Intel CEO Pat Gelsinger released a holiday video celebrating his first year with the company. The video mainly thanks the employees for a good year, but there is one particular remark that caught a lot of attention. “Lake Alder. All of a sudden boom! We’re back in the game,” Gelsinger said. Marlet.
That’s unlikely to be true, and I suspect Gelsinger knows it. I’m not saying this because I have any special or particular knowledge about how Zen 4 will compare to future Intel microarchitectures. That’s an unlikely statement based on how AMD and Intel have faced each other for the past 31 years. Alder Lake is a good CPU microarchitecture and we expect it to be very competitive on desktop, server and mobile. That doesn’t mean Intel has permanently put AMD in the rearview mirror.
It’s all happened before
Gelsinger is a longtime Intel veteran, having joined Intel in 1979. He’s been around long enough to see Intel and AMD rise and fall. Gelsinger earned his degree in electrical engineering while working on 80286. Andy Patrizio credits Gelsinger’s work on the 80386 is key to saving Intel after the i432APX failure. The up-and-coming engineer went on to lead the 80486 design team before moving into management in the Pentium Pro era.
Pat Gelsinger was a relatively high-ranking Intel employee back when Intel was doing everything in its power to kill off AMD and other x86 licensees. In 1991, Intel sued AMD in the Supreme Court to try to invalidate its x86 license. This successfully delayed the introduction of AMD’s 80386 for years, costing the company tens of millions of dollars. The bad blood between the two companies had been legendary for over a decade.
The K6 was the processor you wanted if you couldn’t afford an Intel chip, but needed floating point performance faster than a pocket calculator. Chips from Cyrix and IDT failed this particular test.
For most of the 1990s, AMD established itself as a second-class processor manufacturer. That didn’t stop him from launching the K7 in 1999 and catching Intel off guard. Intel came back with the Pentium 4 Northwood a few years later. AMD then took the lead again with the Athlon 64 in 2003 and cemented it with the Athlon 64 X2 in 2005. Just over a year later, Intel’s Core 2 Duo took over the performance crown , which Intel subsequently owned from 2006 to 2017.
All my life is a circle
If there was ever a moment to conclude that Intel had “definitely” put AMD in its rearview mirror, it was 2012 – early 2017. For five years, AMD racked up huge quarterly losses while Intel continued to iteratively improve its processors. As of March 1, 2017, Intel could claim to have dominated the processor market for approximately 21 of the past 26 years and the entire past decade. On March 1, 2017, the only company actually competing with Intel in the processor market was Intel. AMD had once again become the low-end budget alternative, with very few wins on high-end systems.
On March 2, 2017, AMD significantly caught up. The Ryzen 7 1800X delivered a 1.52x IPC improvement over Piledriver and made eight cores the high-end mainstream desktop standard. AMD followed the Ryzen 7 1800X with several processors which further improved its competitive position against Intel. When Intel retaliated with the 14nm++ and the 8700K, AMD retaliated with the 2700X and so on, until we get to today. In 30 years, neither processor manufacturer has managed to gain a permanent advantage over the other, despite Intel’s size advantage. For almost five years, AMD has continued to nibble away at Intel’s position.
AMD: The Cockroach and/or the Corporate Tardigrade*
All of the above is ancient history today and AMD facing Intel in 2022 is not the AMD it faced in 2012 or 2002. AMD is no longer drowning in debt. It doesn’t struggle with its own node transitions. It is not legally required to use a foundry partner who cannot ship state-of-the-art material. It’s not about the distraction of turning its manufacturing arm into an independent business or fixing a broken microarchitecture. It’s not racing to design a new processor before it runs out of money.
That’s not to say AMD won’t face challenges in 2022. Intel has set an aggressive launch schedule for the next few years. He said he intends to regain overall performance and process node leadership by 2025. Alder Lake is currently very competitive with Ryzen 5000.
Additionally, Intel and AMD face a renewed competitive challenge from ARM over the next few years, courtesy of Qualcomm, Apple, and various hyperscale server vendors. Intel’s new hybrid core strategy puts pressure on AMD to demonstrate that it can match or exceed Alder Lake performance without efficiency cores. AMD’s overall position in the GPU market is weaker to Nvidia than its position in the CPU market to Intel. As Intel plans to enter the GPU market, AMD may need to invest more aggressively in this space.
The fight between Intel and AMD widens, not over. In the future, AMD’s Xilinx will compete with Intel’s Altera in the FPGA market. Intel’s Alchemist GPUs will go up against GeForce and Radeon. The idea that Intel has definitely left AMD behind with Alder Lake doesn’t hold up to scrutiny.
*-Depending on what you think of AMD. Both are very hard to kill.
Objects in the rearview mirror are closer than they appear
AMD has literally never been in a better position to meet Intel’s challenge. The company has now been profitable every year since 2018. Although this may not be a record again, AMD was once infamous for its unreliable financial performance. Under the leadership of Lisa Su, this is no longer the case.
The quickest way for Intel to return pole position to AMD would be to assume that Alder Lake guarantees that outcome. Intel has never quite gotten rid of its main competitor, regardless of its advantage at the time. Gelsinger is probably aware of this. His message was probably more meant to inspire than to be evaluated by the press.
It is important to note that there is no sign of a short-term return to the old status quo. Both manufacturers are announcing very healthy quarterly results thanks to the PC boom. This should continue until 2022-2023. In the past, Intel’s better results have often had negative impacts on AMD. In an environment where both companies sell all the chips they can make, this is less likely to happen.
I don’t know if Alder Lake Mobile or Ryzen 6000 Mobile will prove to be the best mobile architecture, but whatever happens, it won’t be the last word. What we actually have for the first time in at least 20 years is of them financially stable and sound processor design companies that get what they pay for.
Now read:
- Intel 12th-Gen Alder Lake: News, Rumors, Release Date, More
- Intel CEO: Alder Lake left AMD in the rearview mirror
- RISC vs. CISC Is the Wrong Lens for Comparing Modern x86, ARM CPUs
- 12th Generation Intel Alder Lake processors now available with power to rival Apple’s M1 Max
- Intel 12th-Gen Core Alder Lake Architectural Benchmark
