White House proposes up to $8.5B to fund Intel’s domestic chip manufacturing

Well before President Joe Biden signed the CHIPS and Science Act into law back in August 2022, Intel has been a cornerstone of U.S. efforts to increase domestic chip manufacturing. This morning, the White House announced an agreement with the Department of Commerce that would deliver the silicon giant up to $8.5 billion to shore […]

© 2024 TechCrunch. All rights reserved. For personal use only.

How much are Nvidia’s rivals investing in startups? We investigated

Over the past couple of years, Nvidia, by far the largest AI chipmaker, has ramped up its investments in startups that propel it deeper into the AI space. According to S&P Global and Crunchbase, the funding and investment database, Nvidia’s startup investments jumped 280% year-over-year from 2022 to 2023, with the company and its VC […]

© 2024 TechCrunch. All rights reserved. For personal use only.

Intel’s modernization strategy could face setback with end of $5.4B Tower deal

In 2021, Intel CEO Pat Gelsinger announced a comprehensive modernization strategy he dubbed IDM (integrated device manufacturing) 2.0. As part of that vision, the company announced a $20 billion investment to build two new Fabs (chip manufacturing facilities) in Arizona along with plans to increase capacity in other parts of the U.S. and Europe.

In a 2022 article, we described Intel’s new approach this way: IDM 2.0 involves a three-pronged approach to semiconductor manufacturing: Intel’s network of global factories, use of third-party capacity and building out Intel Foundry Services, moving the company beyond simply producing Intel-branded chips, but helping meet the growing needs for custom chips.

As part of that shift, Intel announced plans to acquire Tower Semiconductor, an Israeli chip manufacturer, for $5.4 billion in February 2022. That deal was killed yesterday when Intel announced it was walking away due to regulatory issues getting the deal approved, particularly in China.

At the time of the Tower acquisition announcement, Patrick Moorhead, lead analyst at Moor Insights & Strategies, said that the deal would have given Intel access to custom silicon it didn’t have — but the pitched political battle between the U.S. and China doomed the deal, and it’s going to be hard for Intel to replace that capacity.

“I am not shocked that China scuttled the Intel-Tower deal. The US and China have been in a tit-for-tat battle for the past 5 years on technology, and, unfortunately, both companies are in the crossfire. Intel now needs to determine how it will fulfill its end-to-end foundry vision,” Moorhead told TechCrunch.

Unfortunately, that vision is going to take a big hit with the end of the Tower deal, says Ray Wang, principal analyst and founder at Constellation Research. “Intel needs foundry capability, especially for 3nm future chips. They need to keep up with TSMC and others, and have fallen behind. Tower has great analog specialty chips that would have complemented Intel’s product line-up,” Wang said.

That means that fab production is going to be key, and Intel will have to focus more on newer chips and foundry capacity, he said.

Gartner analyst Raymond Paquet says that the failed deal is going to limit Intel’s modernization approach. “This is a setback for Intel and their foundry business. Tower would have provided Intel with more expertise and foundry customers,” he said. Acquiring these customers is going to be more difficult for Intel moving forward, and Intel will have to hire in-house expertise it would have gotten from Tower, Paquet said.

Let’s not forget this is not the first time a big chip deal has been canceled because of regulatory scrutiny. In February 2022, Nvidia announced that it was walking away from the deal to acquire Arm for $40 billion.

In 2021 we saw a flurry of chip company mergers including that failed Nvidia deal, as well several multi-billion acquisitions that survived the regulatory process. That included AMD buying Xilinx for $35 billion, SK Hynix nabbing Intel’s memory unit for $9 billion, and Analog Devices acquiring Maxim for $21 billion.

The ongoing political battles between China and the U.S., especially around semiconductors, could prevent those kinds of high profile deal from happening again anytime soon. “Given the circumstances of this deal not being completed, semiconductor M&A will be more difficult for everyone [moving forward],” Paquet said.

Intel and Tower Semiconductor cancel $5.4B merger over regulatory hurdles

Intel has called time on its plan to acquire contract chipmaker Tower Semiconductor, citing its inability “to obtain in a timely manner the regulatory approvals required under the merger agreement.”

Chip giant Intel first announced it was planning to buy the Israeli company for $5.4 billion way back in February last year, a move designed to bolster its own contract chip-making business with enhanced manufacturing capacity and intellectual property, while also giving it a wider global reach.

Indeed, Intel revealed plans to invest $20 billion in two new Arizona factories some two years ago, while also confirming a new offshoot called Intel Foundry Services (IFS) dedicated to manufacturing chips designed by other companies. It indicated a major expansion vertical for the company, one that Intel CEO Pat Gelsinger dubbed “IDM (integrated device manufacturing) 2.0.” This was essentially a multi-pronged approach to building semiconductor chips, spanning Intel’s own network of factories, third-party factories, and building out its fledgling foundry services.

Forging relationships with companies already deeply integrated in the foundry space would allow Intel to accelerate these plans. Tower Semiconductor had been manufacturing analog chips for hundreds of companies across the industrial spectrum for some two decades, making it an ideal acquisition target for Intel.

Obstacles

Although Intel hasn’t discussed any of the specifics around its regulatory obstacles — in China, or elsewhere — more than a year on from the original acquisition announcement, concerns started to mount that the deal could be in peril due to pushback in China. Indeed, Gelsinger made several personal visits to the country with a view toward building relationships with industry and government, but it seems that this was insufficient to get the deal over the line.

While it may have been technically possible to conclude the acquisition without China’s approval, China represents a major part of Intel’s business and strategy, meaning getting the greenlight from regulators there was essential.

As a result of all this, Intel said it will have to pay a termination fee of $353 million to Tower Semiconductor, whose shares have dropped more than 11% in light of this news.

“Our foundry efforts are critical to unlocking the full potential of IDM 2.0, and we continue to drive forward on all facets of our strategy,” Gelsinger said in a press release. “We are executing well on our roadmap to regain transistor performance and power performance leadership by 2025, building momentum with customers and the broader ecosystem and investing to deliver the geographically diverse and resilient manufacturing footprint the world needs.”

Intel slashes divdend by nearly two-thirds to shore up cash as chip giant braces for a tough year

In January, when Intel reported one of its worst financial quarters in years, the chip giant worked to keep up investor confidence by holding its dividend steady at $0.365/share. Less than a month later, it’s singing a very different tune. The company today announced that it was revising its dividend to $0.125 per share, down nearly two-thirds, as part of a bigger effort to conserve cash amid the very tough economic environment, and how it’s playing out specifically in the tech sector.

The dividend cut underscores the darker outlook Intel has for the year ahead. The company’s dividend has been level at $0.36 for many quarters and has dipped below $0.30 since 2017, and while dividends do not impact non-investors — can be used to keep investors happy even through rockier patches, such as a bad stock decline or disappointing earnings, and also simply to keep the stock at a premium overall: Intel’s paid out some $80 billion in dividends since 1992 — they are also a bellwether of the company’s bigger state.

Intel is explaining the cut in the context of bigger efforts at the company to cut up to $3 billion this year, and up to $10 billion per year by 2025 — which it will be doing by phasing out certain operations, laying off employees, reducing compensation from executives and making other cuts. It’s also taking a bigger bet on its own tech by building out its own internal foundry, which will take some investment (and comes with its own risk of course), alongside the always-clear-and-present threat of competition in the area of cutting-edge chip design. CEO Pat Gelsinger said the latter of these are still on track.

“Prudent allocation of our owners’ capital is important to enable our IDM 2.0 strategy and sustain our momentum as we rebuild our execution engine,” he said in a statement today. “We remain on track to deliver five nodes in four years and continue to expand the IFS (Intel Foundry Services) customer base. We are well into the ramp of 13th Gen Intel® Core™ and 4th Gen Intel® Xeon® Scalable processors, and we look forward to the launch of Meteor Lake and Emerald Rapids in 2023 and Granite Rapids and Sierra Forest in 2024.”

Intel in October 2022 was reportedly gearing up for thousands of job cuts in the quarter ahead. A spokesperson today confirmed that while it has reduced its workforce it has yet to confirm an exact number of people impacted. The company at the end of 2022 employed nearly 132,000 people. It’s also cut compensation for executives and managers, including a 25% cut for Gelsinger himself.

The company last quarter saw revenues decline 32% on the year before to $14 billion, which also missed analysts’ estimates. All eyes are now on how the company will be doing in the year ahead with current and future orders. Gelsinger dismissed recent reports alleging chip delays as “rumors” in a call today.

Intel slashes divdend by nearly two-thirds to shore up cash as chip giant braces for a tough year by Ingrid Lunden originally published on TechCrunch

Intel unveils high-end 13th-gen 24-core processors plus N-series workhorse to fill the the Pentium and Celeron gap

Intel is taking a more subdued approach to CES these days — forgoing a splashy event staged in a big hotel showroom in the wake of Covid-19, and a wider change in PR strategy after years of making bullish investments in next-generation tech like drones and moonshots like Volocopter and using them as showpieces at those events. Remember the year when Intel imported a whole Volocopter aircraft on to the stage, and placed its then-CEO into it, for its “first US flight”?

Yet the Vegas mega-show remains a key moment for Intel. It’s not just a bellwether for the state of the consumer electronics industry, but it’s an important marketing opportunity as a swathe of consumer electronics companies size up and buy in components for their devices. Today, the company unveiled a host of news related to processors and computer specifications using them, including a new 13th generation of its Intel Core processor, an all-new 24-core processor, the i9, and — addressing the fact that there is over-penetration of computers among business and developed world users — a new N-series specifically for what it describes as “entry-level” education and mainstream laptops, desktops and edge-native applications.

The breadth here is intentional: Intel made its name decades ago for its revolutionary approach to computer processors, which helped usher in a new generation of smaller devices, but it has arguably met some very stiff competition at the higher end of the market, and some would say missed the boat on mobile years ago. These new releases aim to address all of this: providing leadership in the bigger processing race of tomorrow but also hoping for a role in the making of devices for the mass market of today, not least after announcing in September 2022 that it would sunset its iconic Celeron and Pentium processor brands.

“The 13th Gen Intel Core mobile processor family delivers unrivaled, scalable performance for leadership platforms across all laptop segments,” said Michelle Johnston Holthaus, executive vice president and general manager of the Client Computing Group at Intel, in a statement. “With our industry-leading technologies and unmatched global partner ecosystem, people can expect a high-caliber mobile experience in new and unique form factors so they can game or create from anywhere.”

The 13th generation Intel core mobile processor family being unveiled today is spearheaded by the i9-13980HX, which is Intel’s first 24-core processor designed for laptops. Intel claims it is now the world’s fastest mobile (that is, laptop) processor clocking up speeds of 5.6 gigahertz (GHz) turbo frequency and 11% faster and 49% faster performance respectively for single-purpose and multitasking usage. As a measure of what the race is like in processors today, this is less about Intel really setting a new bar as much as it is about keeping up: it notes in a disclaimer that it’s worlds-fastest claim is only valid as of December 2022.

The 24 cores are divided up into 8 Performance-cores and 16 Efficient-cores, it says, and also are complemented by 32 threads and “enhanced Intel Thread
Director” with memory support of up to 128 gigabytes total covering two classes of SDRAM, DDR5 (up to 5,600 megahertz) and DDR4 (up to 3,200 MHz). The state of features today expected by consumers in these devices is laid bare too with a wide range of other support including superfast Wi-Fi 6E (Gig+) support; Bluetooth LE Audio and Bluetooth 5.2 support for faster speeds, multiple devices and lower power consumption (so critical given earlier Bluetooth does drain battery); Thunderbolt 4 support for 40 gigabits per second transfer speeds; and more.

The H-, P- and U-series mobile processors are addressing IoT, “enthusiast” and thinner devices. Intel says that more than 300 models from Acer, Asus, Dell, HP, Lenovo, MSI, Razer, Republic of Gamers, Samsung and others are going to be released this year based on the them.

All of the processors in the 13th generation will also include a Movidius vision processing unit (VPU), built in collaboration with Microsoft to integrate closely with its Windows Studio Effects to handle processing of more AI-based tasks to speed up overall CPU and GPU performance of machines. That collaboration is a notable mark of how hardware and software have had to tie up closer to evolve, and how hardware is becoming increasingly a software play, for more complex applications and faster speeds. Without its own chip-based vertical strategy in-house, Microsoft is an obvious partner.

“Together with Intel we continue to innovate to deliver powerful PC performance and experiences with Windows 11 and all of the products Intel is announcing today,” said Panos Panay, EVP and product head for Microsoft, in a statement. “We’re excited for customers to benefit from substantial optimizations, like improved Windows support for Intel Hybrid Guided Scheduler, and meaningful new experiences, like with the Intel Movidius VPU unlocking a new era of AI acceleration, starting with Windows Studio.”

Intel is describing its new N-series chips, meanwhile, as a direct replacement for the Penium and Celeron lines. “Purpose-built” for the education segment, entry-level computing and IoT edge-native applications, this also means that they will be marketed as more cost-effective and aimed and overall lower-priced and lower-specced devices, while being more modern than the previous generations and being a more evolutionary product for the company.

With new Gracemont-based cores, Intel 7 process technology means 28% better application performance and 64% better graphics performance at the peak compared to the older (now sunset) processors; up to 10-hours of HD video playback (if nothing else is being used) with better camera and display support as well as upgraded WiFi and Bluetooth (they are based also on the i3 tech). Intel said that some 50 new ChromeOS and Windows designs from Acer, Dell, HP, Lenovo and ASUS are due to be launched this year based on these chips.

IoT is also getting addressed with these new N-series chips, which will be appearing, Intel said, in devices used in retail signages, kiosks, point of sale systems, portable medical imaging devices, office automation equipment like copiers, and in safety and security devices.

In addition to the chip news, Intel has also continued iterating on its laptop and portable computing specifications, this year with new developments called Intel Evo.

These are based on the new 13th generation processor and focus on extended battery life to improve both the speed of charging but also how long devices can run unplugged; improved performance for videoconferencing and other video and collaboration applications; and better bridging between laptops and other keyboard computing and mobile handsets and tablets, which it’s terming “Intel Unison.” Again, in the endgame of vertical integration, this was an essential move for Intel, in an environment where those who do still use laptops are always doing it in complement with handsets, something that device makers are keen to make as easy as possible, not least to lose those users as customers of the former products.

Intel Evo will also work with hardware made by accessory providers, covering Thunderbolt 4 docks, monitors, storage and wireless headsets, mice, keyboards and other access points. Whether those will ultimately feel like gimmicks or buggy hardware that no one ultimately uses remains to be seen: at the end of the day, the easiest and most foolproof tools tend to win the day.

Read more about CES 2023 on TechCrunch

Intel unveils high-end 13th-gen 24-core processors plus N-series workhorse to fill the the Pentium and Celeron gap by Ingrid Lunden originally published on TechCrunch

Eliyan raises $40M from Intel and Micron to build chiplet interconnects

Increasingly, as Moore’s law rears its ugly head, computer chip developers are adopting “chiplet” architectures to scale their hardware’s processing power. Chiplets are Lego-like integrated circuit blocks designed to work with other, similar chiplets to form complex, stackable chips that boost performance while maintaining a similar physical footprint. Chiplets offer a number of advantages over conventional designs. But assembly issues — as well as challenges in balancing cost, performance, power consumption and time to market — often plague them in the early phases.

Aiming to overcome the hurdles in chiplet creation, Ramin Fajadrad, Syrus Ziai and Patrick Soheili founded Eliyan, a chiplet interconnect startup, in 2021. Eliyan’s technology — dubbed NuLink — connects chiplet components using standard chip packaging, leading to what the company claims are faster-performing and more energy-efficient chips.

“The focus is on developing a way to enable a more high-performance, lower-power and lower-latency interconnect for chiplet architectures, which experts agree is the only path to continuing to scale Moore’s law,” Farjadrad told TechCrunch in an email interview. “We use our technology in standard packaging, thus saving time, cost and development effort compared to more advanced packing that other interconnect schemes require. In addition, our approach has sustainability benefits by reducing material costs and waste in the manufacturing process and lowering energy consumption for high-performance compute chips.”

Eliyan’s roots are in a previous startup, Aquantia, that Marvell acquired in 2019. Farjadrad says the technology has been under development since 2017; he co-started Aquantia and served as the startup’s chief engineer for nearly 15 years. Prior to co-founding Eliyan, Farjadrad spent several years at Marvell as CTO and VP of the company’s networking and automotive division. Ziai is a former Qualcomm engineering VP, while Soheili was previously VP of business development at semiconductor firm eSilicon.

While Eliyan hasn’t launched its technology commercially yet — it expects the first silicon to hit the market in Q2 2023 — the company claims to have achieved the last step before manufacturing, a tape-out, using semiconductor manufacturer TSMC’s 5 nm process. “Process” in chip lingo refers to an architectural platform; TSMC began mass-producing 5 nm chips in 2020.

“Eliyan’s technology enables processors by allowing them to scale in performance and power to be more readily and practically manufacturable,” Farjadrad said. “The world will always need more computing power, and Eliyan is enabling a critical aspect of making sure scaling will happen for any type of high-performance computing application.”

The fact that Eliyan’s tech has yet to reach market might give some would-be customers pause. But the startup has notable investors in the chip world behind it, including Intel and Micron, who alongside Cerberus and Celestra contributed to Eliyan’s $40 million Series A tranche that closed today.

With the capital, Eliyan plans to continue chasing after a chiplet market that could be worth $50 billion in 2024 — specifically by ramping up testing and implementation. Farjadrad wouldn’t name clients, but said that Eliyan, which currently has a 21-person staff, is in discussions with “big semi companies, hyperscalers and AI processor startups.”

“We’re dealing with the challenges and realities of physics in designing and manufacturing advanced chips … [but we’re] in a high-demand market,” Farjadrad said. “Our technology will ultimately lead to faster, more efficient and cheaper high-performance computing to run data centers, cloud computing AI, graphics and more.” 

Eliyan raises $40M from Intel and Micron to build chiplet interconnects by Kyle Wiggers originally published on TechCrunch

Mobileye IPO warns of potential potholes in the road to autonomous driving

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 core business 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.

Mobileye IPO warns of potential potholes in the road to autonomous driving by Jaclyn Trop originally published on TechCrunch