Mobileye, Intel’s automated driving division, filed Friday for what is expected to be the year’s largest IPO, but its success is far from guaranteed.
The Israeli company, acquired by Intel five years ago for $15.3 billion, touts a broad vision: An autonomous future “where congestion is seen only in history books.” But its S-1 filing with the U.S. Securities and Exchange Commission underscores its precarious position in the ever-evolving self-driving vehicle industry.
Founded in 1999, Mobileye has benefited from its first-mover advantage, supplying automakers with computer vision technology to power their advanced driver assistance systems (ADAS). Now, as Mobileye expands its business model, it faces a proliferating number of rivals — from every side — in the wild and woolly world of automated vehicle technology.
The company’s list of competitors in its S-1 extends beyond the “Tier 1” suppliers in its corebusiness to now include robotaxi developers like Argo AI, Aurora, Auto X, Baidu, Cruise, Momenta, Motional, Waymo and Zoox, as well as what it describes as “consumer AV” competitors Apple, Sony and former customer Tesla.
TechCrunch pored through the S-1 to identify the speed bumps and bright spots in its pursuit to dominate autonomous driving.
Vertical integration
In the filing, Mobileye warned that its historical reliance on a handful of automaker partners may jeopardize future revenue. For the first six months of the year, Mobileye reported that 76% of its revenue was derived from eight automakers. But now big spenders such as General Motors and Mercedes-Benz are starting to develop their own autonomous driving systems in-house.
In pursuit of faster and more efficient AI system development, Intel, Arm and Nvidia today published a draft specification for what they refer to as a common interchange format for AI. While voluntary, the proposed “8-bit floating point (FP8)” standard, they say, has the potential to accelerate AI development by optimizing hardware memory usage and work for both AI training (i.e., engineering AI systems) and inference (running the systems).
When developing an AI system, data scientists are faced with key engineering choices beyond simply collecting data to train the system. One is selecting a format to represent the weights of the system — weights being the factors learned from the training data that influence the system’s predictions. Weights are what enable a system like GPT-3 to generate whole paragraphs from a sentence-long prompt, for example, or DALL-E 2 to create photorealistic portraits from a caption.
Common formats include half-precision floating point, or FP16, which uses 16 bits to represent the weights of the system, and single precision (FP32), which uses 32 bits. Half-precision and lower reduce the amount of memory required to train and run an AI system while speeding up computations and even reducing bandwidth and power usage. But they sacrifice some accuracy to achieve those gains; after all, 16 bits is less to work with than 32.
Many in the industry — including Intel, Arm and Nvidia — are coalescing around FP8 (8 bits) as the sweet spot, however. In a blog post, Nvidia director of product marketing Shar Narasimhan notes that the aforementioned proposed format, which is FP8, shows “comparable accuracy” to 16-bit precisions across use cases including computer vision and image-generating systems while delivering “significant” speedups.
Nvidia, Arm and Intel say they’re making their FP8 format license-free, in an open format. A white paper describes it in more detail; Narasimhan says that the specs will be submitted to the IEEE, the professional organization that maintains standards across a number of technical domains, for consideration at a later date.
“We believe that having a common interchange format will enable rapid advancements and the interoperability of both hardware and software platforms to advance computing,” Narasimhan.
The trio isn’t pushing for parity out of the goodness of their hearts, necessarily. Nvidia’s GH100 Hopper architecture natively implements FP8, as does Intel’s Gaudi2 AI training chipset.
But a common FP8 format would also benefit rivals like SambaNova, AMD, Groq, IBM, Graphcore and Cerebras — all of which have experimented with or adopted some form of FP8 for system development. In a blog post this July, Graphcore co-founder and CTO Simon Knowles wrote that the “advent of 8-bit floating point offers tremendous performance and efficiency benefits for AI compute,” asserting that it’s also “an opportunity” for the industry to settle on a “single, open standard” rather than ushering in a mix of competing formats.
Nvidia, the world’s largest maker of artificial intelligence chips, is at the heart of a new round of U.S. tech sanctions targeting China.
Nvidia noted in an SEC filing that the U.S. government had imposed new export restrictions on two of its most advanced AI chips to China, its second-largest market after Taiwan making up 26% of its revenues in 2021.
The ban could cost Nvidia as much as $400 million in potential sales to China in the third quarter, the firm said.
The export control also bars Nvidia from shipping the chips to Russia, though the company said it doesn’t currently sell to the country.
The U.S. government said the move “will address the risk that the covered products may be used in, or diverted to, a ‘military end use’ or ‘military end user’ in China and Russia.” But the ban is in practice crimping a wide array of businesses and organizations using the silicons beyond military uses.
The two chips in question are the Nvidia A100 and H100 graphic processing units. A100 is designed to provide high-performance computing, storage, and networking capabilities for industries spanning healthcare, finance, and manufacturing, explains Chinese e-commerce and cloud computing giant Alibaba, a user of A100.
H100 is the firm’s upcoming enterprise AI chip that is expected to ship by the end of this year and has part of its production done in China.
Nvidia’s engagement with China won’t be completely severed. The U.S. government has granted permission for Nvidia to keep manufacturing H100 in China, Nvidia said in another filing, though access by Chinese customers will still be restricted.
The ban is “sci-tech hegemony”, snapped China’s foreign ministry spokesperson Wang Wenbin in a regular press conference on Thursday. “The US seeks to use its technological prowess as an advantage to hobble and suppress the development of emerging markets and developing countries.”
The U.S.’s move to bar China’s access to its high-end technologies has in turn accelerated the latter’s pursuit of independence. Huawei has been doubling down on smartphone chip development ever since Washington put it on an export blacklist over national security concerns in 2019. A swathe of domestic semiconductor startups is netting hefty investments from VCs and government-guided funds.
While China may still be a generation behind in producing the most sophisticated chips, the country is gradually sharpening its edge in lower-end, specialized semiconductors, such as neural processing units that give phone cameras a boost. It remains to be seen what ripple effect the Nvidia ban will create.
After initial setbacks with its earlier iteration, the CHIPS and Science Act of 2022 sailed through bothchambers of Congress, early this week. The bill, which held uncharacteristically broad bipartisan support, hit some unexpected roadblocks in its last 24 hours in Congress, but still managed a wide margin of victory at 64-33 in the Senate and 243-187 in the House. The bill was signed into law during an event on the White House’s South Lawn.
The President was joined by Speaker of the House Nancy Pelosi, Senate Majority Leader Chuck Schumer, Gina Raimondo and SparkCharge founder/CEO Joshua Aviv, who spoke about the act’s impact on his Syracuse-based EV startup. “For years, my industry has been at the mercy of the supply chain,” Aviv said, before explaining the company manufacturers its parts in Buffalo, NY. “This new law gives people like me a chance and allows us to grow our businesses.”
Biden noted,“Today is a day for builders. Today America is delivering,” before adding that the bill is “a once in a generation investment in America itself.” He went on to note that the country has gone from producing ~40% of semiconductors to less than 10%, due to rampant manufacturing outsources.
The final version of the act sets aside $52.7 billion for the research, development and domestic manufacturing of semiconductors. $39 billion of that goes toward incentivizing manufacturers, with $2 billion being used to create existing/legacy chips for automotive and defense — a shortage of the former has hamstrung carmakers unable to move unfinished vehicles. Another $13.2 billion will be invested in R&D and workforce development.
Last week, the president joined Michigan Governor Gretchen Whitmer virtually — canceling a planned trip to Hemlock township — home of Corning-owned Hemlock Semiconductor — after once again testing positive for COVID-19 over the weekend. Following remarks from Michigan Senators Debbie Stabenow and Gary Peters celebrating hopes for a return to domestic manufacturing, Whitmer signed an executive order focused on the act.
“It’s important for us to make things in America,” the governor said, before signing an executive directive. “We’re making a once in a century investment in industry and bringing the supply chain for China to Michigan.” The choice of location was clear, given Michigan has come to represent both the rise — and decline — of U.S.-based manufacturing as an epicenter for the auto industry. It also highlights how essential the semiconductor has become, across industries.
The president clearly sees the CHIPS Act as an important piece in a broader plan to bring down the rapid inflation that has contributed to his low job approval, explaining that the rising price of cars has been a major contributor. He also shook off accusations that he lacked proper fiscal oversight. “[T]he bill is not handing out a blank check to companies,” Biden said. “This bill has guardrails that are going to protect taxpayers’ dollars and the interests of the American workers, small businesses, and the communities they’re in. It means companies partnering with community colleges, and technical schools that offer training and apprenticeship programs, and working with small and minority-owned businesses.”
The bill requires recipients to demonstrate significant worker and community investments, including opportunities for small businesses and disadvantaged communities, ensuring semiconductor incentives support equitable economic growth and development.
These funds also come with strong guardrails, ensuring that recipients do not build certain facilities in China and other countries of concern, and preventing companies from using taxpayer funds for stock buybacks and shareholder dividends. It will also support good-paying, union construction jobs by requiring Davis-Bacon prevailing wage rates for facilities built with CHIPS funding.
The president has counted himself among its biggest supporters in the final few weeks, sensing a much-needed win amid inflation and other strong headwinds. He noted in a statement following its passage:
The CHIPS and Science Act is exactly what we need to be doing to grow our economy right now. By making more semiconductors in the United States, this bill will increase domestic manufacturing and lower costs for families. And, it will strengthen our national security by making us less dependent on foreign sources of semiconductors. This bill includes important guardrails to ensure that companies receiving tax payer dollars invest in America and that union workers are building new manufacturing plants across the country.
CHIPS received large bipartisan support, amid an ongoing supply chain/chip shortage, fueled by pandemic-related shutdowns, climate crises and global conflict. Others, including chipmakers like Intel, championed the act as way to bring manufacturing jobs back to the U.S., during a time when chip fabrication occurs in Asia — and Taiwan specifically.
The legislation hasn’t been without its detractors. Senator Bernie Sanders was the only senator who caucuses with the Democrats who voted against it. “In my view, it’s not good enough to tell us that you worry about the deficit when it comes to addressing the needs of working families or the elderly or the children or the sick and the poor,” the Vermont senator said in prepared remarks. “You might want to worry about the deficit when you’re giving huge amounts of corporate welfare to very large and profitable corporations.”
Utah Senator Mike Lee concurred, stating, “The poorer you are, the more you suffer. Even people well-entrenched in the middle class get gouged considerably. Why we would want to take money away from them and give it to the wealthy is beyond my ability to fathom.” Opposition grew among House Republicans, following news of a deal between Senate Democrats and Joe Manchin. More recent criticism has emerged following the bill’s passage in both congressional bodies, including precisely how the moment will be spent.
U.S. chip makers have, obviously, been quite bullish about the influx of funding designed to jumpstart U.S. fabs. Intel initially delayed groundbreaking on its $20 billion chip plant outside of Columbus, Ohio. Critics called the move a “stunt,” aimed at pressuring the legislative branch into passing CHIPS. CEO Pat Gelsinger pushed back on the suggestion that the industry was seeking handouts, noting, “The rest of the world is moving rapidly despite the inability of Congress to get this finished.”
In a briefing issued this morning, the White House noted a pair of initiatives being announced by Boise-based semiconductor firm Micron and San Diego mobile chip powerhouse, Qualcomm,
Micron is announcing a $40 billion investment in memory chip manufacturing, critical for computers and electronic devices, which will create up to 40,000 new jobs in construction and manufacturing. This investment alone will bring the U.S. market share of memory chip production from less than 2 percent to up to 10 percent over the next decade.
Qualcomm and GlobalFoundries are announcing a new partnership that includes $4.2 billion to manufacture chips in an expansion of GlobalFoundries’ upstate New York facility. Qualcomm, the leading fabless semiconductor company in the world, announced plans to increase semiconductor production in the U.S. by up to 50 percent over the next five years.
The U.S. Senate earlier this week approved the CHIPS Act, which includes $52 billion to subsidize domestic semiconductor production. It still has to struggle its way through the bureaucracy (here’s a quick refresher), but clearing the Senate is a huge and important step toward the American chip fabrication industry getting a serious chunk of cash.
Personally, I’m psyched that this may be happening for a few reasons. Yes, yes — supply chain problems and chip shortages have been the bane of everyone’s life for a hot minute, and onshoring some of these fabrication materials, tools, and know-how will go a long way toward making the U.S. less reliant on external manufacturing, and more resilient in general.
That’s all good and well, but let’s be honest: $52 billion isn’t exactly a sachet of coppers and nickels, but chip manufacturing is expensive. The last planned chip factories I can remember are the $19 billion plant Intel is building in Germany and the $20 billion plant the company is building here in the U.S. If that’s the price tag of a factory, the subsidy builds two and a half plants. That means jobs, but it doesn’t exactly turn the U.S. into a chip-fab juggernaut overnight.
Far more than new factories, I’m most excited about the possibility of history repeating itself. Intel’s choice to build a $20 billion chip fabrication facility in Columbus, Ohio, along with the more recent news of the potential cash injection into the industry, could set the stage for a startup ecosystem boost.
Following a strong showing of bipartisan support in the Senate, the CHIPS Act staved off last minute attempts by the GOP House leadership to derail its passage. Congress’ lower chamber approved the $280 billion bill by a vote of 243-to-187. Twenty-four House Republicans backed the act, which is destined for the desk of staunch proponent, Joe Biden.
While the bill has received broader support for its efforts to kickstart domestic semiconductor production, a deal between Senator Joe Machin and the Democrats on an unrelated bill seemingly put it in jeopardy over the past 24 hours.
GOP Congressman Frank Lucas expressed the sentiments of many in his party in a floor speech earlier today, noting,
House Republicans have been working in good faith this entire time to come to consensus legislation that can be passed by both chambers. But time and time again we’ve been thwarted as Democratic leadership has moved the goalposts, shut down the process, and chosen their divisive, partisan policies over a smart bipartisan bill that would benefit our country for generations. I understand why people on my side of the room are furious. I share those concerns. And I’ve been around here long enough to know that this is not the way to do things.
For better or for worse – and it’s very clearly for the worse – the CHIPS and Science Act has been irrevocably tied to a massive tax hike and spending spree in reconciliation.
The Senate wrote a $280 billion blank check to be handed out to whomever Biden wants.
For the record, I'm a NO on the CHIPS bill. I was a NO last week, a NO yesterday, and a NO today. All Republicans should reject it.
On Twitter, House Republican leader Kevin McCarthy clarified that he was opposed to the bill from the outset, writing, “The Senate wrote a $280 billion blank check to be handed out to whomever Biden wants. For the record, I’m a NO on the CHIPS bill. I was a NO last week, a NO yesterday, and a NO today. All Republicans should reject it.”
The bill had previously come under scrutiny from a small but vocal number of congresspeople from both sides of the aisle, who accused it of supporting “corporate welfare,” for the billions of tax payer dollars that would go to chip production.
Biden has been one of the bill’s strongest supporters in recent weeks, touting what the $58 billion for chip manufacturing will mean for U.S. jobs and security in an industry that have been sidelined by myriad global events. In recent decades, a majority of semiconductor fabrication has occurred in Asia — specifically Taiwan. Such geographic concentration has made the industry far more susceptible to external forces — of which there have been plenty of late, from Covid-related shutdowns to weather to global conflicts, both cold (US/China) and hot (Russia/Ukraine).
Congratulations to the House and Senate for voting to approve #CHIPSAct funding. This investment will shape the future of America’s leadership in semiconductor manufacturing and innovation. We are excited to move full speed ahead to start building #IntelOhio!
Intel, in particular, stands to gain from the bill’s passage. The component maker had previously put groundbreaking on hold for its $20 billion Ohio fabrication facility in a move many labeled a “stunt,” after earlier versions of the bill failed to gain congressional approval. “Congratulations to the House and Senate for voting to approve CHIPS Act funding,” CEO Pat Gelsinger wrote following the vote. “This investment will shape the future of America’s leadership in semiconductor manufacturing and innovation. We are excited to move full speed ahead to start building.”
Biden reacted to the bill’s passage in real-time, during an economic roundtable with a group of CEOS. The President basked in the win, following the vote, noting, “The CHIPS and Science Act is exactly what we need to be doing to grow our economy right now. By making more semiconductors in the United States, this bill will increase domestic manufacturing and lower costs for families. And, it will strengthen our national security by making us less dependent on foreign sources of semiconductors. This bill includes important guardrails to ensure that companies receiving tax payer dollars invest in America and that union workers are building new manufacturing plants across the country.”
The President added, “I look forward to signing this bill into law and continuing to grow our economy from the bottom up and middle out for working families all across the country.”
Patent trolling, critics say, is guided by one principle alone: money. Yet tackling it remains a complex task with many angles. Today, a consortium called LOT — set up to help improve how the tech world, and the IP industry at large, handle trolling — is launching a new front in its efforts. It’s formed a new group called Adapt, an acronym for Advancing Diversity Across Patent Teams, with a mission to identify DEI issues in the IP industry and build programs to address them.
There is a specific but potentially very effective thread that links the worlds of diversity and inclusion with that of intellectual property.
DEI is a strong theme in the world of tech, which has traditionally not been great at inclusivity and has been making efforts to set that on a better course. The legal world is equally problematic on that front. As a result, the world of patent litigation acts almost like a force multiplier: you typically need engineering (or other scientific or technical) degrees as well as law degrees to practice IP law, Microsoft assistant general counsel Judy Yee — who is involved in Adapt — said in an interview, resulting in a even smaller subsets out of already-small diversity pools.
Case in point, research from the American Bar Association from 2020 found that 22% of patent attorneys and agents are women, 6.5% are racially diverse, and just 1.7% are racially diverse women.
The benefits of improving those ratios are of course important just as a matter of doing what is right and equitable for all people, and giving not just more opportunities but more people the knowledge and empowerment to realize those opportunities. As with the tech industry’s wish to become more inclusive, it’s about putting people in places to make decisions and build things for audiences that are diverse, too. Representing that at the point of service building and provision is critical in making sure that products are fit for that purpose.
Adapt believes that the same goes for the intellectual property industry.
“The reality is that with inclusive innovation, when we pull people in from marginalized communities, we create the best ideas and values and products,” said Micheal Binns, who is associate general counsel at Meta overseeing patent portfolio strategy for Meta’s Family of Apps (which refers to Facebook, Instagram, Messenger, and WhatsApp), and is also working on the Adapt project.
LOT’s interest in bringing this about comes partly from the fact that it’s one of the relatively rare instances of cooperation and collaboration between companies that might otherwise compete fiercely against each other — and might at some point find themselves in IP disputes of their own although these are considered separate from troll activities run by so-called patent enforcement agents, businesses set up with the primary purpose of amassing patents and then going after companies that they believe violate them in order to win settlements. That, plus the organization’s wish simply to do better in its own realm, gives it a unique place to advocate for programs that can have larger ramifications.
The group will be kicking off its first activities starting in September. They will include building a database of DEI programs and a directory for volunteer and sponsorship activities from DEI organizations: the idea here is to give companies keen to build more DEI initiatives access to outside resources to do so. It will also run a mentorship program aimed at underrepresented IP professionals specifically. Lastly, it will be building a platform and forum for measuring DEI analytics and sharing knowledge related to that.
Initially, Yee noted, Adapt will be focused on the legal aspect of IP — that is, organizations’ attorneys and wider teams working in patent law — but clearly there is an opportunity also to extend that to inventors, those building products and bringing them to legal teams to help secure IP claims.
“We decided for the first phase to focus on the patent profession, but a lot of our efforts bleed into that space,” she said of technical professionals, inventors and those whose aspirations might lead them into that area. “There is a pipeline.”
Patent trolls were a very regular presence in the landscape of technology a decade or so ago, and although you may not hear about patent cases every day nowadays, they are definitely not disappearing.
“Patent trolls are on the rise,” Ken Seddon, the CEO of LOT who previously worked for companies like Apple, Intel and ARM in patent litigation. He cites the currently tough economic climate, with inflation on the rise and the effects of the pandemic leading those who might hold IP assets to make more efforts to enforce them, or more often sell them off to PAEs (patent enforcement entities, the more euphemistic name for trolls), who are in turn making moves to enforce them against operating technology companies.
“The Supreme Court, White House and Congress have far bigger issues to deal with today, so my sense is that they may have given up on patent reform for the high tech industry.” But groups like LOT seem to paint a picture for why that may merit rethinking: it’s picked up 1,200 new members since February 2021, Seddon said, with companies including behemoths like TikTok owner ByteDance but also a lot of much smaller companies.
And that feeds well into the mission both at LOT and Adapt:
“We wanted to demystify that you need to be a large company to tackle diversity,” Binns said. And by that logic, the push for more inclusion will also play into helping fight IP abuse and misuse.
The U.S. Senate voted in favor of the chips bill (64-32) on Tuesday to bolster local semiconductor manufacturing. The bipartisan bill will pave the way for providing nearly $280 billion in various incentives for U.S-based chip making.
The bill, popularly known as the CHIPS-Plus package, has three major incentives for chipmakers: $52 billion of assistance in setting up fabs and manufacturing units, which also includes $2 billion for legacy chipmaking essential to the auto and defense industry; 25% in tax relief for investment in local semiconductor manufacturing at roughly $24 billion; and $200 billion grant for research in the area.
A version of this bill passed the Senate last year but was stalled in the House, and now it faces that final hurdle again before the August recess.
The Semiconductor Industry Association, a trade group representing chipmakers, with members like AMD, ARM, Intel, and Broadcom, welcomed the Senate’s vote.
“Today’s bipartisan vote is a vital step toward enactment of legislation that will strengthen American chip production and innovation, economic growth and job creation, and national security. We thank the bipartisan group of congressional champions for their leadership in advancing the CHIPS Act, applaud today’s Senate vote, and urge swift final passage in both the Senate and House. America has a historic opportunity to re-invigorate domestic chip manufacturing, design, and research, and Congress should seize it before it’s too late,” it said in a statement.
The U.S. has been pushing this bill and efforts to increase local semiconductor manufacturing after the pandemic and global chip shortage affected many industries including smartphones and the automotive sector. In an interview with CNBC on Tuesday, Senator Todd Young mentioned reliance on South Korea and Taiwan for critical chips and said that “we just can’t be that dependant on countries so far away.”
In the past few years, many companies have committed to setting up manufacturing in the U.S., including TSMC, which is building a $12 billion fab in Arizona. Last year, Samsung committed to a new manufacturing site in Texas with a $17 billion investment. South Korea’s SK Group, which held virtual talks with Presiden Joe Biden earlier this week, said it’ll invest $22 billion in the U.S. for development in sectors like semiconductors, green energy, and bioscience projects.
Intel’s on-going push to regain traction in chip manufacturing scored a nice vote of confidence this week, as Taiwan’s MediaTek agree to a “strategic partnership” with the company. The deal finds MediaTek utilizing the Intel Foundry Services process to produce future chips, as the company looks to broaden its global manufacturing footprint.
Intel has been pushing to increase its own chip production in the United States, including the upcoming opening of a massive facility outside of Columbus, Ohio. As of last year, more than 60% of the global foundry revenue was based in Taiwan alone. These past few years have demonstrated precisely why concentrating some much important production in a single geographic locale can be downright disastrous.
While it’s true that actions taken now will likely have little impact on the current chip shortage, it’s not difficult to imagine global crises becoming more norm that exception in the coming years.
Intel launched IFS last year in a bid to take advantage of that demand, and while the specifics of the deal have not been disclosed, a major partner like MediaTek – at very least – a sign that things are moving in the right direction for the new service.
“MediaTek has long adopted a multi-sourcing strategy. We have an existing 5G data card business partnership with Intel, and now extend our relationship to manufacturing smart edge devices through Intel Foundry Services,” MediaTek Senior Vice President NS Tsai said in a release. “With its commitment to major capacity expansions, IFS provides value to MediaTek as we seek to create a more diversified supply chain. We look forward to building a long-term partnership to serve the fast-growing demand for our products from customers across the globe.”
The deal could also help move the needle on the $52 billion chip manufacturing subsidies bill set for a final vote in the Senate this week.
It’s been a long time coming, but Google Cloud today announced its first Arm-based VMs, following AWS, with its Graviton instances, and Azure, which also recently launched Arm VMs. But while AWS built its own custom chips, Google Cloud is following Azure’s lead here by using chips from Ampere. These new VMs, which are now in preview, will join Google Cloud’s line of Tau VMs under the ‘Tau T2A’ moniker. This line launched almost exactly a year ago, using AMD Milan processors, to offer a better price/performance ratio.
“We are excited to extend the rich choices we already offer with Intel and AMD and enter the Arm ecosystem to provide our customers with even more choice and flexibility. We have support for a broad ecosystem of operating systems, databases, programming languages and other tools,” Sachin Gupta, Google Cloud’s VP and GM for infrastructure, said in a press briefing ahead of today’s announcement.
The new chips will come in pre-defined SKUs with up to 48 vCPUs, each with up to 4GB of memory. The VMs will offer up to 32 Gbps of networking bandwidth and support the usual range of storage options available in the Google Cloud ecosystem. Google says these CPU specs will make these machines useable for a wide range of workloads, including as web servers and for running containerized microservices, data-logging applications and more.
Like the AMD-powered Tau chips, Google sees these as its price-performance optimized solutions. A 32-core Tau T2A VM in Google Cloud’s us-central1 region will cost $1.232 per hour, for example.
Users will be able to use the likes of RHEL, CentOS, Ubuntu and Rocky Linux on these machines, in addition to Google’s own Container-Optimized OS for running containerized applications. At this point, Arm support has become table stakes for most OS and software vendors, which in turn also greatly enhances the usefulness of these VMs (and those of Google’s competitors).
The new VMs are now available in a small number of regions, including us-central (Iowa – Zone A, B, F), europe-west4 (Netherlands – Zone A, B, C) and asia-southeast1 (Singapore – Zone B, C), but will come to other data centers over time.
“Ampere Altra Cloud Native Processors were designed from the ground up to meet the demands of modern cloud applications,” said Jeff Wittich, Chief Product Officer, Ampere Computing. “Our close collaboration with Google Cloud has resulted in the launch of the new price-performance optimized Tau T2A instances, which enable demanding scale-out applications to be deployed rapidly and efficiently.”
In addition to using these VMs as part of Google Cloud’s Compute Engine, Google also now supports them as part of its Kubernetes Engine, the Dataflow stream and batch processing service and Batch, a new fully managed job scheduler for batch job Google is also launching today. “This new capability will benefit major use cases for throughput-oriented computing such as weather forecasting and electronic design automation,” said Gupta. “The primary purpose of this new service is to offer unprecedented flexibility in time, location and cost of cloud capacity for batch jobs.”
