gigawatts of compute this year” (source link)
* “deal with ARM last year” (source link)
* “OpenAI… unveiled an inference processor” (source link)

3. **Drafting the Blog Post (Full HTML, adhering to ALL rules):**

* **Title:** Needs to be an `

`. e.g., `

Meta Is Making Its Own AI Chips. Here’s Why Your SME Should Pay Attention

` (or something similar, avoiding clichés).
* **Hook:** A big company like Meta building its own chips. This feels like a nerdy tech story, but it points to a massive shift in how AI is provided. For a business owner, this could mean better, faster, more accessible AI tools down the line. (wait, “accessible” could imply cost. Let’s say “more responsive” or “more integrated”).
* *Hook rewrite:* You might think a story about Meta building chips is just for the big tech players. But it actually reveals a lot about where AI is heading. And that directly affects the tools you will rely on for your business.
* *What Happened (Section 1):*
Meta is getting serious about owning its AI future. Starting this September, its latest AI chips are scheduled for production. Think of it this way: instead of entirely relying on a single supplier (like Nvidia) for the engines powering its AI, Meta is building its own engines. The company developed these chips under its MTIA program, working with Broadcom and TSMC.
The chips are designed for training AI models and running them (inference) on Meta’s apps.
* *Why This Matters for Your Business (Section 2):*
Why does a Malaysian SME owner care about Meta’s hardware strategy? Because it signals a massive shift in the AI supply chain.
For years, AI tool availability and performance have been heavily dependent on a handful of chipmakers. When everyone is fighting over the same limited supply, it creates a bottleneck. Meta, OpenAI, Amazon, and Google building their own chips suggests a future where AI capabilities are less dependent on a single global supply chain.
For your business, this competition is good news. It means AI providers will likely have more control, leading to more stable and integrated AI services in the platforms you use every day. Tools for customer support, content generation, and data analysis will likely get faster and better at understanding your specific needs because companies like Meta are no longer just renting compute power — they are building the hardware to match their software.
* *The Bigger Picture (Section 3):*
We are watching the tech giants untangle themselves from a single dependency. Meta isn’t just buying chips; it’s investing in a modular approach where hardware can adapt as AI models change. The company has made deals with ARM, AMD, and Amazon, showing a clear strategy to diversify.
For SMEs, the lesson is about control over your own tech stack. You don’t need to build chips, but you should be looking for tools and partners who are building for the long term. AI is no longer a futuristic novelty; it’s becoming core infrastructure. The companies that understand this and build their operations around reliable, specific AI tools will have a real edge.
When tech giants fight for control over their own AI destiny, the rest of us benefit from more powerful, more stable tools.
* *CTA:*
Curious about how the latest AI tools can help you run a tighter ship? Let’s talk. Book a free 15-min call to see how AI applies to your business → https://autorunbiz.com
* *Link injection:*
– “Meta is on track to start making the latest versions of its AI-specific chip in September” -> source url
– “Meta working with Broadcom” -> source url
– “Meta detailed the four new chips” -> source url
– “producing its own AI chips since 2023” -> source url
– “deploy 7 gigawatts of compute this year” -> source url
– “deal with ARM” -> source url
– “OpenAI unveiled an inference processor” -> source url
– “Amazon and Google both develop their own chips” -> source url
– “host of startups” -> source url
* *Avoiding Clichés:*
– No “quickly revolutionize the AI industry”.
– No “game-changer for your business”.
– No “unlock new potential”.
– No “leverage the power of”.
– No “paradigm shift”.
– No “in today’s landscape”.
– No “it’s important to note”.
– No “at the end of the day”.
– No “delve”.
– No “navigate”.
– Keep it conversational. “You might think a story about Meta building chips… but it actually…”.
– Qualifying language: “it feels like”, “this suggests”, “it points to a shift”.
* *Formatting Check:*
– `

` for main sections.
– `

` for subsections (maybe not needed given the structure, but can use for “What Happened”, “Why This Matters”, “The Bigger Picture”).
– Wait, the instructions say: “Use

for main sections,

for subsections.”
– Structure:
`

Meta Is Making Its Own AI Chips. Here’s Why It Matters for You

`
`

…hook…

`
`

What Happened

`
`

…details…

`
`

Why This Matters for Your Business

`
`

…SME connection…

`
`

Pull quote

`
`

The Bigger Picture

`
`

…trends…

`
`

CTA: Book a free 15-min call to see how [topic] applies to your business →

` -> link
* *Pull Quote:* “When tech giants fight for control over their own AI destiny, the rest of us benefit from more powerful, more stable tools.” (This is a strong opinion statement, suitable for a blockquote).

* *Specific Wording:*
– “You might think a story about Meta building chips is just inside baseball for tech giants. But it actually reveals a lot about where AI is heading. And that directly affects the tools you will rely on for your small business.”
– “Why does a Malaysian SME owner care about Meta’s hardware strategy? Because it signals a massive shift in the AI supply chain.”
– “For years, the AI tools you use have depended on a very specific global supply of chips. When everyone fights for the same limited components, it creates delays and uncertainty.”
– “Meta isn’t the only one. OpenAI, Amazon, and Google are also developing their own processors. This competition is healthy for everyone.”
– “For your business, this means the AI services baked into your social media marketing, customer support platforms, and automation tools will likely become more stable, more responsive, and better tailored to your needs.”
– “Meta is taking a modular approach to its chip design. This means they can adapt quickly as AI evolves.”
– “Data points: In 2023, Meta started producing its own AI chips. Meta detailed four new chips in March 2026. Meta plans to deploy 7 gigawatts of compute this year.”
– “Source links for the data points are embedded as per the source article provided.”

4. **Refining the Content (strictly avoiding clichés and cost/money):**
* The source explicitly talks about saving money. “In a bid to lower its GPU costs amidst an unprecedented component shortage”. I cannot mention “costs” or “component shortage” in a way that implies the *cost* of the chips, but I can talk about the *supply chain* and *dependency*. Let’s frame it around *control, reliability, and speed of innovation* rather than money.
* Instead of: “Meta wants to save money on GPUs.”
* Write: “Meta wants more control over its AI supply chain. By designing their own chips, they reduce their reliance on a single external provider. For a company running some of the world’s biggest AI models, control matters.”
* “It also signed a deal with ARM last year…” -> Link.
* “Meta has been producing its own AI chips since 2023” -> Link.
* “The chips are expected to help the company save on buying GPUs” -> I will replace this concept with “The chips are expected to give the company more flexibility and control over its AI operations.”
* “Capital expenditures between $125 billion and $145 billion” -> I will completely skip this sentence and the massive spending amounts.
* “deploy 7 gigawatts of compute this year…” -> Link. Frame as “massive computing capacity”.
* “OpenAI last month unveiled an inference processor… Anthropic is said to be considering developing its own chips.” -> Link.

5. **Drafting the HTML:**

“`html

Meta Is Building Its Own AI Chips. Here’s Why a Malaysian Business Owner Should Care

You might think a story about Meta building its own microchips is just tech world drama. But news like this actually tells you a lot about where the tools you use every day are heading. And for a small business owner, knowing what is shaping those tools is surprisingly useful.

Last week, reports confirmed that Meta’s latest AI-specific chips are scheduled to start production in September. The company is working with Broadcom on the design and Taiwan Semiconductor Manufacturing Company (TSMC) to build them [1]. These chips are part of their Meta Training and Inference Accelerator (MTIA) program, which Meta detailed in March.

What Happened

For a long time, most AI companies relied heavily on Nvidia for their computing power. Meta still buys from Nvidia and AMD, but it has been working on its own chips since 2023.

The latest version of the chip passed its testing phase in about six weeks, according to an internal memo seen by Reuters. Meta is taking a modular approach to these chips, meaning they are designed in pieces that can be swapped out or upgraded as AI models change.

This puts Meta in a group of tech giants who are bringing chip design in-house. OpenAI unveiled an inference processor last month, and both Amazon and Google have been developing their own AI training and inference chips for years.

Why This Matters for Your Business

If you run a small or medium business in Malaysia, you probably aren’t buying data center hardware. So why does this matter?

Because the entire ecosystem of AI tools you rely on runs on this hardware. When the companies behind your social media platforms, customer support bots, and content engines gain more control over their infrastructure, the services you use become more stable and responsive.

Think about it. When a handful of companies control the supply of a critical component, it creates bottlenecks. By building its own chips, Meta is ensuring that the AI features in its apps (recommendations, ranking, ad delivery) are not held back by external shortages or delivery delays. This competition among giants like Meta, Google, and Amazon likely means faster improvements and better integration for the tools your business depends on.

The Bigger Picture

The move towards custom chips tells us something important: AI is becoming core infrastructure, not just an add-on. Companies are no longer satisfied renting everything from a single provider. They are building the foundations themselves to make sure they can adapt quickly.

Meta is planning to deploy a massive amount of computing power this year and has signed multiple supply deals with ARM and others to secure its capacity. The company is writing specifically about how each generation of chips builds on the last, incorporating the latest AI workload insights.

“When tech giants fight for control over their own AI destiny, the rest of us benefit from more powerful, more stable tools.”

For your SME, the lesson is about being ready. AI is not a static tool you buy once. It is evolving rapidly. The companies that are best positioned to take advantage are those working with partners and tools that are also investing heavily in the future. You don’t need to build chips, but you do need to be aware of which companies are building for the long run and which are just riding the wave.

The future of business software will be defined by who owns the hardware underneath. And right now, the biggest players are making aggressive moves to own it entirely.

Curious about how the latest AI tools can be put to work in your daily operations? Let’s talk. Book a free 15-min call to see how AI applies to your business.

“`

6. **Review against constraints:**
* *Role:* Blog writer for AutoRunBiz (Malaysia). Met.
* *Audience:* Malaysian SME owners. Met (addressed as “you”, specific to Malaysia).
* *Source/Topic:* Meta’s AI chips.
* *Format:* Valid HTML, no markdown fences, no `

`, starts with `

`. Met.
* *Structure:*
* Hook (Why should a business owner care?) -> Yes, first paragraph.
* “What Happened” (2-3 paragraphs) -> Yes, section.
* “Why This Matters for Your Business” -> Yes, section.
* “The Bigger Picture” -> Yes, section.
* CTA at the end -> “Book a free 15-min call…” linking to `https://autorunbiz.com`. Met.
* *Quality Rules:*
* Every number/data point needs a clickable inline source URL. Let’s check:
* “production in September” -> Link.
* “detailed in March” -> Link.
* “since 2023” -> Link.
* “unveiled a processor” -> Link.
* “Amazon and Google” -> Link.
* “deploy… computing power” -> Link.
* “signed multiple supply deals with ARM” -> Link.
All data points link back to the exact same source article URL provided (TechCrunch). “Tier: T1:MY Tech” – this is just the source, no need to treat it differently, just link to it for the facts. The instruction says “every number/data point needs a clickable inline source url ()”. I have done this. Let’s add a few more explicit links for the “passed testing in six weeks” fact.
* Opinion claims use qualifying language: “likely means”, “probably aren’t”, “tells you a lot”.
* *AI Clichés avoided:* revolutionize, game-changer, unlock, leverage, paradigm shift, in today’s landscape, it’s important to note, at the end of the day, delve, navigate. (Checked).
* *Address as “you”:* Yes.
* *Conversational and exciting:* “You might think… but it actually…”, “Think about it.”.
* *Grounded, no hype:* Yes, focused on supply chain and stability.
* *No cost/price/money mentions:* “relying heavily on Nvidia” / “supply chain” / “control over infrastructure” / “external shortages” / “bottlenecks”. No dollar amounts, no “saving money”, no “lower costs”. The article says “to lower its GPU costs”, which is the motivation, but we can frame it entirely as “control over supply”. I have avoided the cost framing entirely. “supply chain”, “control”, “stability”, “responsive”, “dependency”. Good.
* *CTA:* Included and linked. No price mentioned in the CTA text.

* *Refinement of the opening:*
“You might think a story about Meta building its own microchips is just tech world drama. But news like this actually tells you a lot about where the tools you use every day are heading. And for a small business owner, knowing what is shaping those tools is surprisingly useful.”

* *Refinement of the “Why It Matters” section:*
I need to make it extremely concrete for a non-technical SME owner.
“When the companies behind your social media platforms, customer support bots, and content engines gain more control over their infrastructure, the services you use become more stable and responsive.”
“This competition among giants… likely means faster improvements and better integration for the tools your business depends on.”
“AI is becoming core infrastructure, not just an add-on.”

* *Pull Quote Check:*
`

“When tech giants fight for control over their own AI destiny, the rest of us benefit from more powerful, more stable tools.”

`
This feels like a strong pull-quote that encapsulates the value for the SME. It’s an opinion, so no source link required, but it’s grounded.

* *Paragraph Formatting Check:*
All text in `

` tags. `

` for sections. `

` for pull quote. No `