Thread regarding Qualcomm Inc. layoffs

To be fair with the management team, this is one area they had a good vision. Just series of bad choices of leadership. They gave RTP so many chances and they keep fumble on it. RTP incompetence galore.

Also, last time I checked... MSFT and NVDA don’t make high performance CPUs.

So... Intel, AMD, ARM and Apple must have really lost out!

Their “loss”... LOL.

Benchmarks under NDA!

Hahahaha...

If they were good, Q would be shouting them from the rooftops. What response would Intel have against good performance??? Are you implying that a good score would make them make better processors (as opposed to anything they have now)? Do you hear yourself? It doesn’t even make sense.

Reality - the performance was sh1t and the project got canned.

4hjq- fact is the Spec benchmarks were only released to customer's under NDA, we didn't want Intel to know for competitive reasons. There were a number of benchmarks released by a third party if you search for them. BTW, you must be some kind of loser to monitor Q layoff threads. Before you ask, I'm waiting to start a new job at MS where a rather large portion of the CPU team landed. Nvidia picked up almost 70 of the Asics team. Yeah, we're a bunch of losers. Well it's Q's loss.

What color is the sky in the Raleigh distortion field?

Still waiting on those submitted SPEC benchmark scores at www.spec.org

4huo - you obviously are not an insider from the team or a customer otherwise you would know the designs were excellent as was the team. You are just a few ntetested in spreading hate. Go crawl back into your hole.

In the case of QDT, potential buyers were all going to want several months to do due diligence. I certainly would if I were dealing with SM. That timeframe wasn't consistent with the short term boost to earnings that was needed due to the panicked promises made during AVGO's bid. By closing down QDT, an immediate cut in opex lifted quarterly earnings and the shutdown costs were financially quarantined as "one off expenses". Anyone who can read a balance sheet knows how this works. That's still a real loss for shareholders though.

@3aui

If you believe that , I have a bridge to sell you.

Nobody turns down free money - not even GD or SM.

Anyone with a modicum of critical thinking skills can infer the real story.

By the way - where are those submitted SPEC benchmark scores? (crickets)

@-2yon there were quite a few potential buyers, however SM and GD insisted on taking qdt "off the book" before the end of fiscal year, that left little or no time to negotiate a term and close the deal with any buyers. The whole process was totally self destructive, what a great leadership!

@UvYgD8W-2imp meant to @UvYgD8W-1png.

@UvYgD8W-1jjt and clearly either executive management didn't plan for this, or the chip was such a dud that no amount of optimization support from Q could save it. Clearly no one in this thread knows which it is.

@1png

If it was such a great business, why were no bids accepted when Q put the team up for sale?

Where are the submitted SPEC scores?

Things that make you go hmmmmm.

The Raleigh distortion field is alive and well even when burning in failure.

Let's be real, bottom line is Q execs didn't want to wait for profitability on this new business. They needed to save $1B to appease shareholders and enrich their own compensation.

1fng-id--t you don't go from 0 to 4M overnight. There is software and other optimization being done for conversion.

1fng - try again Intel data center gross margins are more like 80%, you know nothing.

Those orders were just a start of the initial product. New business take time to ramp to profitability. Those saying that the designs were sh-t don't know what they are talking about. The designs were better than Intel on perf and power.

Facebook has 4 million servers.

20K is a piss in the ocean.

Let’s say 60K servers as you suggested at a $2000 ASP. This brings in $72M at 60% gross margin ( these are Intel margins) so right off the bat, I’m being very generous.

Gross margin does not count R&D, so that has to be paid. You have 800 QDT engineers burning 150K per Head per year ( this includes overheads of the building rent, benefits etc ) so that alone burns 120M per year - already losing money.

But of course, employees are not the only R&D costs. You need to pay for the EDA tools, the computers, the travel, the utilities and on and on...

Rumors were that the burn rate was closer to $500M. There is no way 60K servers were going to cut it.

Like someone else said - you needed more zeros there.

@1jjt

The “too expensive” quote was from @ohv

You might want to re-read it you RTP dumb f--k.

AW was sh!t.

If it was great, someone would have bought it to fund its development or bought the team that made it.

I know SM and the execs have made bad decisions, but throwing out a high margin, profitable business that hurts Intel doesn’t make sense in any universe other than the one where AW and follow-ons were un-economic (AKA total sh!t).

Anyways, enjoy trying to find a job at a competitive processor company. I’m sure the small number of good engineers already have offers but the bulk of RTP is just subpar.

@UvYgD8W-1jjt If you're trying to argue 60K units is enough to prove a solid product, that number is missing more than a few zeroes.

It could still be the case of a decent chip stabbed in the back by id--tic decisions in the sales and software orgs. Intel has been on the other side of this again and again. Not surprised if efforts at Q fare similarly now.

@qdt - You mean the 20k in orders from Facebook? Or the 40k from cloudflare. This had nothing to do with the technology. As a person tasked to measure performance of AW vs Skylake, you are full of c-ap and don’t know what you are taking about. AW owned TCO. Period.

Too expensive - wtf are you talking about. The money had already been spent. It was just marketing now until SM and CA came up with this awesome strategy change.

A classic case of financial engineering for super short term gains. What a disaster.

Yes - top notch team. Now at Nvidia, Intel, Microsoft and Ampere. Smartest most decent engineers I ever worked with. Seriously, that talent will end up eating Q’s lunch.

I am so tired of id--ts who don’t know what the f--- they are talking about. STFU unless you’ve done your homework.

@ohv

Do you hear yourself?

It was too expensive?

What kind of logic is that?

If there were orders on the horizon to pay for it, then why cancel it?

Here’s the real answer.

Nobody wanted it in enough volumes to be economically viable.

That’s the truth.

In other words - it svcked.

uir- maybe you are not in the know or maybe you are just a complete a--hole. Both the Amberwing and Firetail (next gen) were better than competitive. Facebook upped their orders significantly. The reason the program was cancelled was because it was too expensive for the shortsighted execs and deemed too much of a long shot. In other words the execs were scared of Intel in this space. Go back to designing your non-competitive ASICS

I agree in full with the OP.

I heard they "borrowed" the original design from a private equity company.

“Top notch team”?

What are you smoking?

Their processor wasn’t competitive so it got cancelled. It’s really not complicated.

Cool story bro