HN Reader

What's the most cost-effective way to run open source models using cloud infrastructure? AWS? Digital Ocean? salad.com? Lambda.ai? Vast.ai? Use case is infrequent and small usage, but needs to be able to scale to match increasing demand.

Did they raise prices because of increased or decreased demand? I can't help but wonder if things are overbuilt and being underutilized.

I guess somewhere this year having your own GPU might be cheaper than renting.

> AWS has spent two decades conditioning customers to expect prices only ever go down. That expectation is now broken.

So long for amazon’s “earn trust” leadership principle

Everything is getting monopolized (oligopolized) for rent extraction! Homes, healthcare, energy, compute. 'Capitalism' FTW! Coming up next: water, air.

- GPU prices rising

- RAM prices rising

- hard drive prices rising

Are we looking at a future where home computers are replaced by thin clients and all the power lies in subscription services?

‘You don't need storage space, use our cloud subscription’

‘You don’t need processing power, stream your games through our subscription service.’

Game publishers have already publicly floated the idea of not selling their games but charging per hour. Imagine how that impact Call of Duty or GTA.

Physical media could easily be killed off. Does my iPhone need 1TB of storage or will they shrink that and force everything through iCloud?

How long before car ownership is replaced with autonomous vehicle car pools? Grocery stores closed to visitors, all shopping done online and delivered to your door by drone.

Headline: hopes you weren't paying attention

Body: The change had been telegraphed: AWS's pricing page noted (and bizarrely, still does) that "current prices are scheduled to be updated in January, 2026," though the company neglected to mention which direction.

These do not seem entirely consistent?

.

> This comes about seven months after AWS trumpeted "up to 45% price reductions" for GPU instances - though that announcement covered On-Demand and Savings Plans rather than Capacity Blocks. Funny how that works.

Assuming I found the right pricing page(s), this new increased price is still lower than those other prices that were lowered.

The “hopes you weren't paying attention” part of the headline seems needlessly inflammatory. All I see is demand rising into a supply-limited market (GPUs and RAM). That doesn’t seem nefarious, just high-school economics.

I think a lot of businesses don’t actually need cloud AI at all. Once workloads stabilize, cloud is mostly a convenience tax. Most business use cases (docs, forecasting, monitoring, support, control systems) don’t need frontier models or hyperscale elasticity. Efficient models running locally are already “good enough”. Continuous inference + data gravity + latency/privacy constraints make owned edge hardware economically and operationally sensible again.

They increased the price of capacity blocks, not on demand. Capacity blocks pricing was promotional with a well defined end date from the day 1. And it was even lower than spot instance lot of times.

The price shock that’s about to hit the AI industry is gonna be very fun to watch.

A few months ago, there was a lot of news lambasting tech companies for extending the depreciation lifespan of GPUs from ~3 years to ~5 years. Do these price hikes suggest a longer lifespan is probably the right way to see how long these GPUs will be valuable?

On Demand GPUs got cheaper because Azure exists. Capacity Blocks got 15% pricier because your migration plan doesn't

Wait Corey Quinn is on El Reg now? That's awesome

Is there a reliable service that plots hourly price per GPU per cloud through time?

AI is choking everything else to death with its resource hogging.

I have a pair of Dell T7500 units for “slow and simple” self-hosting, and just snagged a trio of 1Tb/64Gb 2018 Mac Minis to join my own surplussed Mac Mini of the same vintage as single-purpose machines.

I have a gigabit symmetrical SOHO fibre connection. It’s time I brought everything back under one roof except for DNS.

To me this all seems like downstream from massive capital expenses on GPU purchases, driving up the demand for memory, etc. It begs the question of how much of the GPU capacity that has been sold/delivered actually being utilized? Are we paying across the board for a bunch of inventory sitting idle?

I'm just going to say: I bet it will eventually come out that RAM prices are yet another instance of illegal cartel collusion. For which all the RAM manufacturers have been convicted. More than once! If past incidents are any indication anytime there is a perceived "hot" market the RAM makers agree to limit production to current levels allowing market demand to increase prices, followed by production cuts if the market starts to soften.

GPU is a function of limited manufacturers and vendor lock-in combined with massive capex required to compete(). Like some (but not all) price inflation during and post COVID: rising prices can become a self-fulfilling prophecy. If all your customers expect prices to rise might as well meet their expectations and get while the gettin' is good.

() Like a new CPU architecture only worse. The amount of engineering required as "table stakes" increases exponentially while at the same time the manufacturing expertise does the same. This suddenly and rapidly raises the barrier to entry. Anyone who can survive the early squeeze can do quite well in such markets.

Demand for AI is growing. It makes sense they’re trying to increase profits.

Sad part is that most people actually won't notice.

I'm tired of AWS full of hidden fees and tricking you into things that you didn't know you were paying for until a month or two goes by.

Random theory:

- it is a practical test of price elasticity

(if demand doesn’t drop much, then it means profit can be increased more)

that just means i can up my rates as well. thanks amazon

I wonder how far we are down the track of enshittification of cloud providers... and how much more is to come.

Even the cloud won’t save you.

Maybe it might be a good idea to squash it with any legal avenue, especially antitrust and data privacy laws that require reasonable and non discriminatory access by end consumers to self-maintained and self-hosted infra?

Someone is learning that modern capitalism actually is bad

I like the stock

"oh you dont' want to use Bedrock? You want to do something that might vaguely compete with our offering on our infra? Fck you, pay me"

Buy your own hardware while you still can.

i would kindly like to curse openai for making the rest of world worse for everyone else. boohoo your ai is so smart? make it smaller.

Imagine trying to build a business in this environment.

But weren't the AWS shills saying that AWS only reduces pricing? I was trusting them blindly! /s

it's last call to get a GPU while there is stock left, no hopes those will become any more affordable in future considering the ongoing DRAM crisis. I am going to get 5070 Ti to replace my 1080 Ti which did a great service over all these years, paid it's worth back two times over during monero mining days and still can play a lot of games on high settings - recently finished RDR2. I sometimes wonder if Nvidia will be still making GPUs useful for gamers in a year or two or just completely shift it's focus to AI accelerators regular people will have no use for. RIP affordable computing at home.

So rather than free market capitalism where AWS would have actually lowered prices after investing heavily in a new datacenter-focused line of Huawei GPUs, the US government took steps to stop capitalism and competition, gave NVIDIA lots of free money (USA! USA!) and now prices are going up.

Recall pictures of Bezos smiling at various Trump events.

Everything Trump admin has done so far with tech reduces competition and tries to pick winners. This price increase is just the beginning.