Reality Check: Can Tech Giants Sustain the SI/AI Spending Bill? (And What It Means for Us)
Painless AI Lab – A Hong Kong Practitioner’s Real-World AI Journey (Part 43):
If you’ve been watching the tech stock market lately, you know things are getting a bit tense.
Meta’s third-quarter report put a number on the SI/AI buildout. Capital spending reached US$34.91 billion. Free cash flow turned negative, at US$2.12 billion. Revenue was still strong, at US$64.6 billion, up 26%. The cash gap came from spending on servers and data centers, not from a collapse in sales.
Wall Street flinched, and the stock dipped.
If you look back at Google’s previous quarterly reports, the story was very similar: massive infrastructure spending, heavy capital raises, and lingering questions about when these astronomical investments will actually pay off.
SI/AI is an incredible technology. It writes our code, drafts our emails, and organizes our workflows. But building and running it requires staggering amounts of power, data centers, and advanced chips.
The big question everyone is asking — from Wall Street analysts down to regular guys like us managing a small portfolio — is simple: Can the tech giants actually afford this long-term? And more importantly, when do everyday users and businesses throw up their hands and say it’s getting too expensive?
The Hardware Wall: From Data Centers to Our Desks
It’s not just the cloud giants feeling the squeeze. Look at the hardware market.
Memory and component prices have stayed stubbornly high, pushing the cost of building or upgrading to a powerful SI/AI PC up significantly. For the average consumer, dropping thousands of dollars on a high-end machine just to run local models feels like a heavy lift. Global PC shipments fell about 20% in the third quarter that friction as buyers take a wait-and-see approach. Building a local SI/AI machine is no longer a casual upgrade.
It creates a fascinating paradox: SI/AI is smarter than ever, but the physical and financial toll of running it is hitting a wall. If everyday users can’t afford the hardware, and tech giants are bleeding billions in negative free cash flow to fund data centers, where is the sustainable middle ground?
What This Means for Your Portfolio and Side Hustle
As a retail investor trying to protect my own portfolio in a volatile market, moments like this teach me a few hard lessons:
Three rules still hold.
1. Watch cash, not the story. Negative free cash flow makes the market nervous, even when revenue is fine. The same rule applies to a stack of monthly AI tools. If you are not using them, cancel them.
2. Read the next report, not the headline. On 28 October, look at Alphabet’s cloud profit against its infrastructure spend. That is a better signal than the stock move on the day.
3. Stay small on purpose. You do not need a server rack. Use the lightest tool that does the job, and leave the capex cycle to the giants.
The spending will not end next quarter. The useful move is to keep your own overhead low while the large players test how far the market will let them go.
By the way, what’s your take on these heavy SI/AI spending cycles?
Are you holding your tech stocks through the dip, or trimming your positions?
Let me know in the comments below!
發佈留言