10 minute read
AI’s Depreciation Trap: How Big Tech’s Accounting Math Could Catch Up to It
- Big Tech’s AI spending spree may be inflating short-term profits as companies stretch depreciation schedules on hardware that’s becoming obsolete faster than ever.
- Once accounting catches up with the real pace of AI hardware turnover, investors could see rising costs and thinner margins that test today’s lofty valuations.
Why it matters: If Big Tech’s profits are being flattered by slow depreciation rates, investors may be underestimating the true cost of maintaining AI infrastructure. As the pace of hardware obsolescence accelerates, future earnings could face a sharp hit when accounting finally catches up to reality. What looks like strong profitability today may be masking a cycle of rising costs and shrinking margins tomorrow - a familiar setup for corrections when optimism runs ahead of economics.
Now the Deep Dive: Today’s AI market may be running headfirst into a wall.
Sure, the massive AI capital spending looks great for investors now (and has definitely boosted growth) - but it may be setting itself up for future downside.
Why? Because of depreciation2 - aka the “wear and tear” of accounting.
When a company buys equipment, it doesn’t expense it all at once - it spreads that cost over several years, assuming the equipment slowly loses value. The longer you say it lasts, the smaller the yearly expense - and the bigger your profit looks.
“So what’s the big deal? Every company does this.”
True. But AI may change things.
See, in this AI arms race, having the best hardware isn’t optional - it’s becoming essential. Falling even one or two generations behind can mean losing 30–70% in efficiency - and with it, your competitive edge. Thus, the lifespan of AI equipment is shrinking faster than many expect.
For instance:
- GPUs - especially in data centers - are often replaced every 18 to 24 months.
- Networking systems are constantly redesigned to handle more data.
- And the move from training to running AI (inference) forces even quicker upgrades.
Think of Big Tech’s AI build-out like owning a fleet of race cars that need constant upgrades. Every year brings a faster engine or better transmission, and last year’s models are suddenly outclassed.
- Compare that to a bulldozer at a mining firm – it’ll still likely bulldoze the same ore even after seven years. Thus, the difference in performance between old and new is relatively small. But in AI? The drop-off between chip generations can be massive.
This is where the math starts to bite.
Say a company spends $100 million on servers it plans to depreciate over seven years - about $14 million a year. But if those servers need replacing after five, the company buys another $100 million in new gear.
Now, depreciation jumps to roughly $34 million a year. Nothing shady - just that assumptions about “useful life” can directly impact future earnings and erase the profit boost that looked so juicy at first.
- Note that this is just a simplified illustration - in practice, firms may write off old assets rather than carry overlapping depreciation, but the effect on total costs is similar via rising depreciation and shrinking margins.
And the stakes could be enormous.
But on paper, Big Tech still assumes its “cars” will last 4-6 years. That keeps the annual “wear and tear” charge smaller, making profits look better - even though the hardware will likely be replaced long before the books say it should.
Keep in mind that these useful-life assumptions were set in a different era. Now, the pace of AI innovation means hardware becomes obsolete far ahead of accounting schedules. Thus, when accounting life lags real life, earnings can be inflated. Profits look stronger, valuations look attractive, and analysts cheer - that is, until the bill for all that new equipment comes due.
Of course, firms like China’s DeepSeek show there’s another path. By squeezing every ounce of performance from older chips, they’ve proven that clever engineering can sometimes beat the newest hardware.
But for most of Big Tech, growth depends on constant reinvestment - not ingenuity. And when billions in assets need replacing years ahead of schedule, the illusion of rising profitability can fade fast.
On the other hand, if these AI investments ultimately deliver higher productivity and profit growth, maybe the trade-off will be worth it - just the growing pains of the next expansion.
If not, the coming depreciation wave could turn this boom into a long, expensive hangover.
Either way - it’s something to mull over.
Figure 1: Bloomberg, November 2025
Extreme Fear Index Flashes Warning - But Liquidity Could Spark the Next Rally
- Markets are gripped by extreme fear just as liquidity pressures may be easing — a setup that’s punished sellers before rewarding patient buyers.
- If government spending and the end of Fed tightening flood cash back into the system, today’s panic could become tomorrow’s rebound.
What you need to know: The Fear & Greed Index is pointing deep into fear right now - showing that markets are jittery, sentiment is weak, and many investors are pulling back4.
Why it matters: When fear dominates like this, stocks often get sold indiscriminately and valuations can fall to levels where the “margin of safety” widens - meaning the upside potential improves because downside risk is lower.
Now the Deep Dive: We all know that sentiment drives markets just as much as earnings do.
- If investors are full of greed and “hopium” – then markets rise regardless of fundamentals.
- And if markets are riddled with fear – they won’t want to touch anything with a 10ft stick.
Well, right now markets are terrified.
To put this into perspective, the CNN Fear & Greed Index - which tracks investors mood on a scale from 0 (Extreme Fear) to 100 (Extreme Greed) - is sitting deep in fear territory. In fact, as of writing this, it’s at 8/100.
For context, this is the lowest reading since the April market chaos amid the trade war.
That tells us two things.
First - panic is dominating. When investors are this nervous, stocks often get sold across the board, regardless of quality. It’s driven by emotional selling. And historically, markets have tended to bottom when fear is high and confidence is low (albeit there’s some real issues in the economy that justify this).
Second - opportunity may be forming. When fear pushes prices down this fast and hard, it can widen what investors call the margin of safety - aka the gap between what something is worth vs. what it costs. And the bigger that gap, the less you risk losing and the more potential you have to gain if sentiment turns around.
- Buying after a big market drop is like falling from five feet instead of twenty - you can still get hurt, but the ground’s a lot closer, so the risk of real damage is relatively less.
Now, I’m not saying this is the bottom (and be wary of anyone who does). But this level of fear has a way of shaking out weak hands and resetting prices lower.
There’s also something else going on beneath the surface here.
I’m talking about liquidity.
I warned about this back in early October in “Bank Reserves Are Shrinking – Is a Liquidity Crunch Next?” – noting that liquidity was drying up fast as the Treasury rapidly rebuilt its cash pile, the Fed's sucked out money through its QT (quantitative tightening) program, and the overnight reverse repo facility was emptied out (I called this the Unholy Trinity).
This matters because when liquidity drains, it can freeze the economy’s pipes - leading to broad sell-offs across everything from small-caps and crypto to gold and Big Tech.
- Said another way, things that looked great when liquidity was flush aren’t looking so great now that less money is sloshing around.
Well, that’s exactly what today’s market feels like - red across the board.
But here’s the upside - the very forces that drained liquidity may now be reversing.
- The U.S. government has reopened and will begin spending down its nearly $1 trillion cash pile, pushing money back into the system.
- The Fed has signaled it will end QT in December, easing another drain on reserves.
If those tides truly reverse, liquidity could flood back into markets just as fear peaks.
Pair that with extreme pessimism, and you may have the seeds of the next upswing.
The point is this - extreme fear can either mark the start of a real crisis - or the moment when the crowd finally runs out of sellers, right before liquidity returns.
As always, time will tell.
Figure 2: CNN.com, November 2025
Over Half of U.S. Homes Are Losing Value — Is Housing Finally Resetting?
- More than half of U.S. homes are falling in value as high mortgage rates and affordability pressures cool a multi-year boom.
- While that’s painful for homeowners, it’s creating new entry points — and perhaps the start of a long-overdue housing reset.
What you need to know: More than half of U.S. homes - 53% - have lost value over the past year, the highest share since 2012, according to Zillow5.
Why it matters: After years of soaring home prices, the housing market is cooling fast and broadly. Rising mortgage rates, affordability crunches, and softening demand relative to more sellers are pushing values lower - and for many Americans, that means the wealth effect that once fueled spending is reversing, especially as a surging amount of homeowners are becoming “underwater”. But on the other hand, this gives would-be home buyers a better price to buy.
Now the Deep Dive: The U.S. housing market hasn’t seen declining home prices like this in over a decade.
As of October 2025, over half of homes have declined in value - a sharp reversal from the pandemic-era boom when nearly everything with a roof was appreciating.
Back then, cheap borrowing, underbuilding, and FOMO (fear of missing out) spurred bidding wars.
But now, it’s the opposite direction.
Here are just some of the worrying numbers:
- >6% mortgage rates + elevated prices have strangled new buyers - with the median home price still up roughly 60% since 2019.
- Affordability has collapsed as the share of median income going toward housing hit 43% in September (anything above 30% is considered unaffordable)6.
- Over 900,000 homeowners are now underwater - meaning they owe more than their homes are worth (the highest level in three years)7.
- Mortgage refinancing rejection rates hit 46% in October, the highest since 2013 (those who want to try and lighten their interest costs can’t even do it).
Thus, the momentum remains to the downside for housing (although it varies by state).
Now, this isn’t necessarily a crash, but a reset after years of unsustainable growth – which is healthy.
But that doesn’t mean the ripple effects aren’t significant.
- Consumer confidence: Housing is the biggest source of wealth for the bottom 90% of households. When values fall, spending can slow - dragging down retail and local economies.
- Credit markets: Home equity lines and refinances dry up, tightening liquidity and the potential disposable income that could’ve been spent elsewhere (aka every $1 saved on housing could’ve been used to buy something else).
- Psychology: The belief that real estate “only goes up” is sowing doubt - changing how people invest and borrow (seem familiar?)
Yet there’s a silver lining in all this.
For first-time buyers priced out during the pandemic, lower prices may finally open doors - literally.
And for long-term housing investors, downturns like this often reset the market, laying the foundation for the next phase of healthy growth.
Still, I’m cautious about how much wealth is tied up in real estate as unaffordability spreads.
History shows that when property bubbles burst, they hit harder than most other downturns - think U.S. 1929, Japan 1991, the U.S. housing crisis in 2008, and China today.
We’re not at those points yet, but it’s something worth watching - especially as housing affordability becomes an increasingly political fault line.
Something to monitor.
Figure 3: Zillow, November 2025
Anyway, who knows how this will all play out?
This is just some food for thought as we watch how these trends develop.
We’ll be keeping a close eye on things. Enjoy the rest of your weekend.
Sources:
- Burry’s Depreciation Gripe Shines Spotlight on Big Tech Profits
- Understanding Depreciation: Methods and Examples for Businesses
- Peter Berezin| Hyperscaler CAPEX estimates | LinkedIn
- Fear and Greed Index - Investor Sentiment | CNN
- 53% of U.S. homes lost value in the past year, the most since 2012 - 2025
- Home Ownership Affordability Monitor - Federal Reserve Bank of Atlanta
- Nearly 900,000 new homeowners are underwater on their mortgages, signaling a troubling shift in the housing market - MarketWatch
Disclosures:
This communication is general in nature and provided for educational and informational purposes only. It should not be considered or relied upon as legal, tax or investment advice or an investment recommendation, or as a substitute for legal or tax counsel. Any investment products or services named herein are for illustrative purposes only, and should not be considered an offer to buy or sell, or an investment recommendation for, any specific security, strategy or investment product or service. Always consult a qualified professional or your own independent financial professional for personalized advice or investment recommendations tailored to your specific goals, individual situation, and risk tolerance.
Information contained in the materials included is believed to be from reliable sources, but no representations or guarantees are made as to the accuracy or completeness of information.
Dunham & Associates Investment Counsel, Inc. is a Registered Investment Adviser and Broker/Dealer. Member FINRA/SIPC. Advisory services and securities offered through Dunham & Associates Investment Counsel, Inc.