By Bolaji Ojo

In market capitalization, Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC) is a distant second to customer Nvidia Corp., the world’s most valuable chipmaker.

Nvidia’s CEO Jensen Huang is on record for admitting his company – without the services provided by TSMC – would either not exist or be a completely different enterprise playing a much smaller role in the electronics value chain.

Not everyone in the electronics market would so bluntly state their reliance on the Hsinchu, Taiwan-based foundry. Apple Inc., the foundry’s biggest and perhaps most important customer accounting for about one-quarter of TSMC’s annual revenue, hardly talks about the supplier it relies upon for the majority of its semiconductor requirements.

The technology world may not admit this reality to themselves, but one company – TSMC – has become its most “invaluable” asset. TSMC is redefining how the industry operates and, in some ways, even its future. This is a fact the industry struggles with, hesitant to openly discuss what could turn into its biggest headache if a major disaster happens in Taiwan or if governments clash over the island’s future.

Read More »TSMC’s ‘Invaluable’ Status Makes it a Target. Change, it Must

By Junko Yoshida

Last Friday’s worldwide IT outage, traced to CrowdStrike, galvanized everyone from corporate board directors and their information security officers to IT managers, cyber security experts and the public at large.

The cause was a faulty update from CrowdStrike, deployed to computers running Microsoft Windows.

The defective updates, which grounded flights, disrupted banking and healthcare services and 911 emergency call centers, made the high-tech industries sit up and wonder:

How did we let this happen?

Read More »CrowdStrike’s Update Downfall: Who Dropped the Ball?

By Peter Clarke

What’s at stake:
At stake is the future of the most important part of the deep technology sector. That is, if you believe that the AI revolution currently permeating the world is going to build on — and be more important than — the transistor, the computer and the Internet. Nvidia is the dominant AI-hardware player seeking to expand in multiple directions. Masayoshi Son, the founder and CEO of SoftBank Group, wants to transform his company into an AI full-stack powerhouse.

Graphcore Ltd. (Bristol, England) has been acquired by Japan’s SoftBank Group, finally confirming rumors of a deal that had circulated for several months.

We don’t know the purchase price but it is expected to have been less than the $700 million of venture capital raised by Graphcore since its formation in 2016; speculations put the price at about $500 million. As such, this feels like a distressed sale despite the positive gloss that CEO Nigel Toon has put on the deal.

Read More »How Does Graphcore Fit into SoftBank’s AI Play?

By Junko Yoshida

The software-defined vehicle (SDV) is “all the rage,” if you are to believe press-release headlines and media coverage (including this publication).

A voice of prudence is Phil Koopman, professor at Carnegie Mellon University, who recently published an article, “Architectural Coupling Killed the Software Defined Vehicle” on his Substack newsletter.

Despite its provocative headline, Koopman writes, “I don’t think the SDV is actually dead.”

However, as carmakers stampede toward the SDV cliff, Koopman warns that they might be throwing together vehicle architectures that inevitably become too complex to manage.

Read More »The SDV and Its Unintended Challenges

By Junko Yoshida

It’s easy to pick on Volkswagen.

The dizzying array of Volkswagen’s recent partnerships, joint ventures and investments is all over the map – in geography and business focus. Worse, all this churn shows no apparent thread, at least to outsiders like myself, that might help VW to knit the tangle together. 

Paired with VW’s frequent reorganizations and management changes in recent years, such deals beg the conclusion that the organization is in disarray, throwing spaghetti on the wall to see what sticks. Maybe things aren’t that bad. Maybe it just looks bad.

Read More »VW’s Software Crisis isn’t VW’s Alone

By Junko Yoshida

What’s at stake:
Eager beavers in the auto industries skipped some fundamental technology and business steps in their rush to bring autonomous vehicles to the market. As they go back to the drawing board, can OEMs redesign an E/E vehicle architecture for SDVs that speeds up automotive innovation? 

There is ample evidence that the heralded rollout of advanced autonomous vehicles has stalled. By any measure, it isn’t “just around the corner.”

Although I hesitate to knock the engineering community’s ambitions to develop vehicle autonomy, I can respectfully suggest that tech suppliers, automakers, thinktanks, politicians, marketers and media come clean on AVs. A collective intellectual dishonesty has blinded us to the reality that autonomous vehicles were never close to being ready for prime time. The result is a pervasive consumer distrust in self-driving cars.

Read More »Industry Pivots from Self-Driving to Software-Defined

By Junko Yoshida

We still don’t know a whole lot about what AI PC exactly does. Nor do we understand what consumers should expect. 

One thing is clear. Afflicted with AI fever, the tech industry that contributed to creating AI PCs and Copilot+ PCs is overtly jubilant. 

Fueled by AI’s explosive growth and pressured by Wall Street, business and tech-world CEOs are eagerly styling themselves as passionate users and advocates for AI.

Intel Corp. CEO Pat Gelsinger compares AI PC’s impact to Wi-Fi’s emergence on Intel’s Centrino platform. Lisa Su, CEO of Advanced Micro Devices Inc., told reporters last week that she’s “an avid user of GPT, Copilot too.” She added that AMD wants “to put AI through the development pipeline, as well as marketing, sales, HR, all of those. It is going to be AI everywhere.”  

Given this enthusiasm among smart executives, seriously, what could go wrong?

Read More »Microsoft’s and Apple’s AI-Mighty Algorithm

By Peter Clarke

What’s at stake:
There was a time when Computex in Taiwan, the land of the PC motherboard manufacturers, was a slightly tired annual event. But not in 2024, the year of the advent of the AI-enabled personal computer – or as Microsoft would brand it, the CoPilot+ PC. This year there are many parts in motion and that means there are likely to be big changes in the winners and losers in the supporting semiconductor cast of characters.

In the mid-1990s when Robin Saxby, then CEO of processor IP licensor ARM, told me that his company’s technology was in 70 percent of all mobile phones I realized that startup company ARM had become a global player. The current CEO, Rene Haas, is claiming that ARM could claim more than 50 percent of the Windows PC market by 2029. If that happens it will likely mean that Intel’s time as a global player is over. That is unless Intel itself becomes a maker of PC processors based on the ARM architecture.

It could happen.

Read More »Microsoft’s CoPilot+ Puts Intel on a Knife-Edge

By Junko Yoshida

What’s at stake:
The new frontier in AI is in multimodal models such as CLIP and Stable Diffusion. Because humans interact with the world through both vision and language, AI researchers believe machines, too, need multimodal channels.  Can something like LLaVA foster a “general-purpose assistant” that effectively follows multimodal vision-and-language instructions aligned with human intent? If so, at what cost?

We live in a tumultuous time.

Business reporters are often pressed to investigate commercial and technological advancements almost immediately after the scientific community breaks new ground. This is particularly true in the field of artificial intelligence.

With little time either for reflection or examination, reporters take scientists’ word on the Next Big Thing, reducing AI journalism to little better than stenography.  Remember when its promoters kept saying the autonomous vehicle is “just around the corner.”

Corporations, obliged to become AI prospectors, now face a similar dilemma. Like the 49ers of the Gold Rush, the first priority for a company hoping to strike it rich, is to stake a claim, only worrying afterward whether its particular AI investment is the mother lode or a dry hole.

Clouding our judgement further is the exponential growth of Nvidia. Money talks.

Read More »It’s a Multimodal World, After All

By Peter Clarke

What’s at stake:
Arm, founded in 1990, more or less invented the business model of licensing circuits and computing architectures as intellectual property. It achieved success in what were then new, embedded markets, such as mobile phones, industrial and automotive electronics. The Arm architecture is now penetrating servers, AI, PCs and high-performance computing. But all is at risk if SoftBank insists on using Arm to compete with customers such as Nvidia. That is unless you consider Arm’s licensing business model will soon be on the wane due to a flight to RISC-V.

Nikkei reported last week that Arm Holdings plc is planning to set up an AI chip division and have an AI processor on sale in the Fall of 2025. And that this would be the behest of majority shareholder SoftBank Group and its CEO Masayoshi Son.

Arm has not previously sought to compete with its licensees and its independence was always seen as a strength. If SoftBank insists on indulging in vision-driven tinkering with the Arm business model, it could hasten the rise of the open-source, extensible RISC-V ISA as the go-to alternative to Arm.

Read More »Could SoftBank’s Son Kill Arm with His AI Vision?