Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Intel returns to profitability after two quarters of losses (cnbc.com)
146 points by bluedino on July 27, 2023 | hide | past | favorite | 60 comments


Is this an accounting trick to look profitable or is it normal and reasonable?

> Effective January 2023, Intel increased the estimated useful life of certain production machinery and equipment from five years to eight years. When compared to the estimated useful life in place as of the end of 2022, Intel expects total depreciation expense in 2023 to be reduced by $4.2 billion. Intel expects this change will result in an approximately $2.5 billion increase to gross margin, a $400 million decrease in R&D expenses and a $1.3 billion decrease in ending inventory values.


>Is this an accounting trick to look profitable or is it normal and reasonable?

It seems normal and reasonable to me with their change in direction to making chips for others. Previously Intel (for the most part) needed to rapidly depreciate hardware and move to the next process node to stay competitive. Now they can continue running a process node far beyond what they would have historically to create a cheaper chip-line for third parties.


They have been talking about being a foundary for external use for a while, years... yet i have not heard of a single customer yet, just an announcement today that intel and ericsson have a partnership.

It feels like yes they want to become a external foundary, but they just havent fully committed to that task yet.


>They have been talking about being a foundary for external use for a while, years

Correct, Gelsinger announced roughly a month after returning to Intel their new strategy (~2 years ago). Believe it or not, pulling apart the financials of a company the size of Intel can't be done overnight - and I would guess legally they needed some customer contracts built on the old process technology before they could depreciate assets.

>yet i have not heard of a single customer yet, just an announcement today that intel and ericsson have a partnership.

Did you spend any time looking? 30 seconds on google shows Mediatek signed up last year:

https://www.forbes.com/sites/patrickmoorhead/2022/08/10/inte...


Yep, i feel like i am just a bit skeptical. I would have just liked to see multiple chips out in the wild (from different companies outside of intel) using the intel fabs.


This smells of mark to market accounting aka Enron book cooking


Arm has announced a partnership with Intel Foundry Services: https://www.intel.com/content/www/us/en/newsroom/news/intel-...


ARM isn't a customer though. It's IP that Intel's potential customers can use.


IBM also.


The A18 node will be the first node that IFS will deliver to external customers. That node was recently pulled in and is scheduled now for H2 2024. What you heard so far, like from Ericsson, are design wins. Presumably Intel is working with many more customers that we don’t know about yet.

Disclaimer: long Intel, recently increased my position.


The CEO allegedly wants them to be an external foundry. His direct reports may not be as enthusiastic.


The article said that foundry did about a quarter of a billion dollars in revenue


That's what, ~5k wafers a month?


Well they only announced that they would be doing this two years ago. You can’t set that up overnight, when your fabs are intimately intertwined with the rest of your operations and you need both to separate proprietary IP and document processes (and set up the ability to work closely with external customers).

A financial quarter of $250MM a couple of years after that is pretty fast.


They had been doing this before that point. Altera had been on Intel's foundry service since 2013.

5k wafers after ten years of foundry service is pretty abysmal.


It is both normal and reasonable and an accounting trick. The entire concept of depreciation is an accounting trick. Albeit a useful trick if you ask me.

Look at cash flows if you don't like accounting tricks.


I came to a conclusion that most accounting trickery is just for tax purposes.


How much of this is due to salary cuts, benefits/bonus cuts, and layoffs? I live in an area with a lot of Intel campuses and I know a ton of people that were laid off, had their salary cut, no bonuses, etc.


David Zinsner, Intel CFO, said, “Strong execution, including progress towards our $3 billion in cost savings in 2023, contributed to the upside in the quarter. We remain focused on operational efficiencies and our Smart Capital strategy to support sustainable growth and financial discipline as we improve our margins and cash generation and drive shareholder value.”

From: https://www.intc.com/news-events/press-releases/detail/1637/...


fwiw Google did the same trick with their servers


Likewise Nvidia...buried in supplemental CFO commentary[1] circa FY23 Q4:

> Starting in fiscal 2024, we are extending the useful lives of a majority of our servers, storage, and network equipment from three years to a range of four to five years, and assembly and test equipment from five to seven years. As a result, we expect favorable impact to operating income of approximately $26 million in the first quarter and $133 million in fiscal 2024 from a reduction in depreciation expense based on equipment balances as of January 29, 2023. The favorable impact will be primarily in operating expense, and to a lesser extent in cost of goods sold.

[1] https://www.sec.gov/Archives/edgar/data/1045810/000104581023...


Could be a benefit from being stuck on the same node for longer


Interesting that this is allowed under GAAP.


If they’re lying about the life of these machines, that will come back to bite them later on when their forecasted expenses don’t match the reality.


To make profits look better. It's why analysts tend to build Thier own p&l from the accounts and focus on the free cash flow instead.


I've always been a fan of AMD, but I must admit, Intel is making some impressive strides in the Machine Learning field, especially with tools like OpenVINO. The process of converting models to ONNX and running them in an accelerated manner on Intel GPUs has become significantly easier compared to using ROCm.

If Intel continues on this trajectory, it wouldn't be a shocker to see them leapfrog AMD and give NVIDIA a run for their money.


In the last year we made a 180 from growth at all costs to profit at all costs. Would be crazy if the market just valued consistent sustainable growth but im sure some economist would tell me this is stupid


Honestly when I see the extremely high power consumption of Nvidia cards, I feel there are some inefficiencies and consequently an opportunity for Intel.

It remains to be seen if Intel can deliver reasonably affordable GPUs with enough hardware efficiency gains while not being too far behind Nvidia on the software tooling front.

But they have a card to play here.


For SYCL and DPC++ to give a NVIDIA a run for their money, they need to also at very least be Fortran and Python friendly (let alone all the other languages that target PTX), have a good graphical GPU debugger like Nsight and Visual Profiler, and tons of helper libraries for compute.


what does openvino run on that is accessible to normal developers? amd has GPU cards for me to use and mi250 mi300 for data centers Nvidia has more of those.

I'd like to have hardware to play with oneapi,openvino,but do not know where the hardware is


Any Intel Arc GPU will run it, as far as I know. I bought an A750 for myself to experiment with Level Zero/SPIR-V and SYCL.


https://www.intel.com/content/www/us/en/developer/articles/s...

here is the hw requirement for oneapi

and my 12 years old core-i7 seems still can play with OneAPI, installing it right now.


Does the A750 work? Ive had a couple of attempts to get their GitHub projects running on an integrated chip without any success


thanks,will buy one


It'll use the built in GPUs on any newer intel CPU. Otherwise, Arc.


I wish they'd let me buy an Arc A310.


Does it have to be A310? There are a couple models of A380 available at Newegg at similar prices to what the A310 was supposedly selling for in other regions.


I think Intel’s return to success is pretty much inevitable at this point. Just look at how much money they are getting in subsidies from countries like the US and Germany. This is coming from a strong geopolitical imperative to bring advanced semiconductor manufacturing back to the West.


All semi manufacturers are getting subsidies.. tsmc and samsung have contributions measured in percentage of gdp of the respective countries. Its hard to believe they are not getting any


Sure, I’m not saying that they won’t be successful too.

But Intel is the only Western company capable of manufacturing the most advanced semiconductors right now, and you better believe that makes them too big to fail.


Finally the government is doing something good; paying chip manufacturers so I can get paid a million dollars a year to use O(n^2) algorithms everywhere!


> This is coming from a strong geopolitical imperative to bring advanced semiconductor manufacturing back to the West.

But didn't they also offshore their CPU design teams?


No; in fact, we're doing more of our core development in the US than at some previous points in our history.

(I work at Intel, but this is just my personal observation.)


Does intel still have the really toxic contractor culture with a TON of h1b's? I don't see them rebounding until they get quality employees and not pay severely under market through contractors.


I work remotely on a small-ish research team. Almost all of my immediate colleagues are in the US or Western Europe. Most (all?) of the folks on P- and E-Core Engineering that I work with are full employees (i.e. not contractors) in the US or Israel. I'm not sure I'd know if they were H1Bs, but they're generally very knowledgeable and hardworking. From my perspective, there's little outsourcing on the core businesses (pun fully intended).

As far as comp goes, mine is competitive with Bay Area FAANG.


No? Please cite a source for this.


By cost cutting measures, replacing the unrealistic dividends phase.. I have no idea why a ex-monopoly taking every action it can think of for accelerating its descent in market share would be seen positively for getting to barely profitable by steps that are no doubt sacrificing future sales.


Taking devil's advocate position here:

1) dividend was not (any longer?) rational, since they needed to reinvest in production process improvements; this is therefore a good CEO decision, which gives confidence to stock owners that CEO can do what needs to be done

2) "five nodes in four years", i.e. acceleration to catch up in process capability, is said to be on track, which (if actually true) is good news

3) this was the "pain for future gain" part of Gelsinger's turnaround plan, so if they can actually make a profit even during this phase, that is good news

Not sure if I believe any of this, but that's my take on why it might be reason to buy the stock. Not that I am, in fact, actually buying the stock.


The market is weird.

It might have jumped because the market was pricing in a faster decline and it beat expectations, causing a correction.

It could also be because earnings reports are times of higher volatility for a stock, and people will use derivatives to make certain bets around it. And then the earnings results can cause them to take decisive action to exit those positions, which might not be happening in the spot market but the spot market can feel the ripple effects of big movements in derivatives.

Or maybe a bunch of meme investors are buying intel because it had good news and there's nothing more sophisticated than that.

The market is weird, and it's especially weird on short time scales. Let's wait a week and see where they land.


Because much of investor behavior is (arguably rationally) not driven by business fundamentals, but how they assume other investors will act. This can lead to counterintuitive and self-fulfilling group think behavior where the metrics might drive the stock price because of an incorrect perception that others value that metric, regardless of actual sentiment.


They've been promoting MBAs to leadership for years, it lead to the down turn and apparently they've figured out a new financial gimmick.


> Investors want to see gross margins expand even as the company invests heavily in manufacturing capability.

Intel should tell such investors to fuck off and let those who believe in intel reap the rewards.

I’m not bullish on Intel myself though I hope they do get their act together.


So every sale went down? And the stock still went up because profits went up?


Likely the lower sales were already priced into the stock price, and it was the stronger-than-expected forecast that caused the stock bump.


CEO is being thrify enough that he's wearing vmware swag


Imagine being down 15% YoY and STILL making 12.9 billion


Intel made money this quarter, but only because of a $2 billion dollar tax rebate. Without that they're still losing money.


I refuse to buy anything Intel until they reach 7nm.


Intel 7 has been shipping since 2021.


Intel 7 is on 10nm.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: