It is admittedly surprising to see that Intel (INTC +1.77%) is having such a tough time ramping its upcoming Broadwell processor into high volumes. It's clear that Intel's aggressive goals for pursuing significantly improved transistor performance, as well as a greater-than-normal density improvement, came at a price – the delay of its next generation PC processor. As a result of this, should you freak out?
It really depends on the competition
The "doomsday" scenario for Intel as a result of this delay goes a little something like this:
- Intel delays its 14-nanometer microprocessor lineup
- This allows the foundries such as Taiwan Semiconductor (TSM 1.04%) and Samsung (NASDAQOTH: SSNLF) to "catch up" to Intel
- Intel's biggest structural advantage -- a semiconductor technology leadership -- goes away
Now, this thesis is largely predicated on the following facts:
- TSMC's and Samsung's 16/14-nanometer processes are roughly comparable to Intel's 14-nanometer process
- TSMC's and Samsung's customers can roll out products based on these processes during 2015
If all of those hold true, then yes, Intel's manufacturing lead will have diminished significantly, and its biggest weapon in the fight against the ARM ecosystem essentially disappears.
That's a ton of "ifs"
According to TSMC's and Samsung's marketing departments, they will go into production on their 16/14 nanometer processes, respectively, early next year. What this means for actual products, however, is an entirely different ballgame. Further, Intel claims that it has a performance/density edge over the foundries' respective processes, though full disclosures at a technical conference such as IEDM should shed some more light on the matter.
That said, as investors, we can try to connect the dots. Intel has traditionally had a significant transistor performance lead at any given time. Further, it has adopted next generation technologies, such as Strained Silicon, High K/Metal Gate, and FinFET well before the foundries have. Does it really stand to reason that, after years of pretty sustained leadership, Intel would all of a sudden see its lead evaporate in just one generation?
It seems more likely that Intel simply got too aggressive with trying to push the boundaries of its 14-nanometer process, and that aggressiveness blew up in the company's face -- forcing the product delays.
Should you freak out about the Broadwell delay?
It's disappointing that Intel is late with Broadwell, particularly after having promised production in 2013, only to retract that claim a month afterward. Further, though Intel's CEO indicated on the most recent conference call that the first Broadwell parts had gone into production during the first quarter of 2014, it's a bit alarming to see leaked internal documents hit the web claiming that the first parts don't go into production until July, with other parts following on much later.
Source: VR-Zone China.
It will be important to pay attention to the upcoming Intel earnings call, as at least one sell-side analyst is likely to ask about this delay if Intel doesn't volunteer an update. Listen to answers to the following questions:
- Will the 14-nanometer delay impact the previously disclosed launch timings of Cherry Trail (14-nanometer Atom for tablets) and Broxton (high end 14-nanometer Atom for premium phones/tablets)?
- Will this delay of Broadwell impact the launch timing for the next generation product, codenamed Skylake for laptops? (Leaked roadmaps suggest Intel is skipping Broadwell in desktops and moving to Skylake directly in Q2 2015. If so, does that push out the 10-nanometer follow-on known as Cannonlake?
If Intel assures investors that the 14-nanometer Atom product line (particularly Broxton) is still on track to what was presented at its Investor Meeting back in 2013, then Intel's competitive position will still be extremely strong. If not, then the Intel thesis gets a bit tougher.
Foolish bottom line
It's not yet time to freak out about the delay of Broadwell. Wait until the July 15 conference call, where Intel's management team will likely update investors on what's going on here. Delays are never fun, and it's easy to get caught up into thinking that Intel failed to execute while the rest of the industry will march ahead unabated.
However, the reality is that if 14-nanometer was hard for Intel, it's probably going to be even harder for the foundries and their fabless customers.
